Search

The advent of deep learning has propelled significant advancements in object detection, thereby enhancing the intelligence of underwater autonomous driving systems. In this paper, we explore the cutting-edge applications of autonomous driving technology in the field of underwater exploration, addressing the pivotal role of target detection in navigating and executing tasks within challenging marine environments. In this study, the object detection capability of such systems is enhanced by integrating deep learning and multisensor fusion technology, especially by combining high-precision sensor data with multitask learning models to achieve efficient and robust detection. Our study has three principal contributions. First, we introduce a novel light perception detection system that combines monocular camera technology with 4D radar. It enriches environmental perception by weaving in radar signals and significantly enhances the accuracy and stability of target detection. Second, we have developed a dual-modal detection framework, named Radar-Picture Detection, which utilizes a parallel sequence prediction method. This approach prioritizes radar signal processing, aiding in the improvement of target detection accuracy in intricate underwater environments. Third, we conducted a comprehensive evaluation of our model's performance using the FloW Dataset, which is specifically curated for identifying floating waste in inland waters through unmanned vessel footage. We not only propel forward the field of target detection for underwater autonomous systems but also establish new avenues and a solid foundation for deploying deep learning and multisensor fusion technology in marine environmental perception. Insights and methodologies from this study are poised to spearhead further developments in autonomous marine exploration, enhancing safety, efficiency, and our understanding of underwater environments. [ABSTRACT FROM AUTHOR]

Object detection is one of the critical technologies in intelligent remote sensing, integrating sensor design, image acquisition, and analysis to locate targets in images, especially crucial for the perception and communication of unmanned aerial vehicle (UAV) systems. In recent years, significant progress has been made in object detection based on deep learning, such as convolutional neural networks (CNNs). However, existing methods, particularly in adapting to remote sensing sensor systems, still face two main challenges: deep learning framework overfitting and inefficient multiscale object detection. To bridge the gap between remote sensing sensor design and image analysis algorithms, we propose a lightweight detector based on context feature adaptive redetecting, ReAC-DETR, an enhanced version of the baseline detector Faster-RCNN. ReAC-DETR further improves with Fast-RCNN by optimizing the feature extraction and fusion branches, making it more suitable for small object detection. It also alleviates the problem of vague object features in remote sensing perception through an object saliency enhancement algorithm. Finally, we propose a spatial context adaptive analysis algorithm to enhance the detector's capability to adjust to multiscale object detection and improve detection precision. ReAC-DETR can provide a more robust detection technology for remote sensing, enabling the system to adapt to various shooting devices, methods, targets, and environments. We tested ReAC-DETR on two challenging datasets and achieved excellent performance. [ABSTRACT FROM AUTHOR]

Clustering is an unsupervised machine learning method grouping data samples into clusters of similar objects, used as a system support tool in numerous applications such as banking customers profiling, document retrieval, image segmentation, and e-commerce recommendation engines. The effectiveness of several clustering techniques is sensible to the initialization parameters, and different solutions have been proposed in the literature to overcome this limitation. They require high computational memory consumption when dealing with big data. In this paper, we propose the application of a recent object detection Deep Learning model (YOLO-v5) for assisting the initialization of classical techniques and improving their effectiveness on two-variate datasets, leveraging the accuracy and reducing dramatically the memory and time consumption of classical clustering methods. [ABSTRACT FROM AUTHOR]

