Search

Keypoint matching plays a vital role in the realm of synthetic aperture radar (SAR) image processing, serving as a crucial component within this domain. Another key fact to remember keypoint matching is a crucial step in change detection and image stitching. In the context of SAR images, a scale-invariant feature transformation (SIFT)-based approach, known as SAR–SIFT, presents a notable advantage by mitigating the impact of speckle noise; however, it cannot yield accurate edge information, and the resulting keypoints are nonuniformly distributed. We propose an edge enhancement and homogeneous spatial key point improved SAR–SIFT framework based on an unbiased difference-ratio (UDR) edge detector (called EEKHI–SAR–SIFT) to solve the above problems. The algorithm relies on the characteristics of edge unbiased localization and constant false alarm rate of UDR edge extraction to reduce the extracted wrong corner information as well as enhance the extraction exactness of keypoints. In addition, adaptive nonmaximum suppression (ANMS) method is applied to homogenize the dense keypoints with a large initial number that are gained by means of the EEKHI–SAR–SIFT algorithm and reduce their local clustering. Finally, a descriptor construction strategy that retains multiscale information is adopted to improve the descriptor uniqueness. Tests using multiple sets of SAR image data from different satellites (Gaofen-3, RADARSAT, and Sentinel-1A) demonstrate that the efficacy of the proposed EEKHI–SAR–SIFT algorithm reduces the root mean square error is about 1–2 pixel lower than the final result of the original SAR–SIFT algorithm.

Multiview satellite images play a crucial role in reconstructing Earth's surface information. Its application in deep-learning-based multiview stereo (MVS) methods has recently exhibited impressive performance. However, there is still ample room for improvement, particularly in the reconstruction of the digital surface model (DSM) within scenes characterized by dense features, high repetition, and significant elevation variations. To address the challenges, we present a novel network, Appli-SatNet. The network initially introduces a multiscale transformer module to emphasize feature correlations within and across multiple images. Subsequently, a novel gated recurrent units iterative optimizer is introduced to derive iterative indices for the cost volume within implicit states. This iterative process enhances the correlation among the cost volumes and enables multiple iterative refinements for prediction elevation maps. Ultimately, this leads to the generation of a high-resolution DSM that matches the resolution of the nadir image. The method is validated on a publicly available TLC SatMVS dataset, which includes plain areas with repetitive textures and inconspicuous features, mountainous regions with significant continuous elevation changes, and urban areas with dramatic elevation variations and high repetition. Comparative to mainstream methods, our approach demonstrated remarkable improvements in reconstruction quality. Moreover, when applied to low-resolution images captured by the Ziyuan-3 01 satellite over various complex terrains, it consistently exhibits robust reconstruction accuracy and superior visual quality, underscoring the applicability of our model.

Qinghai Lake, the largest inland saltwater lake in China, is an important water body that maintains the ecological security of the northeastern Tibetan Plateau. Global navigation satellite system reflectometry technology (GNSS-R) is rarely used in lake detection, especially in lake boundary detection under the circumstances of spaceborne. This article attempts to implement lake boundary detection using cyclone global navigation satellite system (CYGNSS) raw intermediate frequency (IF) data for the first time. First, the second-order frequency-locked loop is used to assist the third-order phase-locked loop closed-loop tracking to successfully generate higher delay resolution and Doppler resolution (67 × 51) delay Doppler maps (DDM) than CYGNSS compressed DDM (11 × 17); second, based on the principle of scattering characteristics of different reflectors of GNSS-R, and the fusion method of signal power ratio and effective area delay distance (DLR), the 2018 Qinghai Lake boundary detection inversion was carried out in the lake area to explore the feasibility of achieving higher spatial resolution (0.7 km) and high-precision lake boundary detection with the raw IF data of CYGNSS. The inversion detection results are compared with the topographic distribution results of Google Earth and global land analysis and discovery, and the inversion result error is about 0.5 km, which is consistent with the theoretical spatial resolution. It proves that spaceborne GNSS-R technology and CYGNSS raw IF data have great potential in the field of lake boundary detection.

When multiple synthetic aperture radar images are stitched together, the intensity disconnects between them can have a significant impact on the mosaic's quality. Many approaches focus on decreasing the intensity differences between images while ignoring the issue of image quality improvement. This study provides an algorithm for color correction and naturalness restoration for multiple images with uneven luminance in order to generate high-quality mosaics. To increase the illuminance component's naturalness, the image is first divided into illuminance and reflectance components, and the illuminance component is subjected to adaptive luminance improvement and contrast enhancement. Using a color consistency optimization approach, the intensity disparities between illuminance components are subsequently minimized. The reflectance and enhanced illuminance components are then combined to produce an improved image. After that, the enhanced image is mosaicked using multiband blending. Finally, the intensity differences between the enhanced images are further decreased using the block-based Wallis transform based on the mosaic. We assessed the proposed method on 402 Sentinel-1 images covering the majority of China's land area to verify its robustness. When compared to similar algorithms, our strategy reduces the color distance by about 36.72%, improves the average gradient by around 89.44%, and increases the patch-based contrast quality index by roughly 32.85%. The experimental outcomes reveal that our approach has considerable advantages in terms of color correction and image quality improvement, both visually and quantitatively.

