Search

• A deep learning algorithm is used to build a comprehensive drought monitoring model. • Multi-source remote sensing data is processed as input data. • The comprehensive drought model has good applicability in the monitoring of meteorological drought and agricultural drought. • The model successfully monitored a drought process in Henan Province, China. • This study provides a new method for the comprehensive assessment of regional drought. Drought is a popular scientific issue in global climate change research. Accurate monitoring of drought has important implications for the sustainable development of regional agriculture in the context of increasingly complex global climate change. Deep learning is a widely used technique in the field of artificial intelligence. However, ongoing on drought monitoring using deep learning is relatively scarce. In this paper, the various hazard factors in drought development were comprehensively considered based on satellite data including Moderate Resolution Imaging Spectroradiometer (MODIS) and tropical rainfall measuring mission (TRMM) as multi-source remote sensing data. By using the deep learning technique, a comprehensive drought monitoring model was constructed and tested in Henan Province of China as an example. The results showed that the comprehensive drought model has good applicability in the monitoring of meteorological drought and agricultural drought. There was a significant positive correlation between the drought indicators of the model output and the comprehensive meteorological drought index (CI) measured at the site scale. The consistency rate of the drought grade of the two models was 85.6% and 79.8% for the training set and the test set, respectively. The correlation coefficient between the drought index of the model and the standard precipitation evapotranspiration index (SPEI) was between 0.772 and 0.910 (P < 0.01), which indicated a strong level of significance. The correlation coefficient between the drought index of the model and the soil relative moisture at a 10 cm depth was greater than 0.550 (P < 0.01), and there was a good correlation between them. This study provides a new method for the comprehensive assessment of regional drought. [ABSTRACT FROM AUTHOR]

Abstract China's environmental problems are not only related to the fundamental interests of the broad masses of the people, but also to China's national security and international image. At present, China's environmental protection work is facing a complex situation. Pollution sources can be divided into natural pollution sources and man-made pollution sources. Natural sources of pollution refer to places where nature releases harmful substances or causes harmful effects to the environment, such as active volcanoes. Man-made pollution source refers to the pollution source formed by human activities, and is also the main object of environmental protection research and control. Among the man-made pollution sources, air pollution sources, water pollution sources and soil pollution sources can be classified according to the main objects of pollution. Among them, air pollution sources and water pollution sources have the greatest impact on human life. Therefore, it has become an important subject worthy of in-depth discussion to take automatic and electronic measures for potential environmental pollution incidents, discover environmental pollution problems in time, reduce the probability of environmental pollution incidents, and even put some major environmental pollution incidents in their infancy. In this paper, deep learning method is used to classify the existing key pollution source video. Water pollution experiments show that the accuracy of video counting reaches 93.1%, which is better than other video processing schemes. The operation time of the system reaches acceptable range, and a solution to meet the real-time requirement is put forward. [ABSTRACT FROM AUTHOR]

• A multi-view field-road classification method was proposed for GNSS trajectories. • Trajectories were converted to trajectory images to encode rich motion information. • Pixel-level visual features were extracted by a specific image segmentation model. The understanding of agricultural machinery, whether it is used "in field" or "on road", plays an essential role in optimizing the efficiency of cross-regional agricultural mechanization services that have been available in China for decades. With the widespread availability of Global Navigation Satellite System (GNSS)-enabled devices, developing a method for automatically identifying the activity associated with each point in GNSS recorded trajectories can significantly enhance the optimization of these services. In this paper, we propose a novel field-road classification method that utilizes two feature vectors to represent each point in a GNSS trajectory. The first feature vector is a statistical feature vector extracted with existing methods, while the second feature vector is a visual feature vector obtained through an image segmentation model. To extract visual features, we first convert each GNSS trajectory into a trajectory image that encodes the motion information (e.g., speed, direction) of each point and the spatiotemporal relationship between points. Then, an image segmentation model specifically for field-road classification was developed, which can effectively extract a visual feature vector for each pixel in a trajectory image, corresponding to each point in the GNSS trajectory. The image segmentation model used in our approach has a robust feature extraction capability that enhances the statistical feature representation of each point by adding pixel-level visual features. We evaluated the effectiveness of our multi-view field-road classification method using trajectories of corn, wheat, and paddy harvesters. The results demonstrated that our method achieved accuracy scores of 94.17%, 90.93%, and 83.43% for respective datasets. Our method outperformed state-of-the-art field-road classification methods, especially for trajectories with high-frequency GNSS acquisition. [ABSTRACT FROM AUTHOR]

[Display omitted] • We developed an ensemble mineral prospectivity mapping method based on attention-based convolutional neural network. • The proposed ensemble method improves mineral prospectivity modeling accuracy by learning the key channel feature attentively. • Mineral potential map of W–Sn deposits in Nanling metallogenic belt was produced using the proposed method. Data-driven mineral prospectivity mapping (MPM) based on deep learning methods has become a powerful tool for mineral exploration targeting in the past years. Convolutional neural networks (CNNs) have shown great success in this field because of their powerful ability to capture the complex spatial geo-anomalies related to mineralization. However, the exploration big data applied to MPM mainly relies on the high dimensions of evidence layers (other than spatial dimensions), namely, a large number of channels. This impedes the extraction of key channel features related to mineralization when using traditional CNNs. In this paper, we developed an ensemble MPM method based on CNN and Attention model: the ATT–CNN method. Specifically, a channel attention layer is added after the convolution operation of the CNN to enhance the extraction of key channel features in complex exploration data, thereby improving the feature extraction ability and prediction accuracy of CNN for MPM. A case study of W–Sn mineral prospectivity modeling in the Nanling metallogenic belt in South China was used to verify the proposed method. To alleviate the issue of training sample scarcity, we used data augmentation methods (including sliding window and random zero noise addition) when training CNN models. The results show that the prediction accuracies of the ATT–CNN model (92.949% and 94.872% using sliding window and random zero noise addition, respectively) are higher than those of the traditional CNN (91.667% and 92.308%, respectively). Moreover, the improved areas under the receiver operating characteristic curves (AUC) of ATT–CNN (0.987 and 0.971) compared to those of the CNN (0.970 and 0.964) suggest that the proposed ensemble method improves the geological generalization of CNN. The high agreement with known deposits suggests that the areas targeted in this study can guide future mineral exploration of the W–Sn mineralization in the Nanling range. [ABSTRACT FROM AUTHOR]

Laser scanning is used as a modern means to capture data from tunnels to assess their condition, but automated processing requires robust component detection and deterioration characterization. In order to segment 3D tunnel point clouds aiming at more accurate results with high time efficiency, this paper describes a point cloud technique that collects actual tunnel scenes and develops an attention-enhanced sampling point cloud network named ASPCNet. In the developed model, the feature embedding module is responsible to process the point cloud data for local features followed by the attention module for enhancing the feature extraction and learning. Additionally, the point downsampling-upsampling structure fully assists the model to strengthen the capability to process point clouds for time efficiency. In the training process, a weighted focal loss is designed to enhance the model learning by eliminating the effect of data imbalance. The developed ASPCNet is trained and then tested on a dataset collected from a cross-river metro tunnel section in China, demonstrating its efficiency and effectiveness. In comparison with different sampling ratios, state-of-the-art methods, and sampling methods, the ASPCNet with a uniform sampling rate of 2 exhibits the best performance, achieving an overall accuracy of 0.9758, a mean Intersection over Union (MIoU) of 0.8988, and an inference time of 4.1 s, demonstrating that the sampling structure involved in this research boosts the time efficiency, the developed model has superior performance, and the sampling method adopted is beneficial to strengthen the model performance. • A deep learning method named ASPCNet is developed for 3D point cloud segmentation. • It consists of several modules to process point clouds for accuracy and time efficiency. • A focal loss is designed to handle data imbalance and boost feature learning. • Effectiveness and efficiency are verified on a dataset with six classes of tunnel point clouds. • The developed method performs excellently with high accuracy and great efficiency. [ABSTRACT FROM AUTHOR]

• Combined with the error reconstruction rate, an adaptive VMD technology is developed and used for data preprocessing. • Two mainstream deep learning methods (LSTM, DBN) are used as sub-series predictors, and the classic PSO algorithm is utilized to determine the number of neurons in the DBN network. • A novel hybrid prediction model based on the PSO-DBN nonlinear combination mechanism is proposed. Wind power forecasting plays a vital role in enhancing the efficiency of power grid operation and increasing the competitiveness of power market. In this paper, a novel hybrid forecasting model is developed by using the decomposition strategy, nonlinear weighted combination, and two deep learning models to overcome the drawbacks of the linear weighted combination and further enhance wind power forecasting accuracy and stability. Firstly, the variational mode decomposition (VMD) technique is employed to decompose the original wind power series to extract local features. Then, the long short-term memory neural networks (LSTM) and deep belief networks based on particle swarm optimization (PSO-DBN) are utilized to construct sub-series prediction models. Finally, the multiple sub-series forecasting models are integrated by a nonlinear weighted combination method based on PSO-DBN to construct a hybrid model for short-term wind power forecasting. To verify the performance of the developed forecasting model, wind speed data of 10 min from a wind farm in Shaanxi Dingbian, China are selected as case studies. The results show that the proposed method in this paper is more effective than other existing methods. [ABSTRACT FROM AUTHOR]