Traditional maritime search and rescue methods mainly rely on manual operation, which takes a long time to identify and results in poor search and rescue results. This paper applies computer vision technology to the field of maritime distress object detection and proposes an improved object detection algorithm YOLOv5s‐EFOE based on the YOLOv5s algorithm. Firstly, the authors change the detection head of the YOLOv5s algorithm to use a mixed channel strategy to build a more efficient decoupling head, reducing the number of 3 × 3 convolutional layers in the middle of the decoupling head to only one. The width of the head is scaled by the width multipliers of Backbone and Neck. Secondly, in conjunction with the SimOTA matching algorithm, the positive samples of the target to be tested are dynamically allocated to improve the recognition ability of different targets. Finally, considering the low pixel value of the maritime distress target from the perspective of unmanned aerial vehicles (UAV), the loss function CIoU in the original YOLOv5s is changed to EIoU. EIoU not only considers the distance and aspect ratio of the centre point, but also considers the true difference between the width and height of the prediction box and the real box, which improves the prediction accuracy of the anchor box. Experiments are conducted on a subset of the public dataset SeaDronesSee. The Recall$Recall$ of the YOLOv5s‐EFOE model proposed in this paper reached 75.9%, mAP0.5$mA{P_{0.5}}$ reached 79.9%, and mAP0.5−0.95$mA{P_{0.5 - 0.95}}$ reached 44.5%. These indicators are superior to the original YOLOv5s model, YOLOv7 series models, and YOLOv8 series models. Compared with the original model, the YOLOv5s‐EFOE model has increased the Recall$Recall$ by 5.4%, mAP0.5$mA{P_{0.5}}$ by 5.6%, and mAP0.5−0.95$mA{P_{0.5 - 0.95}}$ by 4.6%, improving the difficult to detect and missed detection situations. This model can be deployed on UAVs and can effectively identify maritime distress target, achieving the purpose of search and rescue. [ABSTRACT FROM AUTHOR]

Accurate detection of algae in microscopic image plays a crucial role in water quality monitoring. However, the existing object detection methods still face challenges in accurately detecting different categories of algae in microscopic image. In order to improve the accuracy, an improved YOLOv5s model is proposed for microalgae detection, by combining a Receptive Field Enhancement (RFE) module, the Wise‐IoU v3 dynamic non‐monotonic focal loss function, and a Dynamic head (Dyhead), which is termed receptive field enhancement wise‐IoU dyhead (RWD)‐you only look once (YOLO). Firstly, to detect microalgae of various scales, the Bottleneck in the C3 module of YOLOv5s is replaced with a more reasonable RFE module. Secondly, Wise‐IoU v3 is applied to enhance detection accuracy by assigning varying weights between high‐quality and low‐quality images. Finally, Dyhead is introduced to enhance the representation capacity of the detection head by integrating three attention mechanisms: scale awareness, spatial awareness, and task awareness. The proposed RWD‐YOLO model significantly enhances the accuracy of algae detection in microscopic image. Specifically, the experimental results on the microalgae dataset show that the RWD‐YOLO achieves an mAP@0.5 of 93.2% and an mAP@0.5:0.95 of 65.1%. Compared to the original YOLOv5s, mAP@0.5 and mAP@0.5:0.95 are improved by 3.7% and 5.7%, respectively. [ABSTRACT FROM AUTHOR]

In this paper, a framework of obstacle avoidance algorithm applied to power line damage safety distance detection is constructed, and its overall architecture and key processes are described in detail. The system design covers three core modules: visual data acquisition and preliminary processing, accurate target recognition and distance measurement, and system error analysis and correction. In the visual data processing chain, we deeply analyze every step from image acquisition to preprocessing to feature extraction, aiming to enhance the adaptability of applications to complex scenes. The target recognition and distance estimation part integrates advanced technology of deep learning to improve the reliability of recognition accuracy and distance estimation. In addition, many common error sources, such as system bias, parallax discontinuity, fluctuation of illumination conditions, etc., are discussed in depth, and corresponding correction strategies are proposed to ensure the accuracy and stability of the system, which provides powerful technical support for achieving efficient and accurate safety monitoring. Specifically, by carefully adjusting the learning rate, convolution kernel size, batch size, pooling layer type, and number of hidden layer nodes, we succeeded in improving the overall accuracy from the initial average of 92.4–95%, and the error rate decreased accordingly. [ABSTRACT FROM AUTHOR]