High-resolution stereo satellite images (HRSSIs) have the potential to provide the accurate height and volume information, playing a crucial role in assessing building collapses during various natural disasters. However, the time-consuming process of three-dimensional (3-D) reconstruction, inadequate vertical accuracy of digital surface model (DSM), and concentrated clustering of buildings pose challenges for collapse assessment focused on buildings. Therefore, we present an improved approach for rapid fine-grained assessment of building collapses. First, the accurate and consistent positioning parameters for HRSSIs are obtained through the combined block adjustment using laser altimetry points, ensuring the generation of DSMs with vertical accuracy exceeding 2 m. Next, a set of rapid 3-D reconstruction techniques is introduced, achieving a significant eightfold improvement in generating DSMs. Subsequently, we deploy an automated workflow for batch processing and registration of open-source building footprints, enabling the accurate extraction of building height changes from dual-time DSMs. Finally, based on the building change image, a large-scale GIS image of building floor-level collapses is generated using connected component detection and threshold classification strategies. These findings have far-reaching implications for postdisaster emergency response, damage assessment, and expeditious reconstruction efforts. In our study, we processed an 800 km2 area in Kahramanmaras Province, Turkey, generating dual-time DSMs within 1 h. This enabled the assessment of floor-level collapses for a total of 48 092 buildings within the area. Validation was conducted on 361 houses in the city center, utilizing Google Street view images as ground truth. Remarkably, our approach achieved a high accuracy rate of 93.27% in floor-level assessment.

Detailed and timely monitoring of the location and intensity of the fire is critical to reducing the destructive impacts of a fire. Satellite imagery platforms, in particular geostationary satellites with high temporal resolution, allow for real-time fire monitoring. However, because of the coarse resolution of geostationary satellite images, even when deep learning models are applied, precision still remains limited. Thus, the prediction models easily fall into a local optimal solution because of the insufficient semantic information from low spatial resolution. Therefore, in this study, we proposed a novel deep learning model, SBT-FireNet, to monitor fire spots from Himawari-8 satellite images. Specifically, the extraction modules of spatial, band, and time-series features were designed and integrated into the model. The spatial feature extraction module served to collate information about fires and their surrounding environment, while the band and time-series features were designed to obtain fire-sensitive band and time information, respectively. The newly structured SBT-FireNet model was tested in four fire-prone areas with high forest cover. The precision of SBT-FireNet in four test areas is 35.2% higher than other methods. The model yielded significant improvements through the combination of the modules of spatial, band, and time-series features and their fire-tailored design. The advantages of the high temporal resolution of geostationary satellite images were fully integrated into the model to ensure that the model monitors the possibility of fire in an automated way in a near-real-time manner.

Accurate surface water mapping is crucial for monitoring and protecting ecosystem environments. During the past decades, several water extraction methods have been presented and significant progress has been made. However, most of the existing research has mainly focused on reducing the interference factors, such as clouds and cloud shadows, building shadows, and mountain shadows. There are relatively few studies concentrated on the fine extraction of water body boundaries, which is of equal importance for surface water mapping. Therefore, in this article, we developed a novel boundary enhancement network, to improve its ability to extract water boundary information. The proposed BE-Net consists of three modules, i.e., SE, SA, and MFF, which are used to focus on water features, water boundary information, and fuse features of different levels, respectively. In addition, the Sobel edge loss function is adopted. The proposed model was tested on six regions with significant differences in water body morphology and the contribution of the Sobel edge loss function and the three modules were investigated. The results demonstrated that: 1) compared with the state-of-the-art methods, the proposed BE-Net achieved the best accuracy in all testing areas and was all above 97%; 2) Compared with the commonly used loss function, the Sobel edge loss function can better capture the detail boundary information, resulting in higher accuracy; and 3) by integrating the SE, SA, and the MFF modules, a robust boundary enhancement network is constructed, and the water extraction accuracy can be significantly improved.

While various methods have been developed to estimate the age of impact craters, such as the crater size frequency distribution and morphology methods. Accurately and efficiently estimating the ages of lunar craters using traditional techniques is challenging due to their complex morphology and large number. As a result, the accuracy of age estimation algorithms for meteorite craters based on deep learning is restricted by factors such as a scarcity of age-labeled data and the complex morphology of these craters. To address these issues, this article presents an enhanced double-branch network for estimating crater ages via semisupervised learning and multisource lunar data. The algorithm consists of three steps: semisupervised training data augmentation, adaptive two-branch feature extraction, and a two-stage crater age classification process. The effectiveness of the improved approach was validated through ablation experiments, resulting in an overall accuracy of 83.7% on the test set of meteorite craters. This is 5.2% higher than the accuracy achieved by the previous deep learning method.