• Plug-in hybrids with dual-mode engines greatly reduce fuel consumption and emissions. • A hierarchical energy management strategy is proposed for close-to-optimal control. • Compared with mainstream control methods under standard and real-world drive cycles. • Generalization and robustness to random initial and terminal conditions are proved. The dedicated hybrid engines (DHEs) with dual-mode combustion technology can drastically reduce the fuel consumption and emissions while guarantee the power density. This paper aims to investigate the optimal control of such DHE-based plug-in hybrid electric vehicles (PHEVs) under real driving conditions, with minimum fuel penalties caused by transient engine dynamics. For this purpose, the benefits brought by artificial intelligent control and traffic preview in terms of energy efficiency can be combined with the advantages of advanced combustion engine. This paper presents a hierarchical energy management strategy (HEMS) to realize the synergy of global and instantaneous optimization. At the cloud level of HEMS, dynamic programming is applied to obtain optimal combustion mode and state of charge reference trajectories in a receding horizon. At the powertrain level, deep reinforcement learning with a ranking-prioritized experience replay algorithm is used to output optimal engine power and combustion mode for the energy management. To evaluate the proposed strategy, a dual-mode engine with homogeneous charge compression ignition and spark ignition systems is tested and mapped, with which the PHEV is modeled in GT-Suite and Matlab/Simulink. Comprehensive experiments are carried out to verify the optimality, generalization and robustness based on a standard driving cycle and a real-world driving cycle in China with GPS data recorded. The results show that the HEMS avoids frequent switching of combustion modes and outperforms the conventional methods by more than 4% and 10% in terms of fuel economy and NOx emissions, respectively, with random initial and terminal conditions. [ABSTRACT FROM AUTHOR]

• A hybrid deep learning model within decomposition-ensemble strategy named DEASeq2Seq is proposed for multistep metro passenger flow prediction. • DEASeq2Seq employs CEEMDAN and RQA to capture intrinsic dynamics. • An attention-based Seq2Seq model is developed to characterize temporal dependencies. • Real-world datasets to verify the proposed model and further model interpretation. Short-term passenger flow prediction has practical significance for metro management and operation. However, the complex nonlinear and non-stationary characteristics make it challenging to detect evolution characteristics of passenger flow. To address this problem, a hybrid short-term metro passenger flow prediction model named decomposition ensemble attention sequence to sequence (DEASeq2Seq) is proposed in this paper. The proposed DEASeq2Seq includes three phases: decomposition, ensemble, and prediction. First, complete empirical mode decomposition with adaptive noise (CEEMDAN) is utilized to decompose the original passenger flow data into several intrinsic mode functions (IMFs) and a residue. Second, recurrence quantification analysis (RQA) is performed to reconstruct the decomposed modes into a stochastic part, a deterministic part, and a trend part via determinism evaluation. Third, a Seq2Seq model with the attention mechanism is proposed to execute multistep prediction for short-term passenger flow and explore the influence mechanism of the reconstructed components on the prediction targets. The real dataset from Chengdu metro, China, is used to verify the proposed model. The experiment results show that the proposed DEASeq2Seq model outperforms the benchmark models. Further model interpretations are conducted to analyze the impacts of decomposition strategy, ensemble strategy, and attention mechanism. [ABSTRACT FROM AUTHOR]

• A detection approach for late-autumn shoots of litchi tree based on UAV remote sensing is provided. • A remote sensing dataset of late-autumn shoots of litchi tree is constructed. • A deep learning model based on the combination of YOLOv5 and Swin-transformer is proposed. • The detection accuracy of late-autumn shoots in remote sensing images is improved. • It will provide technical support for other researchers on late-autumn shoots and orchards manager. Litchi is one of the most common economic fruits in southern China, however, the growth of late-autumn shoots of litchi hinders flower bud differentiation and reduces yield of fruit. The early identification of the late-autumn shoots is of great significance for orchard management to control shoots and then increase fruit yield. At present, the identification of late-autumn shoots still relies on manual methods, which is not suitable for smart orchard management in a large area due to low recognition efficiency and high subjectivity. Therefore, a convenient, fast and cost-effective method is urgently needed. In response to this problem, the paper proposes a method based on the combination of unmanned aerial vehicle (UAV) remote sensing and object detection algorithm to detect late-autumn shoots. For this purpose, a remote sensing dataset of late-autumn shoots of litchi is first constructed by UAV. An improved YOLOv5 algorithm called YOLOv5-SBiC is then developed for late-autumn shoots identification. In the YOLOv5-SBiC algorithm, the transformer module is introduced to speed up the convergence of the network and improve detection accuracy, the attention mechanism module is employed to help the model extracting details, and BiFPN is used to better solve the multi-scale problem in detecting and then improve the recognition effect of small-sized objects. In addition, CIOU is selected as the loss function of bounding boxes regression to achieve high-precision localization of the boxes. The test results demonstrate that the recognition accuracy of YOLOv5-SBiC reaches a relatively high value of 79.6%, which is 4.0% higher than that (75.6%) of the original YOLOv5 algorithm and 15.9% higher than that (63.7%) of the pure transformer algorithm. It's also demonstrated that YOLOv5-SBiC is more competitive than the mainstream target detection algorithms in the dataset of late-autumn shoots. [ABSTRACT FROM AUTHOR]

[Display omitted] • The measurement errors in long-time series remote sensing images are corrected by statistical and spatio-temporal prediction model. • The ecological degradation of the Miyun Reservoir basin from 1985 to 2021 is mainly caused by anthropogenic factors. • Through statistical theory, the ecological development trend in the experimental area is analyzed, and the conclusion is verified by land use evolution. Miyun Reservoir, located in the Miyun District, Beijing, China, is the largest comprehensive water conservancy project and is an important ecological protection area in the North China region. Changes within the basin are the driving factors affecting the ecosystem in the watershed; therefore, it is important to analyze the changes in the ecological environment of Miyun Reservoir. For the analysis of a long time series of image data remotely sensed by satellite, the outliers caused by atmospheric, lighting, and sensor measurement errors are significant, and it is difficult for traditional algorithms to effectively recover the true image value. To address this, this paper proposes a theoretical model for predicting spatio-temporal variation based on deep learning to identify and correct invalid and anomalous values in extended time series data. This study corrected and analyzed the results of Remote Sensing based Ecological Index inversion of Landsat data of the Miyun Reservoir watershed from 1985 to 2021. The findings and conclusions of this study are important for the analysis of long time series image data from satellite remote sensing and for improving regional ecological evaluation and sustainable development planning. [ABSTRACT FROM AUTHOR]

With the rapid development of economy, air pollution occurs frequently, which has a huge negative impact on human health and urban ecosystem. Air quality index (AQI) can directly reflect the degree of air pollution. Accurate AQI trend prediction can provide reliable information for the prevention and control of air pollution, but traditional forecasting methods have limited performance. To this end, a dual-scale ensemble learning framework is proposed for the complex AQI time series prediction. First, complete ensemble empirical mode decomposition adaptive noise (CEEMDAN) and sample entropy (SE) are used to decompose and reconstruct AQI series to reduce the difficulty of direct modeling. Then, according to the characteristics of high and low frequencies, the high-frequency components are predicted by the long short-term memory neural network (LSTM), and the low-frequency items are predicted by the regularized extreme learning machine (RELM). At the same time, the improved whale optimization algorithm (WOA) is used to optimize the hyper-parameters of RELM and LSTM models. Finally, the hybrid prediction model proposed in this paper predicts the AQI of four cities in China. This work effectively improves the prediction accuracy of AQI, which is of great significance to the sustainable development of the cities. • A dual-scale ensemble framework is proposed for air quality index prediction. • CEEMDAN technology is adopted to obtain multiple intrinsic mode functions. • An improved WOA algorithm based on dimension learning-based hunting is proposed. • Long short-term memory network is used to predict high-frequency components. • Regularized extremum learning machine is used to predict low-frequency components. [ABSTRACT FROM AUTHOR]

Reliable weather forecasting is quite important for sectors with weather-dependent decision-making. To capture unknown and complex dependencies from multiple spatiotemporal factors, i.e., autocorrelated, meteorological causal, meteorological spatial factors, this paper proposes a spatiotemporal feature extraction and fusion network inspired by humans' multi-level cognition process, namely, SFEF-Net, including the data pre-processing stage, transfer pre-training stage, spatiotemporal feature extraction stage, and spatiotemporal feature fusion stage. Specifically, inspired by humans' empirical analysis, the 24 Solar Terms knowledge and Pearson correlation coefficient analysis are utilized to respectively select input timestamps and influencing factors in the first stage. Inspired by humans' associative learning, a Long Short-Term Memory Network (LSTM) layer with powerful remember capability is trained via weather data from multiple neighbored monitoring stations in the second stage, and further is transferred for better feature extraction. Inspired by humans' independent analysis, every factor is separately captured and analyzed via the pre-trained LSTM layer in the third stage, which also coincides with the philosophical idea of divide and conquer. Inspired by humans' comprehensive consideration, extracted features are fused level by level via the attention mechanism with dynamic weights in the fourth stage, which avoids invalid dependencies between multiple features and overfitting. The real-world daily average air temperature forecasting in Shaanxi Province, China is taken to demonstrate the SFEF-Net beyond 12 baseline methods. Additionally, forecasts of drastic change period, ablation analysis, multiple steps, and different stations further present the effectiveness of the SFEF-Net. [ABSTRACT FROM AUTHOR]