As a fundamental computer vision task, crowd counting plays an important role in public safety. Currently, deep learning based head detection is a promising method for crowd counting. However, the highly concerned object detection networks cannot be well applied to this problem for three reasons: (1) Existing loss functions fail to address sample imbalance in highly dense and complex scenes; (2) Canonical object detectors lack spatial coherence in loss calculation, disregarding the relationship between object location and background region; (3) Most of the head detection datasets are only annotated with the center points, i.e. without bounding boxes. To overcome these issues, we propose a novel Mask Focal Loss (MFL) based on heatmap via the Gaussian kernel. MFL provides a unifying framework for the loss functions based on both heatmap and binary feature map ground truths. Additionally, we introduce GTA_Head, a synthetic dataset with comprehensive annotations, for evaluation and comparison. Extensive experimental results demonstrate the superior performance of our MFL across various detectors and datasets, and it can reduce MAE and RMSE by up to 47.03% and 61.99%, respectively. Therefore, our work presents a strong foundation for advancing crowd counting methods based on density estimation. [ABSTRACT FROM AUTHOR]

Background: Ripeness is a phenotype that significantly impacts the quality of fruits, constituting a crucial factor in the cultivation and harvesting processes. Manual detection methods and experimental analysis, however, are inefficient and costly. Results: In this study, we propose a lightweight and efficient melon ripeness detection method, MRD-YOLO, based on an improved object detection algorithm. The method combines a lightweight backbone network, MobileNetV3, a design paradigm Slim-neck, and a Coordinate Attention mechanism. Additionally, we have created a large-scale melon dataset sourced from a greenhouse based on ripeness. This dataset contains common complexities encountered in the field environment, such as occlusions, overlapping, and varying light intensities. MRD-YOLO achieves a mean Average Precision of 97.4% on this dataset, achieving accurate and reliable melon ripeness detection. Moreover, the method demands only 4.8 G FLOPs and 2.06 M parameters, representing 58.5% and 68.4% of the baseline YOLOv8n model, respectively. It comprehensively outperforms existing methods in terms of balanced accuracy and computational efficiency. Furthermore, it maintains real-time inference capability in GPU environments and demonstrates exceptional inference speed in CPU environments. The lightweight design of MRD-YOLO is anticipated to be deployed in various resource constrained mobile and edge devices, such as picking robots. Particularly noteworthy is its performance when tested on two melon datasets obtained from the Roboflow platform, achieving a mean Average Precision of 85.9%. This underscores its excellent generalization ability on untrained data. Conclusions: This study presents an efficient method for melon ripeness detection, and the dataset utilized in this study, alongside the detection method, will provide a valuable reference for ripeness detection across various types of fruits. [ABSTRACT FROM AUTHOR]

Defect detection in pharmaceutical blister packages is the most challenging task to get an accurate result in detecting defects that arise in tablets while manufacturing. Conventional defect detection methods include human intervention to check the quality of tablets within the blister packages, which is inefficient, time-consuming, and increases labor costs. To mitigate this issue, the YOLO family is primarily used in many industries for real-time defect detection in continuous production. To enhance the feature extraction capability and reduce the computational overhead in a real-time environment, the CBS-YOLOv8 is proposed by enhancing the YOLOv8 model. In the proposed CBS-YOLOv8, coordinate attention is introduced to improve the feature extraction capability by capturing the spatial and cross-channel information and also maintaining the long-range dependencies. The BiFPN (weighted bi-directional feature pyramid network) is also introduced in YOLOv8 to enhance the feature fusion at each convolution layer to avoid more precise information loss. The model's efficiency is enhanced through the implementation of SimSPPF (simple spatial pyramid pooling fast), which reduces computational demands and model complexity, resulting in improved speed. A custom dataset containing defective tablet images is used to train the proposed model. The performance of the CBS-YOLOv8 model is then evaluated by comparing it with various other models. Experimental results on the custom dataset reveal that the CBS-YOLOv8 model achieves a mAP of 97.4% and an inference speed of 79.25 FPS, outperforming other models. The proposed model is also evaluated on SESOVERA-ST saline bottle fill level monitoring dataset achieved the mAP50 of 99.3%. This demonstrates that CBS-YOLOv8 provides an optimized inspection process, enabling prompt detection and correction of defects, thus bolstering quality assurance practices in manufacturing settings. [ABSTRACT FROM AUTHOR]