Orthorectification reflects a large amount of real and objective information, such as the characteristics of images and the geometric accuracy of maps. Conducting a large batch of orthorectification is a process with high time cost owing to the pixelwise correction each image. A common approach is to use graphics processing unit (GPU) parallel computing to accelerate orthorectification processing. However, most of the existing GPU acceleration studies have adopted experimental testing methods to determine thread parameters, which are inapplicable to different GPUs and affect the GPU acceleration performance. We put forward an adaptive calculation method for GPU thread parameters based on the performance parameters of different GPUs and by simultaneously blocking the image automatically according to the GPU memory space. We used 112 ZY-3 images to test the adaptive GPU and compare it to a general GPU. The experimental results show the following: first, for a single ZY-3 image, the GPU acceleration by the adaptive calculation method presented in this article is 43.22% higher than that by the general GPU, and the correction time is 34.41 times faster than that of the central processing unit. The result of the automatic image blocking was the same as that of the artificial blocking. Second, the experimental performance on four different GPUs indicated that all GPUs exhibited a significant speed boost. Third, for large-batch images, the GPU acceleration by the adaptive GPU was 32.6% higher than that by the general GPU, which provides an adaptive optimization strategy for large-batch image orthorectification.

The output speed of a high-resolution camera with a fast frame rate is high, reaching 800 MB/s or even more. Consequently, the abundant data recorded can easily cause data overflow during transmission and frame loss during storage. To address this problem of data overflow and frame loss while achieving reliable acquisition rates, herein, we design a high-speed video adaptive transmission and storage method based on the camera link transmission protocol. The system uses the camera link transmission protocol and high-speed cameras for data acquisition. During data transmission, memory space is employed as the data buffer pool, an adaptive memory buffer pool is designed using the hardware and acquisition parameters, and a circular read/write buffer, combined with a block device read/write method, is used inside the buffer pool for data transmission. For data storage, input/output completion port asynchronous storage mixed with frame sequence number filtering is employed to write data, and the adaptive buffer of the hard disk is used to achieve efficient data writing. The experimental results reveal that the proposed method can realize efficient and stable writing of data on conventional industrial cameras based on the camera link interface, effectively addressing the problem of frame loss during data storage. Moreover, the stored data can be transferred efficiently with less resource occupation, thereby meeting the application requirements of high-precision vision measurement systems and demonstrating broad application value.

Cyclone Global Navigation Satellite System (CYGNSS) launched in recent years, provides a large amount of spaceborne GNSS Reflectometry data with all-weather, global coverage, high space-time resolution, and multiple signal sources, which provides new opportunities for the machine learning (ML) study of sea surface height (SSH) inversion. This article proposes for the first time two different CYGNSS SSH inversion models based on two widely used ML methods, back propagation (BP) neural network and convolutional neural network (CNN). The SSH calculated by using Danmarks Tekniske Universitet (DTU) 18 ocean wide mean SSH (MSSH) model (DTU18) with DTU global ocean tide model is used for verification. According to the strategy of independent analysis of data from different signal sources, the mean absolute error (MAE) of the BP and CNN models’ inversion specular points’ results during 7 days is 1.04 m and 0.63 m, respectively. The CLS 2015 product and Jason-3 data were also used for further validation. In addition, the generalization ability of the model, for 6 days and 13 days training sets, was also evaluated. For 6 days training set, the prediction results’ MAE of the BP model is 11.59 m and 5.90 m for PRN2 and PRN4, and the MAE of the CNN model is 1.37 m and 0.97 m for PRN2 and PRN4, respectively. The results show that BP and CNN inversions are in high agreement with each product, and the CNN model has relatively higher accuracy and better generalization ability.

Light and color uniformity is essential for the production of high-quality remote-sensing image mosaics. Existing color correction methods mainly use flexible models to express the color differences between multiple images and impose specific constraints (e.g., image gradient or contrast constraints) to preserve image texture information as much as possible. Due to these constraints, it is usually difficult to correct for the differences in texture between images during image processing. We propose a method that can optimize the luminance, contrast, and color difference of remote-sensing images. In the YCbCr color space, this method processes the chrominance and luminance channels of the image. This is conducive to reducing the influence of the different channels. In the luminance channel, the block-based Wallis transform method is used to optimize the luminance and contrast of the image. In the chromaticity channel, to optimize the color differences, a spline curve is used as a model; the color differences are formulated as a cost function and solved using convex quadratic programming. Moreover, considering the efficiency of our method, we use a graphics processing unit to make the algorithm parallel. The proposed method has been tested on several challenging datasets that cover different topographic regions. In terms of visuals and quality indicators, it shows better results than state-of-the-art approaches.