One of the key and difficult points in seismic data processing is seismic data denoising. Under the influence of the acquisition environment, the collected seismic data usually have low signal-to-noise ratio (SNR) and low resolution. In desert region of western China, different from other regions, the random noise there has complex characteristics of non-Gaussian, non-stationary and non-linearity. Its main frequency is quite low and its spectral overlap with those of effective signals. Moreover, large amount of data and intelligent requirements make traditional denoising methods encounter difficulties. In addition, the existing deep learning denoising methods also reveal two problems: first, they can only effectively suppress simple seismic noise such as Gaussian noise. But for the case of desert noise, they usually mistakenly judge noise as effective signals and destroy signal characteristic structure. Second, what is more important, in supervised-based methods paired noisy and pure data are usually used as training set. However, there is no noise-free data in actual desert seismic data, which severely limits the processing performance of supervised-based networks. Thus, a new parameter-shared variational auto-encoding adversarial network (PS-VAAN) is proposed for desert seismic data denoising in this paper. This new method includes two encoders, two generators and two discriminators. It can realize domain conversion from noisy data domain to pure data domain. We construct relatively complete unpaired pure and noisy data training sets for training, and design reconstruction loss function and adversarial loss function to optimize network parameters. Moreover, cycle-consistency loss is introduced to make sure two domains mapping into the same space. Compared with the state-of-art denoising methods on synthetic and actual seismic records, the proposed method not only has superior denoising ability, but also has strong feature extraction and fitting ability to recover effective signals with almost no energy loss. • Desert seismic noise has the characteristics of low frequency and non-stationary. • This paper designs a parameter-shared variational auto-encoding adversarial network. • The network use unpaired noisy and pure training sets to optimize parameters. • The proposed method has a strong ability to recover effective signals from noise. [ABSTRACT FROM AUTHOR]

The research of traffic revitalization index can provide support for the formulation and adjustment of policies related to urban management, epidemic prevention and resumption of work and production. This paper proposes a deep model for the prediction of urban Traffic Revitalization Index (DeepTRI). The DeepTRI builds model for the data of COVID-19 epidemic and traffic revitalization index for major cities in China. The location information of 29 cities forms the topological structure of graph. The Spatial Convolution Layer proposed in this paper captures the spatial correlation features of the graph structure. The special Graph Data Fusion module distributes and fuses the two kinds of data according to different proportions to increase the trend of spatial correlation of the data. In order to reduce the complexity of the computational process, the Temporal Convolution Layer replaces the gated recursive mechanism of the traditional recurrent neural network with a multi-level residual structure. It uses the dilated convolution whose dilation factor changes according to convex function to control the dynamic change of the receptive field and uses causal convolution to fully mine the historical information of the data to optimize the ability of long-term prediction. The comparative experiments among DeepTRI and three baselines (traditional recurrent neural network, ordinary spatial–temporal model and graph spatial–temporal model) show the advantages of DeepTRI in the evaluation index and resolving two under-fitting problems (under-fitting of edge values and under-fitting of local peaks). [ABSTRACT FROM AUTHOR]

• A deep learning COVID-19 diagnostic system is developed using state-of-the-art deep learning architectures. • A detailed analysis of different methods used for detecting COVID-19 is presented in this paper. • Chest X-rays from different sources are collected to build a robust classification model. • The developed diagnostic model provided efficient results, even with a limited number of input images. • The results produced in this work are validated by an expert radiologist. In recent months, a novel virus named Coronavirus has emerged to become a pandemic. The virus is spreading not only humans, but it is also affecting animals. First ever case of Coronavirus was registered in city of Wuhan, Hubei province of China on 31st of December in 2019. Coronavirus infected patients display very similar symptoms like pneumonia, and it attacks the respiratory organs of the body, causing difficulty in breathing. The disease is diagnosed using a Real-Time Reverse Transcriptase Polymerase Chain reaction (RT-PCR) kit and requires time in the laboratory to confirm the presence of the virus. Due to insufficient availability of the kits, the suspected patients cannot be treated in time, which in turn increases the chance of spreading the disease. To overcome this solution, radiologists observed the changes appearing in the radiological images such as X-ray and CT scans. Using deep learning algorithms, the suspected patients' X-ray or Computed Tomography (CT) scan can differentiate between the healthy person and the patient affected by Coronavirus. In this paper, popular deep learning architectures are used to develop a Coronavirus diagnostic systems. The architectures used in this paper are VGG16, DenseNet121, Xception, NASNet, and EfficientNet. Multiclass classification is performed in this paper. The classes considered are COVID-19 positive patients, normal patients, and other class. In other class, chest X-ray images of pneumonia, influenza, and other illnesses related to the chest region are included. The accuracies obtained for VGG16, DenseNet121, Xception, NASNet, and EfficientNet are 79.01%, 89.96%, 88.03%, 85.03% and 93.48% respectively. The need for deep learning with radiologic images is necessary for this critical condition as this will provide a second opinion to the radiologists fast and accurately. These deep learning Coronavirus detection systems can also be useful in the regions where expert physicians and well-equipped clinics are not easily accessible. [ABSTRACT FROM AUTHOR]

Conventional time-series prediction methods for natural gas consumption mainly focus on capturing the temporal feature, neglecting static and dynamic information extraction. The accurate prediction of natural gas consumption possesses of paramount significance in the normal operation of the national economy. This paper proposes a novel method that resolves the deficiency of conventional time series prediction to address this demand via designing a hybrid deep learning framework to extract comprehensive information from gas consumption. The proposed model captures static and dynamic information via encoding gas consumption as matrices and extracts long-term dependency patterns from time series consumption. Subsequently, a customised network is proposed for information fusion. Cases from several different regions in China are studied as examples, and the proposed model is compared with other advanced approaches (such as long short-term memory (LSTM), convolution neural network long short-term memory (CNN-LSTM)). The mean absolute percentage error is reduced by a range of 0.235%–10.303% compared with other models. According to the comparison results, the proposed model provides an efficient time series prediction functionality. It is also proved that, after effectively extracting comprehensive information and integrating long-term information with static and dynamic information, the accuracy and efficiency of natural gas consumption prediction are greatly promoted. A sensitivity analysis of different modules combination is conducted to emphasise the significance of each module in the hybrid framework. The results indicate that the method coupling all these modules leads to significant improvement in prediction accuracy and robustness. [ABSTRACT FROM AUTHOR]

The Hangjinqi area is a significant gas exploration area from the northern Ordos Basin, an important petroliferous basin in China. Although a considerable amount of achievements has been made in the natural gas exploration field, the controversy about the hydrocarbon generation characteristics of source rocks has raged unabated for ten years. In addition, few studies have systematically investigated hydrocarbon expulsion potential in the Hangjinqi area. In this paper, we adopted a series of new approaches to assess the hydrocarbon generation and expulsion potential of source rocks in Hangjinqi. Specifically, we analyzed the rock pyrolysis, kerogen composition, and biomarker data and applied Convolution Neural Network (CNN) in the field of deep learning to determine organic matter abundance, type, maturity and depositional environment. Then, two techniques, the hydrocarbon generation kinetics method and the hydrocarbon generation potential method, were implemented to obtain both the hydrocarbon generation and expulsion quantity. The results reveal that the new technique CNN is suitable for TOC prediction, and the performance is improved largely compared to traditional methods, e.g., Back Propagation neural network (BP) and △log R. Rock pyrolysis data, kerogen macerals and biomarker data reveal that the source rocks are gas-prone and dominated by type Ⅲ. By the integrating hydrocarbon generation potential method and hydrocarbon generation kinetics method, the hydrocarbon generation quantity of the study area is 20.35 × 1012m3. The results of the hydrocarbon potential method indicate that about 23.7 × 108t and 0.74 × 108t hydrocarbons were expulsed from coal and mudstone, respectively. According to comprehensive analysis, the hydrocarbon generation and expulsion quantities in the Hangjinqi area are extensive, indicating good prospects for natural gas exploration. These findings have significant implications for the understanding of natural gas resources potential and the distribution of favorable natural gas areas. • We propose Convolutional Neural Network (CNN) to calculate TOC content. • Two methods were used to assess the hydrocarbon generation and expulsion potential. • The debated issue of hydrocarbon generation potential in Hangjinqi was resolved. • Hydrocarbon expulsion quantity in Hangjinqi was calculated for the first time. • Comprehensive geochemical data in Hangjinqi area are provided. [ABSTRACT FROM AUTHOR]

• Our method has the potential to extract water with high accuracy on a large scale. • Our method realizes the conversion from unlabeled learning to noisy label learning. • The model performance is significantly higher than traditional methods. • Our method provides a new idea for unsupervised learning. Large-scale monitoring of surface water bodies is of great significance to the sustainable development of regional ecosystems. Remote sensing is currently the main method of global earth observation. On the one hand, the traditional water index is simple and efficient, but it relies on a fixed global threshold, which leads to low accuracy for water extraction. On the other hand, deep learning has achieved state-of-the-art results in classification of spectral data, but it consumes a substantial amount of manpower and time to label sufficient high-quality samples. Furthermore, spectral characteristics of water bodies in different areas vary greatly due to changes in atmospheric conditions and viewing geometry. In this paper, we propose a new accurate water extraction framework based on unsupervised deep learning and NDWI of multispectral images. Binarized NDWI images are used to identify potential water bodies, and deep learning training is performed using these pseudo samples and labels. This process realizes the conversion from unlabeled learning to noisy label learning. First, we proposed a simple and fast binarization algorithm to segment as many real water bodies as possible from NDWI images. Then a set of water confidence assessment rules was constructed from the four aspects of the spectrum, shape, agglomeration, and range. The water segments were scored and sorted to make the model start learning from easy samples and gradually advance to complex samples. Finally, the adjusted co-teaching learning strategy is adopted to filter errors introduced by noisy labels through peer networks with different learning capabilities during training. We tested the accuracy of our method using Gaofen Image Dataset (GID), Sentinel-2 and Landsat images of several water bodies in China. Compared with other methods, our method improved the F1 score by 18.1–40.3% and 6.8–22.2% for GID and Sentinel-2 images, respectively. In addition, our method is more stable in long-term water monitoring. The proposed method has the potential to be used for extracting water bodies with high accuracy on a large scale, especially in areas with complex environments and a lack of deep learning samples. And it provides a new idea for unsupervised learning in the current remote sensing field by fully combining remote sensing expertise and spectral information of ground objects. [ABSTRACT FROM AUTHOR]