In lunar exploration missions, path planning for lunar rovers using digital elevation models (DEMs) is currently a hot topic in academic research. However, research on path planning using large-scale DEMs has rarely been discussed, owing to the low time efficiency of existing algorithms. Therefore, in this article, we propose a fast path-planning method using a distributed tile pyramid strategy and an improved A* algorithm. The proposed method consists of three main steps. First, the tile pyramid is generated for the large lunar DEM and stored in Hadoop distributed file system. Second, a distributed path-planning strategy based on tile pyramid (DPPS-TP) is used to accelerate path-planning tasks on large-scale lunar DEMs using Spark and Hadoop. Finally, an improved A* algorithm was proposed to improve the speed of the path-planning task in each tile. The method was tested using lunar DEM images. Experimental results demonstrate that: in a single-machine serial strategy using source DEM generated by the Chang'e-2 CCD stereo camera, the proposed A* algorithm for open list and closed list with random access feature (OC-RA-A* algorithm) is 3.59 times faster than the traditional A* algorithm in long-distance path planning tasks and compared to the distributed parallel computation strategy using source DEM generated by the Chang'e-2 CCD stereo camera, the proposed DPPS-TP based on tile pyramid DEM is 113.66 times faster in the long-range path planning task.

In the coastal areas of China, the eutrophication of seawater leads to the continuous occurrence of red tide, which has caused great damage to Marine fisheries and aquatic resources. Therefore, the detection and prediction of red tide have important research significance. The rapid development of optical remote sensing technology and deep-learning technology provides technical means for realizing large-scale and high-precision red tide detection. However, the difficulty of the accurate detection of red tide edges with complex boundaries limits the further improvement of red tide detection accuracy. In view of the above problems, this paper takes GOCI data in the East China Sea as an example and proposes an improved U-Net red tide detection method. In the improved U-Net method, NDVI was introduced to enhance the characteristic information of the red tide to improve the separability between the red tide and seawater. At the same time, the ECA channel attention mechanism was introduced to give different weights according to the influence of different bands on red tide detection, and the spectral characteristics of different channels were fully mined to further extract red tide characteristics. A shallow feature extraction module based on Atrous Spatial Pyramid Convolution (ASPC) was designed to improve the U-Net model. The red tide feature information in a multi-scale context was fused under multiple sampling rates to enhance the model’s ability to extract features at different scales. The problem of limited accuracy improvement in red tide edge detection with complex boundaries is solved via the fusion of deep and shallow features and multi-scale spatial features. Compared with other methods, the method proposed in this paper achieves better results and can detect red tide edges with complex boundaries, and the accuracy, precision, recall, and F1-score are 95.90%, 97.15%, 91.53%, and 0.94, respectively. In addition, the red tide detection experiments in other regions with relatively concentrated distribution also prove that the method has good applicability.

Sea surface height (SSH) retrieval based on spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) usually uses the GNSS-R geometric principle and delay-Doppler map (DDM). The traditional method condenses the DDM information into a single scalar measure and requires error model correction. In this article, the idea of using machine learning methods to retrieve SSH is proposed. Specifically, two widely-used methods, principal component analysis combined with support vector regression (PCA-SVR) and convolution neural network (CNN), are used for verification and comparative analysis based on the observation data provided by Techdemosat-1 (TDS-1). According to the DDM inversion method, ten features from TDS-1 Level 1 data are selected as inputs; The SSH verification model based on the Danmarks Tekniske Universitet (DTU) 15 ocean wide mean SSH model and the DTU global ocean tide model is used as output verification of SSH. For the hyperparameters in the machine learning model, a grid search strategy is used to find the optimal values. By analyzing the TDS-1 data from 31 GPS satellites, the mean absolute error (MAE), root-mean-square error (RMSE) and coefficient of determination (R2) of the PCA-SVR inversion model are 0.61 m, 1.72 m, and 99.56%, respectively; and the MAE, RMSE, and R2 of the CNN inversion model is 0.71 m, 1.27 m, and 99.76%, respectively. In addition, the time required to train the PCA-SVR and CNN inversion models is also analyzed. Overall, the technique proposed in this article can be confidently applied to SSH inversion based on TDS-1 data.

Accurate recognition of circular marks is crucial for calibration, object tracking, and three-dimensional reconstruction in videogrammetry. However, most existing studies were designed under single or relatively simple scenes. When the existing algorithms are applied to more complex scenarios, it will result in higher false detection and miss-detection rate. In this article, we present a high-precision recognition method based on a novel deep learning model, circular-MarkNet (CMNet) to solve this problem. The proposed network consists of three main steps: first, circular marks are detected using the improved YOLOv4 model to narrow the search region of the circular contour; the contour of the circular marks is then extracted based on the saliency object detection model BASNet; and finally, least square fitting is used to calculate the central pixel coordinate of the identified contour on the saliency map. The proposed method was tested under three complex scenarios with different characteristics and disturbances. The experimental results demonstrated that: the proposed CMNet can effectively recognize of circular marks within complex scenes, which reveals the superiority and generalization ability of the proposed method; the improved YOLOv4 can significantly enhance the detection accuracy of circular marks, which is crucial to the subsequent saliency courter detection and circle center identification; and CMNet achieved the best performance, with an RMSE of 0.0713 pixel, compared to the state-of-the-art methods.

Deep learning detection methods use in ship detection remains a challenge, owing to the small scale of the objects and interference from complex sea surfaces. In addition, existing ship detection methods rarely verify the robustness of their algorithms on multisensor images. Thus, we propose a new improvement on the “you only look once” version 3 (YOLOv3) framework for ship detection in marine surveillance, based on synthetic aperture radar (SAR) and optical imagery. First, improved choices are obtained for the anchor boxes by using linear scaling based on the k-means++ algorithm. This addresses the difficulty in reflecting the advantages of YOLOv3's multiscale detection, as the anchor boxes of a single detection target type between different detection scales have small differences. Second, we add uncertainty estimators for the positioning of the bounding boxes by introducing a Gaussian parameter for ship detection into the YOLOv3 framework. Finally, four anchor boxes are allocated to each detection scale in the Gaussian-YOLO layer instead of three as in the default YOLOv3 settings, as there are wide disparities in an object's size and direction in remote sensing images with different resolutions. Applying the proposed strategy to ``YOLOv3-spp” and ``YOLOv3-tiny,” the results are enhanced by 2%–3%. Compared with other models, the improved-YOLOv3 has the highest average precision on both the optical (93.56%) and SAR (95.52%) datasets. The improved-YOLOv3 is robust, even in the context of a mixed dataset of SAR and optical images comprising images from different satellites and with different scales.

Fire is an important ecosystem process and has played a complex role in terrestrial ecosystems and the atmosphere environment. Sometimes, wildfires are highly destructive natural disasters. To reduce their destructive impact, wildfires must be detected as soon as possible. However, accurate and timely monitoring of wildfires is a challenging task due to the traditional threshold methods easily be suffered to the false alarms caused by small forest clearings, and the omission error of large fires obscured by thick smoke. Deep learning has the characteristics of strong learning ability, strong adaptability and good portability. At present, few studies have addressed the wildfires detection problem in remote sensing images using deep learning method in a nearly real time way. Therefore, in this research we proposed an active fire detection system using a novel convolutional neural network (FireCNN). FireCNN uses multi-scale convolution and residual acceptance design, which can effectively extract the accurate characteristics of fire spots. The proposed method was tested on dataset which contained 1,823 fire spots and 3,646 non-fire spots. The experimental results demonstrate that the FireCNN is fully capable of wildfire detection, with the accuracy of 35.2% higher than the traditional threshold method. We also examined the influence of different structural designs on the performance of neural network models. The comparison results indicates the proposed method produced the best results.

The COVID-19 pandemic is the most serious catastrophe since the Second World War. To predict the epidemic more accurately under the influence of policies, a framework based on Independently Recurrent Neural Network (IndRNN) with fine-tuning are proposed for predict the epidemic development trend of confirmed cases and deaths in the United Stated, India, Brazil, France, Russia, China, and the world to late May, 2021. The proposed framework consists of four main steps: data pre-processing, model pre-training and weight saving, the weight fine-tuning, trend predicting and validating. It is concluded that the proposed framework based on IndRNN and fine-tuning with high speed and low complexity, has great fitting and prediction performance. The applied fine-tuning strategy can effectively reduce the error by up to 20.94% and time cost. For most of the countries, the MAPEs of fine-tuned IndRNN model were less than 1.2%, the minimum MAPE and RMSE were 0.05%, and 1.17, respectively, by using Chinese deaths, during the testing phase. According to the prediction and validation results, the MAPEs of the proposed framework were less than 6.2% in most cases, and it generated lowest MAPE and RMSE values of 0.05% and 2.14, respectively, for deaths in China. Moreover, Policies that play an important role in the development of COVID-19 have been summarized. Timely and appropriate measures can greatly reduce the spread of COVID-19; untimely and inappropriate government policies, lax regulations, and insufficient public cooperation are the reasons for the aggravation of the epidemic situations. The code is available at https://github.com/zhhongsh/COVID19-Precdiction. And the prediction by IndRNN model with fine-tuning are now available online (http://47.117.160.245:8088/IndRNNPredict).