In the desert region of northwest China, the frequency of wind-sand disasters is high. All types of concrete buildings built in this area face severe wind erosion due to high wind speed, resulting in varying degrees of wind-erosion damage to concrete. To accomplish intelligent identification of concrete wind-erosion damage, a concrete wind erosion experiment was conducted in the laboratory, and a concrete wind-erosion damage dataset was generated under the interference of water stains, scratches, shooting distance, and background noise. This paper combined with transformer theory to improve YOLO-v4 and proposed an object detection algorithm called MHSA-YOLOv4 suitable for wind-erosion damage of concrete. The results demonstrate that MHSA-YOLOv4 exhibits improved object detection performance than YOLO-v3, improved YOLO-v3, and YOLO-v4. On the test set, ACC, Precision, Recall, and mAP of MHSA-YOLOv4 are 91.30%, 91.52%, 92.31%, and 0.89, respectively. MHSA-YOLOv4 can accurately identify wind-erosion damage of concrete images under different test conditions, which reflects strong robustness. The applicability of computer vision technology to the intelligent identification of wind-erosion damage on concrete has been verified. • The dataset of wind-erosion of concrete was established to simulate the spalling of concrete. • A object detection algorithm called MHSA-YOLOv4 was proposed based on Transformer. • MHSA-YOLOv4 can detect concrete wind-erosion damage with high-precision. [ABSTRACT FROM AUTHOR]

It will be beneficial to devise an effective approach for short-term macroeconomic forecasting. Existing traditional statistics-based macroeconomic forecasting mainly focuses on exploring feasible methods for improving the accuracy of long-term predictions. However, the performance of short-term predictions was far less impressive under the impact of unexpected incidents. Furthermore, some deep learning-based approaches can achieve fine-grained variable frequency forecasting and preliminarily demonstrate effective results, but the interpretability is still controversial. Therefore, how to consider both the performance and the interpretability has already become a universal concern and an urgent unsolved problem. In this paper, we identified the above issue and proposed an interpretable data-driven approach, named EcoForecast 1 1 The code and data are available at https://github.com/navfour/ecoforecast. , for short-term macroeconomic forecasting based on the N-BEATS (neural basis expansion analysis for interpretable time series forecasting) neural network. To the best of our knowledge, EcoForecast is the first interpretable purely data-driven unified normative scheme for macroeconomic forecasting that achieves variable forecast frequencies and prediction domains, surpassing traditional statistics-based and deep learning-based approaches in performance or interpretability. EcoForecast used a three-level hierarchical signal encoding, including the fully connected neural network (FCNN) level, the block level, and the stack level. The FCNN level implemented both forecast and backcast information extraction for the temporal prediction and parameter learning of context. Block levels were connected by residuals so that the block's backcast could be sequentially filtered on the input. The stack level was used to form the top-level system of EcoForecast, where each stack was constrained to specialize in different inductive functions. To some extent, EcoForecast can balance effectiveness, efficiency, generalizability and interpretability while conditions such as forecast frequency and time window change, even when unexpected incidents occur. Based on the actual macroeconomic data for China from 1992 to 2022, the data-driven EcoForecast demonstrated high stability in different sequence learning scenarios and the accompanying high-accuracy performance. This stability was reflected in smaller prediction error expectation and variance, tolerance of fewer input samples, and robustness across prediction domains. The experimental results indicated that EcoForecast improved the accuracy up to 3.94 times compared with the traditional BVAR. In the robustness test, EcoForecast required only a quarter of the data to achieve 2.51 times smaller forecast errors than the BVAR while also improving the accuracy of varied macroeconomic indicators such as the Purchasing Managers' Index (PMI) and national electricity generation (ELEC)forecasting by 2.38 and 1.45 times. EcoForecast had a high sensitivity to the emergence of economic inflection points and adapts quickly when the economic environment changes, thus demonstrating a performance that exceeds traditional solutions in GDP forecasting during epidemics and PMI forecasting during economic turmoil. Interfacing with traditional economics research, the interpretable EcoForecast can uncover the trends and cycles of economic change, from which the conclusions are validated with the actual economics practice in China. Our findings can provide a new possible research direction for short-term macroeconomic forecasting. [ABSTRACT FROM AUTHOR]

Traditional construction progress tracking relies on labor-intensive activities with time lags, potential man-made errors, and inefficient progress management, which demands for an innovative and automated progress tracking approach. This paper describes a deep learning method that utilizes image segmentation to automatically evaluate the wall construction progress of an entire floor with the progress results streamlined to BIM. The approach was applied to a case study in China for assessing plastering construction activities with high segmentation accuracy (mean average precision = 96.8%). Further improvement of Mask Region-Based Convolutional Neural Networks (Mask R-CNN) and evaluation of its superiority over other models have also been discussed. This study provides both theoretical and practical references for unmanned supervision of progress tracking and intelligent schedule management. • Described an evaluation framework to recognize wall progress of an entire floor. • Proposed an image segmentation method to calculate wall construction area. • Improved Mask R-CNN algorithm to increase segmentation accuracy. • Discussed the influence of image light on segmentation accuracy. • Applied the framework to a case study with mean Average Precision of 96.8%. [ABSTRACT FROM AUTHOR]

Integrating return prediction of traditional time series models in portfolio formation can improve the performance of original portfolio optimization model. Since machine learning and deep learning models have shown overwhelming superiority than time series models, this paper combines return prediction in portfolio formation with two machine learning models, i.e., random forest (RF) and support vector regression (SVR), and three deep learning models, i.e., LSTM neural network, deep multilayer perceptron (DMLP) and convolutional neural network. To be specific, this paper first applies these prediction models for stock preselection before portfolio formation. Then, this paper incorporates their predictive results in advancing mean–variance (MV) and omega portfolio optimization models. In order to present the superiority of these models, portfolio models with autoregressive integrated moving average's return prediction are used as benchmarks. Evaluation is based on historical data of 9 years from 2007 to 2015 of component stocks of China securities 100 index. Experimental results show that MV and omega models with RF return prediction, i.e., RF+MVF and RF+OF, outperform the other models. Further, RF+MVF is superior to RF+OF. Due to the high turnover of these two models, this paper discusses their performance after deducting the transaction fee cased by turnover. Experiments present that RF+MVF still performs the best among MVF models and omega model with SVR prediction (SVR+OF) performs the best among OF models. Moreover, RF+MVF performs better than SVR+OF and high turnover erodes nearly half of their total returns especially for RF+OF and RF+MVF. Therefore, this paper recommends investors to build MVF with RF return prediction for daily trading investment. • Compares the performance of machine learning and deep learning in stock preselection. • Combining return prediction of machine learning and deep learning in portfolio formation. • Emphasis on advancing portfolio optimization with return prediction. • Advanced mean–variance model with random forest forecasts performs the best. [ABSTRACT FROM AUTHOR]

Total organic carbon (TOC) content is significant for "sweet spots" prediction and resource calculation. Traditional physical methods and machine learning methods, however, only use labeled samples (core TOC samples and well logging data) to predict TOC, resulting in a performance bottleneck. Meanwhile, many oilfields have considerable unlabeled data (well logging data). Therefore, this paper proposed a pretrain-and-finetune few-shot learning framework, called Contrastive Learning-Convolutional Neural Network (CL-CNN), to predict TOC using unlabeled data. CL has been a research hotspot in unsupervised learning in recent years, with the principle of reducing the distance between similar samples and increasing the distance between different samples in the feature space. CNN is distinguished by its strong generalization ability and high prediction accuracy, as well as its local connectivity and weight sharing. The CL-CNN model first uses CL to pretrain the initial parameters of the CNN model on a large number of unlabeled datasets and then finetunes the resultant CNN model using the labeled data. The framework combines the advantages of CL and CNN to further improve the prediction accuracy and generalization ability. The proposed method was applied to predict the TOC of shale gas in the Sichuan Basin, China, using 121 core samples and 1224 well logging data from the JY 10 well (121 logging data for supervised learning and 1103 logging data for unsupervised learning). Besides, JY 10–10 well data were used to validate the CL-CNN model. Experimental results demonstrated that the CL-CNN model outperformed many machine learning methods, including the support vector machine (SVM), Gaussian process regression (GPR), random forest (RF), and CNN models, indicating the feasibility and effectiveness of the proposed method. Because of this, the CL-CNN model can also be used as a general prediction tool for reservoir parameters such as porosity and permeability. • Contrastive learning was used to learn the general attributes of the unlabeled data. • Contrastive learning-convolutional neural network (CL-CNN) was proposed to predict TOC content. • Unsupervised pretraining and supervised finetuning were used for few-shot TOC prediction. • CL-CNN obtained promising results for predicting TOC and outperformed many methods. • The plane distribution of TOC in the study area was predicted by CL-CNN accurately. [ABSTRACT FROM AUTHOR]