A timely and accurate damage assessment of buildings after an earthquake is critical for the safety of people and property. Most of the existing methods based on classification and segmentation use two-dimensional information to determine the damage level of the buildings, which cannot provide the multi-view information of the damaged building, resulting in inaccurate assessment results. According to the knowledge of the authors, there is no related research using the deep-learning-based 3D reconstruction method for the evaluation of building damage. In this paper, we first applied the deep-learning-based MVS model to reconstruct the 3D model of the buildings after an earthquake using multi-view UAV images, to assist the building damage assessment task. The method contains three main steps. Firstly, the camera parameters are calculated. Then, 3D reconstruction is conducted based on CasMVSNet. Finally, a building damage assessment is performed based on the 3D reconstruction result. To evaluate the effectiveness of the proposed method, the method was tested in multi-view UAV aerial images of Yangbi County, Yunnan Province. The results indicate that: (1) the time efficiency of CasMVSNet is significantly higher than that of other deep learning models, which can meet the timeliness requirement of post-earthquake rescue and damage assessment. In addition, the memory consumption of CasMVSNet is the lowest; (2) CasMVSNet exhibits the best 3D reconstruction result in both high and small buildings; (3) the proposed method can provide detail and multi-view information of damaged buildings, which can be used to assist the building damage assessment task. The results of the building damage assessment are very similar to the results of the field survey.

The accurate and timely identification of the degree of building damage is critical for disaster emergency response and loss assessment. Although many methods have been proposed, most of them divide damaged buildings into two categories—intact and damaged—which is insufficient to meet practical needs. To address this issue, we present a novel convolutional neural network—namely, the earthquake building damage classification net (EBDC-Net)—for assessment of building damage based on post-disaster aerial images. The proposed network comprises two components: a feature extraction encoder module, and a damage classification module. The feature extraction encoder module is employed to extract semantic information on building damage and enhance the ability to distinguish between different damage levels, while the classification module improves accuracy by combining global and contextual features. The performance of EBDC-Net was evaluated using a public dataset, and a large-scale damage assessment was performed using a dataset of post-earthquake unmanned aerial vehicle (UAV) images. The results of the experiments indicate that this approach can accurately classify buildings with different damage levels. The overall classification accuracy was 94.44%, 85.53%, and 77.49% when the damage to the buildings was divided into two, three, and four categories, respectively.

The feasibility of Antarctic sea ice detection based on shipborne global positioning system reflectometry (GPS-R) technology is shown in this paper. Because the permittivity of sea water is much higher than that of sea ice, the reflected left-handed circular polarized (LHCP) GPS signal (RL) reflection coefficient of sea water is markedly higher than that of sea ice. The polarization ratio of RL to the direct right-handed circular polarized (RHCP) GPS signal (DR) is used to distinguish between sea water and sea ice in this paper. The experiment was performed on the ship “XueLong” for approximately 9 days from December 2014 to January 2015 during the 31st Chinese National Antarctic Research Expedition (CHINARE 31). The sea ice concentration data with a 25 km × 25 km spatial resolution derived from the National Snow and Ice Data Center (NSIDC) are used for validation and some pictures of sea ice taken from “XueLong” are shown for comparison. The polarization ratios (RL/DR) are calculated, and the correlation coefficient between the polarization ratios (RL/DR) and the sea ice concentrations is −0.66.

Global navigation satellite system reflectometry technology (GNSS-R) is rarely used for river flow velocity inversion, and in particular, there is currently no research using the BeiDou Navigation Satellite System reflectometry technology (BDS-R) for river flow velocity inversion. In this paper, a carrier phase observation of river flow velocity inversion model is proposed. The interference phase is the integral of the Doppler frequency. The raw intermediate frequency (IF) data sets are processed through an open-loop method to obtain the Doppler frequency observation generated by river flow and then realize velocity inversion. The shore-based river current measurement was conducted on the south bank of Dashengguan Yangtze River in Nanjing city, Jiangsu Province, for nearly two hours on 22 April 2021. After realizing the inversion of river flow velocity in GPS L1, the combined inversion of BDS B1I GEO satellite and IGSO satellite is realized for the first time, which demonstrates the feasibility of river flow velocity inversion using BDS reflected signals. Compared with the real river flow velocity, the GPS L1 PRN 4 (1st period) inversion precision reaches up to 0.028 m/s (mean absolute error, MAE) and 0.036 m/s (root mean square error, RMSE). In parallel, BDS GEO 2 inversion precision can reach 0.048 m/s (MAE) and 0.063 m/s (RMSE), and BDS IGSO 10 inversion precision is 0.061 m/s (MAE) and 0.073 m/s (RMSE). These results illustrate that satellite elevation change rate and distance between specular points and current meter may have a negative effect on the accuracy of river flow velocity inversion. Specular points obstructed by obstacles or too far from the velocity meter may introduce uncertain error in both MAE and RMSE. Neither the satellite elevation nor the signal strength has an obvious correlation with inversion precision, which is consistent with the theoretical principle.