• Lead conversion based on lead I was developed, which are easily measured by wearable devices. • More than one beat signal generation is made by proposed method. • To minimize the distortion of the original signal, raw signals are used. • Diagnosable abnormal signals are generated. Electrocardiogram (ECG) is measured in various ways. The three main ECG measurement methods include resting ECG, Holter monitoring, and treadmill method. In standard ECG measurement methods, multiple electrodes are attached to the limb and chest. Limb and chest leads measure the frontal and sagittal planes of the heart, respectively. In this case, ECG signals are measured briefly up to 10 seconds. To measure ECG signals based on a single lead, wearable devices have been developed that could measure long-term ECG signals daily. ECG signals are vectors in the heart, which is a three-dimensional structure. Therefore, a single-lead measurement lacks detailed information. The objective of this study was to synthesize multiple ECGs from a single-lead ECG using a generative adversarial network (GAN). We trained our model with two independent datasets and one combined dataset. For experiment 1, the PTB-XL dataset was used as the training set, and the China dataset was used as the test set. For experiment 2, the China dataset was used as the training set, and the PTB-XL was used as the test set. Optimized GAN models were obtained for each experiment and evaluated. The Fréchet distance (FD) score and mean squared error (MSE) were used for evaluation. The FD and MSE scores for experiments 1 and 2 were 7.237 and 0.024, and 8.055 and 0.011, respectively. We proposed a method to overcome the limitations of modern ECG measurement methods. Low FD and MSE scores not only indicate the possibility but also the similarity between synthesized ECG and reference ECG when compared in ECG paper format. This indicates that the proposed method can be applied to wearable devices that measure single-lead ECG. [ABSTRACT FROM AUTHOR]

• A deep learning incorporated data augmentation imputation framework for structural health monitoring system in multi-sensor failure is proposed, which only requires 1–2 days' incomplete dataset to impute missing data. • Generative adversarial network (GAN) is adopted to establish the correlation between the same type of sensors with less parameter adjustment. Combining GAN with the data augmentation outperforms conventional models in data imputation. • Long short-term memory (LSTM) network is used to establish the correlation between the strain and dynamic displacement sensors to impute the missing data of displacement sensors. Sensors, as an important part of structural health monitoring systems (SHMSs), will be abnormal sometimes due to their deterioration or environment effect, which will result in data loss during the health monitoring process of the structures. Data loss often happens in real monitoring applications, especially for wireless monitoring systems. Missing data, especially the long-term continuous missing data, will have a great impact on structural damage detection and condition evaluation. Usually, the long-term continuous missing data of the sensors are interpolated by traditional methods such as the correlation methods, which use a lot of normal monitoring data to build models and impute the missing data. However, in practice, many SHMSs in China have been in service for about 20 years or more, and many sensors installed have become faulty. It is usually difficult to obtain enough dataset fit for above methods. In this paper, a novel data imputation framework based on deep learning and data augmentation technique is therefore proposed, which enables the application of data modeling and missing data imputation based on the less remaining data when multiple sensors fail. Data imputation can be made between the same type of sensors (STSs) and also different types of sensors (DTSs). Generative adversarial network (GAN) based deep learning method and data augmentation technique are used for the imputation between the STSs; while long short-term memory (LSTM) network method is used for data imputation between the DTSs. The proposed methods are verified on the dataset of a real concrete bridge located in China, and results show that the proposed method achieves good performance. [ABSTRACT FROM AUTHOR]

For intelligent risk prediction and optimization in the shield tunnel excavation under uncertainty, this paper proposes a data-driven decision support framework based on the integration of deep neural network and gradient descent technique (DNN-GDO). More specifically, the DNN architecture for modeling non-linear relationships between influential factors and potential risk is superior in concurrently extracting and learning features from multi-variable inputs, while the GDO technique as a solution for the multi-objective optimization problem can be directly applied in the established DNN model to provide suggestions for reducing or preventing potential risks in the tunneling environment. To reveal the effectiveness of the developed approach, it is performed in a practical tunnel construction case in the Wuhan metro system, China. Sixteen decision variables belong to four categories of tunnel design, geology, the tunnel boring machine (TBM) operation, and surrounding building are taken into account, which are closely relevant to two significant risks named the accumulative ground settlement and building tilt rate. It has been found that the end-to-end learning pipeline based on DNN can yield high accuracy in the multi-output regression task, resulting in the mean-square error (MSE) of 0.737 and 0.002 for the two target risks, respectively. GDO is proven useful in not only finding the optimal solutions for effective risk control but also uncovering the key features for a better explanation of the optimization process. It has been proved that optimal solutions derived from GDO can flexibly guide to adjust input variables of interest to achieve the minimal risk degree. Besides, cutter rotating speed and ground volume of TBM along with the relative horizontal distance of buildings turn out to be the top 3 critical features contributing more to the tunnel-induced risk mitigation, which deserve more attention. In brief, the novel DNN-GDO decision tool supports to formulate meaningful strategies objectively and efficiently for tunneling safety assurance, which no longer largely relies on time-consuming and laborious manual assessment. • A data-driven approach based on deep neural network and gradient descent technique is proposed. • Gradient descent technique aims to provide suggestions for preventing potential risks in tunnels. • A practical tunnel construction case in China is applied for demonstration purposes. • Deep learning model can yield a high accuracy in the multi-output regression prediction task. • The derived optimal solutions can flexibly guide to adjust input variables to achieve minimal risk. [ABSTRACT FROM AUTHOR]

This paper explains the application of concept mapping to help foster a learning-centred approach. It investigates how concept maps are used to measure the change in learning following a two-week intensive undergraduate Marketing Principles course delivered to 162 Chinese students undertaking a Bachelor of Business Administration programme in China. Using four scoring procedures (breadth of knowledge, relational quality, structural quality and holistic quality), student learning is assessed in terms of prior and new knowledge, as well as improvement in knowledge structure. Concept map scores are correlated with conventional measures of overall student performance. The results provide strong evidence for improvement in students’ ability to externalise new learned concepts resulting from intensive instruction. Incorporating concept maps as a contemporary method of learning in the curriculum potentially enriches student learning, provides a measure of the impact of teaching on students’ learning and adds value to students’ higher education experience. [ABSTRACT FROM PUBLISHER]

• A novel neural network with two residual blocks is proposed to improve the feature extraction ability of high-dimensional and redundant spectral data. Due to the high dimension of tobacco spectral data, a deep neural network is used to extract the features of spectral data. Two residual blocks are added into the proposed neural network to do identity mapping, which can map the data in a shallow layer to a deep layer. The identity mapping can effectively solve the issue of gradient disappearance that can increase the feature quantity of a deep neural network. • The balance and suppression factors are added to the loss function of the proposed neural networks, which solve the issue in training unevenly distributed data. The spectral data size of tobacco collected from different tobacco cultivation regions varies. In the training process of a neural network, the balance and suppression factors add more weight to small-size data samples. Therefore, the neural network can have the same classification ability for each data sample. • A parametric rectified linear unit (PReLU) function is used to add linear factors to the negative input, which can adaptively learn parameters in the network. In the network, batch normalization (BN) operation is used to speed up the network training and improve the generalization ability of the network. In addition, the exponential decay learning rate is applied to the control of learning speed. As a deep learning optimization method, dropout is used to avoid overfitting, improve the generalization performance of network, and make the model training converge faster. Near-infrared (NIR) spectroscopy techniques have been widely used to classify tobacco cultivation regions. NIR spectroscopy of tobacco leaves involves a large number of correlated features, so it is difficult to find the connection between spectral data and tobacco cultivation regions. This paper proposes a novel classification model of tobacco cultivation regions that integrates residual network (ResNet) and NIR spectroscopy techniques. As the number of neural network layers increases, the network may have issues such as network degradation, gradient disappearance, and the reduction of sample recognition rate. The proposed model applies a residual module to a neural network which effectively solves or alleviates the vanishing gradient issues caused by the increase of network depth. This paper also adds balance and suppression factors to the loss function to solve the issues caused by uneven sizes of tobacco samples collected from different cultivation regions in the training process. In the proposed method, tobacco samples are marked as internal and external samples respectively during the training process. Internal samples are collected from the corresponding cultivation regions in the north, northeast, and northwest of Guizhou Province, China. External samples are collected from other cultivation regions. The weight distributions of internal and external samples can be adjusted by experimental results to improve the identification accuracy of the proposed solution. The size of training samples determines the generalization ability of the network and affects the experimental results. A parametric rectified linear unit (PReLU) function is integrated into the network, in which the parameters of a linear unit are adaptively learned to further improve the identification accuracy of the proposed solution. Compared with current mainstream methods, the experimental results confirm that the proposed model is superior in accurately identifying different cultivation regions of tobacco leaves. [ABSTRACT FROM AUTHOR]

Owing to the importance of coalbed methane (CBM) as a source of energy, it is necessary to predict its future production. However, the production process of CBM is the result of the interaction of many factors, making it difficult to perform accurate simulations through mathematical models. We must therefore rely on the historical data of CBM production to understand its inherent features and predict its future performance. The objective of this paper is to establish a deep learning prediction method for coalbed methane production without considering complex geological factors. In this paper, we propose a multivariate long short-term memory neural network (M-LSTM NN) model to predict CBM production. We tested the performance of this model using the production data of CBM wells in the Panhe Demonstration Area in the Qinshui Basin of China. The production of different CBM wells has similar characteristics in time. We can use the symmetric similarity of the data to transfer the model to the production forecasting of different CBM wells. Our results demonstrate that the M-LSTM NN model, utilizing the historical yield data of CBM as well as other auxiliary information such as casing pressures, water production levels, and bottom hole temperatures (including the highest and lowest temperatures), can predict CBM production successfully while obtaining a mean absolute percentage error (MAPE) of 0.91%. This is an improvement when compared with the traditional LSTM NN model, which has an MAPE of 1.14%. In addition to this, we conducted multi-step predictions at a daily and monthly scale and obtained similar results. It should be noted that with an increase in time lag, the prediction performance became less accurate. At the daily level, the MAPE value increased from 0.24% to 2.09% over 10 successive days. The predictions on the monthly scale also saw an increase in the MAPE value from 2.68% to 5.95% over three months. This tendency suggests that long-term forecasts are more difficult than short-term ones, and more historical data are required to produce more accurate results. [ABSTRACT FROM AUTHOR]

• Peach variety detection using VIS-NIR spectroscopy. • Detection and classification without handcrafted feature extraction. • Best detection accuracy was 100% in the validation and 94.4% in the test. More than 1000 peach varieties with significant differences in qualities are cultivated in China. Distinguishing peach varieties is not only needed by peach dealers, but also demanded by products processing enterprises and consumers. Detecting the varieties of fruit is one of the most important analyses. Over the past decades, there are many linear or nonlinear methods have been proposed, such as principal component analysis, partial least squares, support vector machine etc. However, these traditional methods commonly need some preprocessing steps including denoising, baseline correction, wavelength selection and so on. Hence, it requires the users to have enough skilled knowledge before they can establish a good performance detection model. To offer information on identifying peach varieties, visual near-infrared (VIS-NIR) diffuse reflectance spectra between 350 and 820 nm were collected for five peach varieties, 100 samples for each variety, the total samples consisted of 500 peaches. Spectral information has shown the rapid non-destructive measurement ability in many research fields. In recent years, the spectral information was researched for detecting the sugar and firmness of fruits, for examples, apple, peach, tomato and so on. Some traditional detection methods using spectra require application specific transforms, expertly designed constraints and model parameters, and have limited detection performance due to their maintenance costs. With the development of machine learning technology in recent years, deep learning which plays an important role in different research projects has won the eyes of fields from both academy and industry. In order to classify peach varieties by analyzing VIS-NIR spectra, a detection method based on deep learning principle is proposed in this paper. In this paper, the method can realize multi-identification peach varieties, by constructing a one dimensional convolution neural network and establishing a VIS-NIR spectra database containing five kinds of peaches. The network model is obtained through training, and then is employed to predict the testing set data. The accuracy of models based on deep learning reached 100% in the validation dataset and 94.4% in the test dataset. This study indicated that peach varieties could be distinguished successfully by using VIS-NIR spectroscopy and deep learning. [ABSTRACT FROM AUTHOR]

• A Convolutional Long Short Term Memory Network is used to predict the flood events based on deep learning techniques. • The spatial and time characteristics of floods in China are well modeled to overcome the shortcomings generated by merely relying on time-series analysis. • Different from traditional methods, the hydrological area is gridded into different watersheds for future processing using image processing methods. Floods cause substantial damage across the world every year. Accurate and timely prediction of floods can significantly minimize the loss of life and property. Recently, numerous machine learning models have been used for flood prediction, showing that their performance is preferable to traditional statistical models. However, the existing models neglect the spatial features of floods, which drive flood generation and concentration. In this paper, the area of interest is divided into grids based on longitude and latitude, and the rainfall and discharge collected by stations are combined into tensors according to station coordinates. Different from one-dimensional time series, our input feature is a two-dimensional time series with spatial information. Hence, combining a Convolutional Neural Network (CNN) with a Long Short Term Memory Network (LSTM), we propose the convolution LSTM (ConvLSTM) to extract spatiotemporal features of hydrological information. The methodology is demonstrated using the hydrological data collected at the Xi County stations, located on the Huai River in Henan Province, China. Numerical results indicate that the relative error of arrival time is within 30%, and the relative error of peak discharge is within 20%, satisfying the 2005 Chinese Water Resource Standard on flood prediction permit error. The experiments also show that the ConvLSTM outperforms the recent models in terms of flood arrival time and peak discharge, thereby proving a promising alternative. [ABSTRACT FROM AUTHOR]

• The industrial chain information is introduced to improve the stock forecast results. • DM test was carried out on the results, conclusion was statistically supported. • An interpreter based on LIME algorithm is constructed to explain the results. • The investment strategy is constructed, and the yield of maturity exceeds that of the buy-and-hold strategy. The prediction of stock price has always been a main challenge. The time series of stock price tends to exhibit very strong nonlinear characteristics. In recent years, with the rapid development of deep learning, the ability to automatically extract nonlinear features has significantly attracted scholars' attention. However, the majority of the relevant studies have concentrated on prediction of the changes of stock market based on the data of the specific stock (e.g., transaction data, financial data, etc.), while those studies ignored the interaction between stocks of different industries, especially the interaction between the stocks of upstream enterprises and downstream enterprises in the industrial chain. This paper aims to propose a combination of transfer learning of industrial chain information and deep learning models, including multilayer perceptron (MLP), recurrent neural network (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU), for stock market prediction. These models are used to predict the trend of the 379 stock market indices by industry in China, and the DM test was employed for validation of the prediction results. It can be concluded that RNNs are not necessarily such an optimal choice for the prediction when dealing with specific time series data, and it could be justified by using the local interpretable model-agnostic explanations (LIME) algorithm. Hence, the MLP was selected to effectively improve the accuracy of the prediction of the stock market indices based on the transfer learning of industrial chain information. The investment strategy is constructed according to the prediction results, and the yield of maturity exceeds that of the buy-and-hold strategy. [ABSTRACT FROM AUTHOR]

• Two micro energy grid optimization models combined with regional energy characteristics are proposed. • An on-line adaptive optimization model and strategy for multiple MEG access to DN is proposed. • An effective measure of using distributed clean energy in mountainous area is provided. • The problem of various forms of energy supply in Southwest China has been solved. The rapid growth of distributed photovoltaic and wind power installed capacity in Southwest China has reduced the operation safety margin of the power grid. Power grid dispatchers face the dilemma of ensuring security of the power grid and the output curtailment of renewable energy. To achieve the full consumption of renewable energy, it is an effective way to make use of the space-time complementary characteristics of different energies by forming micro energy grids. By connecting to the distribution network, the energy among the micro energy grids can be transferred and distributed in the form of electricity. Therefore, this study builds collaborative operation optimisation models of distribution networks with multi-micro energy grids and proposes an adaptive dynamic real-time optimisation algorithm based on pre-training and online deep learning techniques for the joint system. A test system was constructed based on the actual equipment parameters and data from the Guizhou power grid. Through the calculation and simulation in the test system, the overload of the distribution network equipment and the network loss were effectively reduced while ensuring the full consumption of wind and photovoltaic power. In addition, the accuracy and effectiveness of the model and algorithm were verified. The results obtained in the case studies suggest that the method proposed in this paper can be applied to southwest China and similar regions. [ABSTRACT FROM AUTHOR]

The occurrence of rockburst can cause significant disasters in underground rock engineering. It is crucial to predict and prevent rockburst in deep tunnels and mines. In this paper, the deficiencies of ensemble learning algorithms in rockburst prediction were investigated. Aiming at these shortages, a novel machine learning model, deep forest, was proposed to predict rockburst risk. The deep forest combines the characteristics of deep learning and ensemble models, which can solve complex problems. To develop the deep forest model for rockburst prediction, 329 real rockburst cases were collected to build a comprehensive database for intelligent analysis. Bayesian optimization was proposed to tune the hyperparameters of the deep forest. As a result, the deep forest model achieved 100% training accuracy and 92.4% testing accuracy, and it has more outstanding capability to forecast rockburst disasters compared to other widely used models (i.e., random forest, boosting tree models, neural network, support vector machine, etc.). The results of sensitivity analysis revealed the impact of variables on rockburst levels and the applicability of deep forest with a few input parameters. Eventually, real cases of rockburst in two gold mines, China, were used for validation purposes while the needed data sets were prepared by field observations and laboratory tests. The promoting results of the developed model during the validation phase confirm that it can be used with a high level of accuracy by practicing engineers for predicting rockburst occurrences. [ABSTRACT FROM AUTHOR]

Underwater glider (UG) is one of the most promising low-power and long-voyage autonomous ocean observation platforms. During the gliding, the roll regulation unit (RRU) frequently works to regulate the disturbed heading due to the influence of time-varying current, biofouling, and cumulative error of RRU. In this paper, an accurate energy consumption model of RRU (E RRU) is established to quantify the impact of RRU on the overall energy consumption, and the trial data collected by the Petrel-L glider, China, in the South China Sea are used to verify the accuracy of E RRU. To minimize the working frequency of RRU, a new roll center compensation method (RCCM) is proposed based on variational mode decomposition and long short-term memory (VMD-LSTM). This study analyzes six classical prediction methods based on deep learning, and the VMD-LSTM at k = 5 method is the optimal one to predict the deviation of roll center for UGs. The results indicate that the proposed RCCM can obtain accurate and reliable compensation value for the roll center and effectively restrain the working frequency of RRU. Compared with the data before compensation, the E RRU of the Petrel-L glider is reduced by approximately 22%. Furthermore, the RCCM can apply to other similar UGs. • A mathematical energy consumption model of roll regulation unit (RRU) for the glider is established. • The VMD-LSTM (k = 5) method is the optimal one to predict the deviation of roll center for the glider. • A roll center compensation method (RCCM) is proposed to minimize the working frequency of RRU. • The energy consumption of RRU with RCCM is reduced by approximately 22% in sea trial. [ABSTRACT FROM AUTHOR]