Due to rapid global warming, the relationship between the mass loss of the Antarctic ice sheet and rising sea levels are attracting widespread attention. The Lambert–Amery glacial system is the largest drainage system in East Antarctica, and its mass balance has an important influence on the stability of the Antarctic ice sheet. In this paper, the recent ice flux in the Lambert Glacier of the Lambert–Amery system was systematically analyzed based on recently updated remote sensing data. According to Landsat-8 ice velocity data from 2018 to April 2019 and the updated Bedmachine v2 ice thickness dataset in 2021, the contribution of ice flux approximately 140 km downstream from Dome A in the Lambert Glacier area to downstream from the glacier is 8.5 ± 1.9 Gt·a−1, and the ice flux in the middle of the convergence region is 18.9 ± 2.9 Gt·a−1. The ice mass input into the Amery ice shelf through the grounding line of the whole glacier is 19.9 ± 1.3 Gt·a−1. The ice flux output from the mainstream area of the grounding line is 19.3 ± 1.0 Gt·a−1. Using the annual SMB data of the regional atmospheric climate model (RACMO v2.3) as the quality input, the mass balance of the upper, middle, and lower reaches of the Lambert Glacier was analyzed. The results show that recent positive accumulation appears in the middle region of the glacier (about 74–78°S, 67–85°E) and the net accumulation of the whole glacier is 2.4 ± 3.5 Gt·a−1. Although the mass balance of the Lambert Glacier continues to show a positive accumulation, and the positive value in the region is decreasing compared with values obtained in early 2000.

To overcome the limitation of poor processing times for long-distance off-road path planning, an improved A-Star algorithm based on terrain data is proposed in this study. The improved A-Star algorithm for long-distance off-road path planning tasks was developed to identify a feasible path between the start and destination based on a terrain data map generated using a digital elevation model. This study optimised the algorithm in two aspects: data structure, retrieval strategy. First, a hybrid data structure of the minimum heap and 2D array greatly reduces the time complexity of the algorithm. Second, an optimised search strategy was designed that does not check whether the destination is reached in the initial stage of searching for the global optimal path, thus improving execution efficiency. To evaluate the efficiency of the proposed algorithm, three different off-road path planning tasks were examined for short-, medium-, and long-distance path planning tasks. Each group of tasks corresponded to three different off-road vehicles, and nine groups of experiments were conducted. The experimental results show that the processing efficiency of the proposed algorithm is significantly better than that of the conventional A-Star algorithm. Compared with the conventional A-Star algorithm, the path planning efficiency of the improved A-Star algorithm was accelerated by at least 4.6 times, and the maximum acceleration reached was 550 times for long-distance off-road path planning. The simulation results show that the efficiency of long-distance off-road path planning was greatly improved by using the improved algorithm.

In view of the difficulty of wind direction retrieval in the case of the large space and time span of the global sea surface, a method of sea surface wind direction retrieval using a support vector machine (SVM) is proposed. This paper uses the space-borne global navigation satellite systems reflected signal (GNSS-R) as the remote sensing signal source. Using the Cyclone Global Navigation Satellite System (CYGNSS) satellite data, this paper selects a variety of feature parameters according to the correlation between the features of the sea surface reflection signal and the wind direction, including the Delay Doppler Map (DDM), corresponding to the CYGNSS satellite parameters and geometric feature parameters. The Radial Basis Function (RBF) is selected, and parameter optimization is performed through cross-validation based on the grid search method. Finally, the SVM model of sea surface wind direction retrieval is established. The result shows that this method has a high retrieval classification accuracy using the dataset with wind speed greater than 10 m/s, and the root mean square error (RMSE) of the retrieval result is 26.70°.

In response to the deficiency of the detection capability of traditional remote sensing means (scatterometer, microwave radiometer, etc.) for high wind speed above 25 m/s, this paper proposes a GNSS-R technique combined with a machine learning method to invert high wind speed at sea surface. The L1-level satellite-based data from the Cyclone Global Navigation Satellite System (CYGNSS), together with the European Centre for Medium-Range Weather Forecasts (ECMWF) and the National Centers for Environmental Prediction (NCEP) data, constitute the original sample set, which is processed and trained with Support Vector Regression (SVR), the combination of Principal Component Analysis (PCA) and SVR (PCA-SVR), and Convolutional Neural Network (CNN) methods, respectively, to finally construct a sea surface high wind speed inversion model. The three models for high wind speed inversion are certified by the test data collected during Typhoon Bavi in 2020. The results show that all three machine learning models can be used for high wind speed inversion on sea surface, among which the CNN method has the highest inversion accuracy with a mean absolute error of 2.71 m/s and a root mean square error of 3.80 m/s. The experimental results largely meet the operational requirements for high wind speed inversion accuracy.