• Dictionary learning is used to model gold prospectivity for gold exploration targeting. • The case study shows that the dictionary learning model is no inferior to the logistic regression model and superior to the one-class SVM model. Machine learning and deep learning anomaly detectors have been successfully used to map mineral prospectivity. However, the establishment of machine learning or deep learning anomaly detection models for mineral prospectivity mapping often requires the determination of a set of initialization parameters in the absence of ground truth data. Improper initialization parameters will degrade the performance of these mineral prospectivity mapping models. Different from most machine learning and deep learning algorithms, dictionary learning is a "white-box" algorithm involving only linear transformations, and building a dictionary learning model requires only an empirical definition of the total number of atoms. Therefore, in this paper, two dictionary learning anomaly detectors were established for mineral prospectivity modeling based on the least angle regression-Lasso (LARS-Lasso) algorithm and the iterative shrinkage-thresholding algorithm (ISTA). The following five-step procedure was proposed for mineral prospectivity modeling using the dictionary learning techniques: (a) an overcomplete dictionary is constructed based on the input data; (b) each data point is transformed into sparse coefficients according to the overcomplete dictionary; (c) the sparse representation of each data point is calculated based on the overcomplete dictionary and the sparse coefficients; and (d) the Euclidean norm of the difference between each data point and its sparse representation is calculated and used as the mineral potential of the data point. The dictionary learning models were established to model gold prospectivity in the Jinchanggouliang district, Inner Mongolia, China, and compared with the logistic regression model and one-class support vector machine model in gold exploration targeting. The result shows that (a) the performances of the dictionary learning models are no less than that of the logistic regression (LGR) model and better than that of the one-class support vector machine (OCSVM) model, and (b) the gold prospective areas differentiated by the established models are strongly consistent with geological and metallogenic characteristics in the study area. Therefore, the dictionary learning algorithms are high-performance mineral prospectivity modeling techniques. It is worth to further test the effectiveness of the dictionary learning algorithms for different types of mineral deposits in different areas in mineral exploration targeting. [ABSTRACT FROM AUTHOR]

• The residual structure is introduced into the Inception structure as a convolutional block, and a plant disease classification network is established. • An improved Convolutional Block Attention Module combing channel attention and spatial attention is embedded in the classification network. • Corn, potato and tomato are identified as a whole and individually by using a classification network. • Designed and developed a web-based disease detection system, which can complete real-time identification of plant diseases on mobile phones and computers. In this paper, we proposed a convolutional neural network based on Inception and residual structure with an embedded modified convolutional block attention module (CBAM), aiming to improve the classification of plant leaf diseases. Corn, potatoes and tomatoes are the most cultivated grains in southern China. The leaves of the three crops are very fragile and sensitive and are susceptible to leaf diseases, such as leaf blight of corn, late blight of potato and mosaic virus of tomato. These diseases cannot be identified at early stages. Therefore, an efficient solution is proposed by deep learning techniques to detect the disease categories of crops, which can effectively prevent the spread of diseases and ensure the normal growth of plants. In this experiment, our model achieved an overall accuracy of 99.55% for the identification of the three diseases of corn, potato and tomato. In addition, we tested the three plants individually. The classification accuracy of our model on corn, potato and tomato was 98.44%, 99.43% and 95.20%, respectively. We have also developed a web-based real-time plant disease classification system and deployed our model. The system had good performance in time and accuracy evaluation metrics. The results of this study showed that our model had fewer parameters, shorter training time, and higher recognition accuracy compared to existing image classification models. [ABSTRACT FROM AUTHOR]

Seismic landslides are the most common and highly destructive earthquake-triggered geological hazards. They are large in scale and occur simultaneously in many places. Therefore, obtaining landslide information quickly after an earthquake is the key to disaster mitigation and relief. The survey results show that most of the landslide-information extraction methods involve too much manual participation, resulting in a low degree of automation and the inability to provide effective information for earthquake rescue in time. In order to solve the abovementioned problems and improve the efficiency of landslide identification, this paper proposes an automatic landslide identification method named improved U-Net model. The intelligent extraction of post-earthquake landslide information is realized through the automatic extraction of hierarchical features. The main innovations of this paper include the following: (1) On the basis of the three RGB bands, three new bands, DSM, slope, and aspect, with spatial information are added, and the number of feature parameters of the training samples is increased. (2) The U-Net model structure is rebuilt by adding residual learning units during the up-sampling and down-sampling processes, to solve the problem that the traditional U-Net model cannot fully extract the characteristics of the six-channel landslide for its shallow structure. At the end of the paper, the new method is used in Jiuzhaigou County, Sichuan Province, China. The results show that the accuracy of the new method is 91.3%, which is 13.8% higher than the traditional U-Net model. It is proved that the new method is effective and feasible for the automatic extraction of post-earthquake landslides. [ABSTRACT FROM AUTHOR]

• Multi-feature & stacking were used for automatic grading cataract into six levels based on the fundus images. • Two support vector machine (SVM) classifiers are used as base-learners and fully connected neural network (FCNN) are used as meta-learner to output the final classification result. • Multi-feature & Stacking achieve higher grading performance and lower volatility than grading using high-level features and texture features. • Automatic cataract grading can be used to help government assist people more accurately in poor areas. The shortage of ophthalmologists in rural areas in China causes a lot of cataract patients not getting timely diagnosis and effective treatment. We develop an algorithm and platform to automatically diagnose and grade cataract based on fundus images of patients. This method can help government assisting poor population more accurately. The novel six-level cataract grading method proposed in this paper focuses on the multi-feature fusion based on stacking. We extract two kinds of features which can effectively distinguish different levels of cataract. One is high-level features extracted from residual network (ResNet18). The other is texture features extarcted by gray level co-occurrence matrix (GLCM). Then a frame is proposed to automatically grade cataract by the extracted features. In the frame, two support vector machine (SVM) classifiers are used as base-learners to obtain the probability outputs of each fundus image, and fully connected neural network (FCNN) are used as meta-learner to output the final classification result, which consists of two fully-connected layers. The accuracy of six-level grading achieved by the proposed method is up to 92.66% on average, the highest of which reaches 93.33%. The proposed method achieves 94.75% accuracy on four-level grading for cataract, which is at least 1.75% higher than those of the exiting methods. Six-category cataract classification algorithm show that Multi-feature & Stacking proposed in this paper helps achieve higher grading performance and lower volatility than grading using high-level features and texture features respectively. We also apply our algorithm into four-level cataract grading system and it shows higher accuracy compared with previous reports. [ABSTRACT FROM AUTHOR]

Power dispatching systems currently receive massive, complicated, and irregular monitoring alarms during their operation, which prevents the controllers from making accurate judgments on the alarm events that occur within a short period of time. In view of the current situation with the low efficiency of monitoring alarm information, this paper proposes a method based on natural language processing (NLP) and a hybrid model that combines long short-term memory (LSTM) and convolutional neural network (CNN) for the identification of grid monitoring alarm events. Firstly, the characteristics of the alarm information text were analyzed and induced and then preprocessed. Then, the monitoring alarm information was vectorized based on the Word2vec model. Finally, a monitoring alarm event identification model based on a combination of LSTM and CNN was established for the characteristics of the alarm information. The feasibility and effectiveness of the method in this paper were verified by comparison with multiple identification models. [ABSTRACT FROM AUTHOR]

The identification of overpass structures in road networks has great significance for multi-scale modeling of roads, congestion analysis, and vehicle navigation. The traditional vector-based methods identify overpasses by the methodologies coming from computational geometry and graph theory, and they overly rely on the artificially designed features and have poor adaptability to complex scenes. This paper presents a novel method of identifying overpasses based on a target detection model (Faster-RCNN). This method utilizes raster representation of vector data and convolutional neural networks (CNNs) to learn task adaptive features from raster data, then identifies the location of an overpass by a Region Proposal network (RPN). The contribution of this paper is: (1) An overpass labelling geodatabase (OLGDB) for the OpenStreetMap (OSM) road network data of six typical cities in China is established; (2) Three different CNNs (ZF-net, VGG-16, Inception-ResNet V2) are integrated into Faster-RCNN and evaluated by accuracy performance; (3) The optimal combination of learning rate and batchsize is determined by fine-tuning; and (4) Five geometric metrics (perimeter, area, squareness, circularity, and W/L) are synthetized into image bands to enhance the training data, and their contribution to the overpass identification task is determined. The experimental results have shown that the proposed method has good accuracy performance (around 90%), and could be improved with the expansion of OLGDB and switching to more sophisticated target detection models. The deep learning target detection model has great application potential in large-scale road network pattern recognition, it can task-adaptively learn road structure features and easily extend to other road network patterns. [ABSTRACT FROM AUTHOR]

In this paper, we consider water surface object detection in natural scenes. Generally, background subtraction and image segmentation are the classical object detection methods. The former is highly susceptible to variable scenes, so its accuracy will be greatly reduced when detecting water surface objects due to the changing of the sunlight and waves. The latter is more sensitive to the selection of object features, which will lead to poor generalization as a result, so it cannot be applied widely. Consequently, methods based on deep learning have recently been proposed. The River Chief System has been implemented in China recently, and one of the important requirements is to detect and deal with the water surface floats in a timely fashion. In response to this case, we propose a real-time water surface object detection method in this paper which is based on the Faster R-CNN. The proposed network model includes two modules and integrates low-level features with high-level features to improve detection accuracy. Moreover, we propose to set the different scales and aspect ratios of anchors by analyzing the distribution of object scales in our dataset, so our method has good robustness and high detection accuracy for multi-scale objects in complex natural scenes. We utilized the proposed method to detect the floats on the water surface via a three-day video surveillance stream of the North Canal in Beijing, and validated its performance. The experiments show that the mean average precision (MAP) of the proposed method was 83.7%, and the detection speed was 13 frames per second. Therefore, our method can be applied in complex natural scenes and mostly meets the requirements of accuracy and speed of water surface object detection online. [ABSTRACT FROM AUTHOR]