Sea ice is one of the most prominent causes of marine disasters occurring at high latitudes. The detection of sea ice is particularly important, and the classification of sea ice images is an important part of sea ice detection. Traditional sea ice classification based on optical remote sensing mostly uses spectral information only and does not fully extract rich spectral and spatial information from sea ice images. At the same time, it is difficult to obtain samples and the resulting small sample sizes used in sea ice classification has limited the improvement of classification accuracy to a certain extent. In response to the above problems, this paper proposes a hyperspectral sea ice image classification method involving spectral-spatial-joint features based on the principal component analysis (PCA) network. First, the method uses the gray-level co-occurrence matrix (GLCM) and Gabor filter to extract textural and spatial information about sea ice. Then, the optimal band combination is extracted with a band selection algorithm based on a hybrid strategy, and the information hidden in the sea ice image is deeply extracted through a fusion of spectral and spatial features. Then, the PCA network is designed based on principal component analysis filters in order to extract the depth features of sea ice more effectively, and hash binarization maps and block histograms are used to enhance the separation and reduce the dimensions of features. Finally, the low-level features in the data form more abstract and invariant high-level features for sea ice classification. In order to verify the effectiveness of the proposed method, we conducted experiments on two different data collection points in Bohai Bay and Baffin Bay. The experimental results show that, compared with other single feature and spectral-spatial-joint feature algorithms, the proposed method achieves better sea ice classification results (94.15% and 96.86%) by using fewer training samples and a shorter training time.

Sea ice is one of the typical causes of marine disasters. Sea ice image classification is an important component of sea ice detection. Optical data contain rich spectral information, but they do not allow one to easily distinguish between ground objects with a similar spectrum and foreign objects with the same spectrum. Synthetic aperture radar (SAR) data contain rich texture information, but the data usually have a single source. The limitation of single-source data is that they do not allow for further improvements of the accuracy of remote sensing sea ice classification. In this paper, we propose a method for sea ice image classification based on deep learning and heterogeneous data fusion. Utilizing the advantages of convolutional neural networks (CNNs) in terms of depth feature extraction, we designed a deep learning network structure for SAR and optical images and achieve sea ice image classification through feature extraction and a feature-level fusion of heterogeneous data. For the SAR images, the improved spatial pyramid pooling (SPP) network was used and texture information on sea ice at different scales was extracted by depth. For the optical data, multi-level feature information on sea ice such as spatial and spectral information on different types of sea ice was extracted through a path aggregation network (PANet), which enabled low-level features to be fully utilized due to the gradual feature extraction of the convolution neural network. In order to verify the effectiveness of the method, two sets of heterogeneous sentinel satellite data were used for sea ice classification in the Hudson Bay area. The experimental results show that compared with the typical image classification methods and other heterogeneous data fusion methods, the method proposed in this paper fully integrates multi-scale and multi-level texture and spectral information from heterogeneous data and achieves a better classification effect (96.61%, 95.69%).

Sea ice is one of the causes of marine disasters. The classification of sea ice images is an important part of sea ice detection. The labeled samples in hyperspectral sea ice image classification are difficult to acquire, which causes minor sample problems. In addition, most of the current sea ice classification methods mainly use spectral features for shallow learning, which also limits further improvement of the sea ice classification accuracy. Therefore, this paper proposes a hyperspectral sea ice image classification method based on the spectral-spatial-joint feature with deep learning. The proposed method first extracts sea ice texture information by the gray-level co-occurrence matrix (GLCM). Then, it performs dimensionality reduction and a correlation analysis of the spectral information and spatial information of the unlabeled samples, respectively. It eliminates redundant information by extracting the spectral-spatial information of the neighboring unlabeled samples of the labeled sample and integrating the information with the spectral and texture data of the labeled sample to further enhance the quality of the labeled sample. Lastly, the three-dimensional convolutional neural network (3D-CNN) model is designed to extract the deep spectral-spatial features of sea ice. The proposed method combines relevant textural features and performs spectral-spatial feature extraction based on the 3D-CNN model by using a large amount of unlabeled sample information. In order to verify the effectiveness of the proposed method, sea ice classification experiments are carried out on two hyperspectral data sets: Baffin Bay and Bohai Bay. Compared with the CNN algorithm based on a single feature (spectral or spatial) and other CNN algorithms based on spectral-spatial features, the experimental results show that the proposed method achieves better sea ice classification (98.52% and 97.91%) with small samples. Therefore, it is more suitable for classifying hyperspectral sea ice images.