• Economic development may drive academic construction safety research. • There has been an increase in construction safety research conducted by developing countries in recent years, probably due to an improvement in their economic performance. • Authors' keywords evidenced the popularity of research on safety management and climate. • There is a new interdisciplinary research trend where construction safety combines with digitalization technologies, with the largest number involving deep learning. The construction industry contributes significantly to economic development worldwide, yet it is one of the most hazardous industries where numerous accidents and fatalities happen every year. Little research to date has shed light on the impact of economic development on construction safety research. In this paper, we conduct an analysis of construction safety articles published in the 21st century via a bibliometrics approach. We have analysed: (1) construction safety in developed and developing countries; (2) the major organisations that have conducted construction safety research; (3) authors and territories of the research and (4) topics in construction safety and future research directions. The largest number of published construction safety documents were published by scholars from the US and China; the total number of published articles by these two countries was 1,125, at 56% of the 2000 articles that were published. Both countries showed high levels of research collaboration. While our results suggest that economic development may drive academic construction safety research, there has been an increase in construction safety research conducted by developing countries in recent years, probably due to an improvement in their economic development. While authors' keywords evidenced the popularity of research on safety management and climate, the network analysis on all keywords, i.e. keywords given by Web of Science and authors, suggest that construction safety research focused on three areas: construction safety management, the relationship between people and construction safety, and the protection and health of workers' impact on construction safety. We found that there is a new interdisciplinary research trend where construction safety combines with digital technologies, with the largest number involving deep learning. Other trends focus on machine learning, Building Information Modelling, and visualisation. [ABSTRACT FROM AUTHOR]

Cloud recognition is a basic task in ground meteorological observation. It is of great significance to accurately identify cloud types from long-time-series satellite cloud images for improving the reliability and accuracy of weather forecasting. However, different from ground-based cloud images with a small observation range and easy operation, satellite cloud images have a wider cloud coverage area and contain more surface features. Hence, it is difficult to effectively extract the structural shape, area size, contour shape, hue, shadow and texture of clouds through traditional deep learning methods. In order to analyze the regional cloud type characteristics effectively, we construct a China region meteorological satellite cloud image dataset named CRMSCD, which consists of nine cloud types and the clear sky (cloudless). In this paper, we propose a novel neural network model, UATNet, which can realize the pixel-level classification of meteorological satellite cloud images. Our model efficiently integrates the spatial and multi-channel information of clouds. Specifically, several transformer blocks with modified self-attention computation (swin transformer blocks) and patch merging operations are used to build a hierarchical transformer, and spatial displacement is introduced to construct long-distance cross-window connections. In addition, we introduce a Channel Cross fusion with Transformer (CCT) to guide the multi-scale channel fusion, and design an Attention-based Squeeze and Excitation (ASE) to effectively connect the fused multi-scale channel information to the decoder features. The experimental results demonstrate that the proposed model achieved 82.33% PA, 67.79% MPA, 54.51% MIoU and 70.96% FWIoU on CRMSCD. Compared with the existing models, our method produces more precise segmentation performance, which demonstrates its superiority on meteorological satellite cloud recognition tasks. [ABSTRACT FROM AUTHOR]

• European option datasets were time-sequenced by maturity to extract time information. • Two deep learning models were built for the option pricing tasks. • Our models can price option accurately without volatility calibration and simulation. • An interpretable method was used to find how models improved the pricing results. • Accumulate local effect range method was proposed to get feature emphasis of models. In this paper, we investigated the feasibility of pricing European options with time-sequencing data processing method and deep learning models, based on two European options, the ETF50 options of China and the S&P 500 options of America. Four competing models were built to verify the improvement of the 1D-CNN and LSTM models on the option pricing task. Methods like cross-validations and statistical tests were also used to make our experiments more robust. Besides, in order to increase the stability and the interpretability of our pricing models, we selected the ALE method to interpret and analyze the behavior of the deep learning models. The empirical results indicate that, in both ETF50 option and S&P500 option pricing tasks, the 1D-CNN and LSTM models had significant advantages in forecasting accuracy and robustness under moneyness, trading date or maturity dimension irrespectively. Especially for the LSTM model, which has robust performance using different kinds of cross-validation methods. With the help of ALE method, we proved that the improved performance brought by the 1D-CNN and LSTM models could be attributed to their capability of capturing time-series information and their different emphasis on input features and lags. [ABSTRACT FROM AUTHOR]

• A linear components extraction method is proposed to identify the daily fluctuating patterns of the solar power series. • A scenarios generation algorithm is proposed to provide forecasting linear trend data. • Temporal distributed deep learning models are developed based on the detected daily fluctuating patterns of the PV power generation. • The proposed algorithm can provide accurate day-ahead solar power forecasts for distributed PV farms without local weather information. In recent years, the integration of distributed photovoltaic has witnessed explosive growth because of its great economic and environmental benefits. However, the intermittence and randomicity of PV power generation are challenging the stable and economic operation of active distribution system, especially when the local historical and forecasting meteorological data are not available. Thus, this paper presents a novel Gated Recurrent Unit based hybrid model for forecasting distributed PV power generations. The proposed method uses locally historical PV power generation data to forecast one-day-ahead solar power time series with five-minute intervals. First, a linear components extraction method is proposed to identify the daily fluctuating patterns of the solar power series. Then, temporal distributed Gated Recurrent Unit models are developed based on the detected daily fluctuating patterns of the PV power generation. A scenarios generation algorithm is proposed to provide forecasting linear trend data for each Gated Recurrent Unit model. Finally, the daily linear series scenarios, the linear and nonlinear parts of the historical data are fed into the Gated Recurrent Unit to forecast the PV generation. Data collected from an actual distributed PV farm in southeastern China are utilized to validate the effectiveness of the proposed model. The results show that the proposed model can provide a nominalized root mean square error of 6.83% and a nominalized mean absolute error of 4.12%. Compared to a single Gated Recurrent Unit forecasting model, the developed model has improved the forecasting accuracy by 62.7%. [ABSTRACT FROM AUTHOR]

Weather radar data can capture large-scale bird migration information, helping solve a series of migratory ecological problems. However, extracting and identifying bird information from weather radar data remains one of the challenges of radar aeroecology. In recent years, deep learning was applied to the field of radar data processing and proved to be an effective strategy. This paper describes a deep learning method for extracting biological target echoes from weather radar images. This model uses a two-stream CNN (Atrous-Gated CNN) architecture to generate fine-scale predictions by combining the key modules such as squeeze-and-excitation (SE), and atrous spatial pyramid pooling (ASPP). The SE block can enhance the attention on the feature map, while ASPP block can expand the receptive field, helping the network understand the global shape information. The experiments show that in the typical historical data of China next generation weather radar (CINRAD), the precision of the network in identifying biological targets reaches up to 99.6%. Our network can cope with complex weather conditions, realizing long-term and automated monitoring of weather radar data to extract biological target information and provide feasible technical support for bird migration research. [ABSTRACT FROM AUTHOR]

• A satellite image-based photovoltaic power forecasting model is proposed. • A nonlinear cloud movement forecasting method based on deep learning is proposed. • An algorithm for intra-hour cloud condition estimation is established. • An algorithm for recognizing the cloud region that blocks sunlight is designed. The rapid variation of clouds is the main factor that causes the fluctuation of photovoltaic power. 1 1 This work was supported by the National Key Research and Development Program of China under Grant 2019YFE0118400. The satellite images contain plenty of information about clouds, applicable for photovoltaic power forecast. However, in practice, two main factors obstruct the application of the satellite images: 1) the relatively low update frequency of the satellite images mismatches the photovoltaic power forecasting frequency, and 2) the cloud region that blocks the sunlight changes significantly with time. In this paper, a novel satellite image-based approach for photovoltaic power forecast is proposed to overcome these obstacles and achieve accurate forecasting results. Firstly, concerning the hourly updated satellite images, a nonlinear cloud movement forecasting model, considering the thickness and shape changes of the cloud, is presented to forecast the hourly variation of the images. Secondly, an active cloud region selection rule is derived based on the changing solar position to dynamically select the cloud region that blocks the concerned photovoltaic power station in a satellite image. Thirdly, a sequential cloud region selection algorithm is provided to estimate the intra-hour variation of the cloud to match the photovoltaic power forecasting frequency. Finally, the photovoltaic power is predicted using the XGBoost algorithm concerning the effects of the cloud and other influencing factors. Testing results show that the proposed method can achieve more accurate photovoltaic power forecasts using the low update frequency satellite images. Meanwhile, the superior performance compared with other benchmarks also verifies the effectiveness of considering cloud information obtained by the proposed method for photovoltaic power forecast. [ABSTRACT FROM AUTHOR]

Significant wave height (SWH) is of great importance in industries such as ocean engineering, marine resource development, shipping and transportation. Haiyang-2C (HY-2C), the second operational satellite in China's ocean dynamics exploration series, can provide all-weather, all-day, global observations of wave height, wind, and temperature. An altimeter can only measure the nadir wave height and other information, and a scatterometer can obtain the wind field with a wide swath. In this paper, a deep learning approach is applied to produce wide swath SWH data through the wind field using a scatterometer and the nadir wave height taken from an altimeter. Two test sets, 1-month data at 6 min intervals and 1-day data with an interval of 10 s, are fed into the trained model. Experiments indicate that the extending nadir SWH yields using a real-time wide swath grid product along a track, which can support oceanographic study, is superior for taking the swell characteristics of ERA5 into account as the input of the wide swath SWH model. In conclusion, the results demonstrate the effectiveness and feasibility of the wide swath SWH model. [ABSTRACT FROM AUTHOR]