Recent advances in computer vision and deep learning techniques have facilitated significant progress in video scene understanding, thus helping film and television practitioners achieve accurate video editing. However, so far, publicly available semantic segmentation datasets are mostly limited to indoor scenes, city streets, and natural images, often ignoring example objects in action movies, which is a research gap that needs to be urgently filled. In this paper, we introduce a large-scale, high-precision semantic segmentation dataset of props in Chinese martial arts movie clips, named ChineseMPD. Specifically, this dataset first establishes segmentation rules and general review criteria for audiovisual data, and then provides semantic segmentation annotations for six weapon props (Gun, Sword, Stick, Knife, Hook, and Arrow) with a summary of 32,992 objects.To the best of our knowledge, this dataset is the largest semantic segmentation dataset for movie props to date. ChineseMPD dataset not only significantly expands the application of traditional tasks of computer vision such as object detection and scene understanding, but also opens up new avenues for interdisciplinary research. [ABSTRACT FROM AUTHOR]

Summary: Agriculture has been affected by several global trends raising the concern for food security. Agri‐food demands are amplifying due to the ever‐escalating population numbers. Thus, the notion of extending the use of smart innovative technology in managing agricultural practices has emerged rapidly over the last decade. Technological innovations have contributed significantly to shape modern agriculture as smart agriculture. Smart agriculture unfolds various benefits such as increased production, real time data and production insights and remote monitoring. The rising advancements in Information and Communication Technologies (ICT) have paved the way for researchers to use these technologies in managing agricultural practices, leading to greater benefits for farmers and society. These innovations are the key to establishing agriculture as a research discipline. The purpose of this article is to conduct a scientometric analysis to study the structure and evolution of research activities in the field of smart agriculture. The scientometric analysis aims to empirically map the scientific knowledge and identify any possible challenges in the field. This study performs elementary analysis to study publication growth over the years, impact analysis to assess the leading journals, authors, and countries, and articles analysis for findings patterns among the citations over the years and among the keyword‐based clusters. There has been a considerable increase of more than 200% in the number of publications from 2011 to 2022. However, around 60% of authors have contributed with a single publication. The findings of the study reveal the research trends and hot topics for future research fronts. Deep learning, digital agriculture, object detection, blockchain, and semantic segmentation have been identified as trending topics in smart agriculture. Comprehensively, an intellectual view of the agriculture domain is presented as a scientific field in this article. [ABSTRACT FROM AUTHOR]

Recent developments in robotics increasingly highlight the importance of sensing technology, especially tactile perception, in enabling robots to effectively engage with their environment and interpret physical interactions. Due to power efficiency and low cost, the triboelectric mechanism has been frequently studied for measuring pressure and identifying materials to enhance robot perception. Nevertheless, there has been limited exploration of using the triboelectric effect to detect curved surfaces, despite their prevalence in daily lives. Here, a triboelectric multimodal tactile sensor (TMTS) of multilayered structural design is proposed to recognize distinct materials, curvatures, and pressure simultaneously, thus decoupling different modalities to enable more accurate detection. By attaching sensors to robotic fingertips and leveraging deep learning analytics, the quantitative curvature measurement provides more precise insights into an object's detailed geometric characteristics rather than merely assessing its overall shape, hence achieving automatic recognition of 12 grasped objects with 99.2% accuracy. The sensor can be further used to accurately recognize the softness of objects under different touch gestures of a robotic hand, achieving a 94.1% accuracy, demonstrating its significant potential for wide‐ranging applications in a future robotic‐enabled intelligent society. [ABSTRACT FROM AUTHOR]

Milk thistle, Silybum marianum (L.), is a well-known medicinal plant used for the treatment of liver diseases due to its high content of silymarin. The seeds contain elaiosome, a fleshy structure attached to the seeds, which is believed to be a rich source of many metabolites including silymarin. Segmentation of elaiosomes using only image analysis is difficult, and this makes it impossible to quantify the elaiosome phenotypes. This study proposes a new approach for semi-automated detection and segmentation of elaiosomes in milk thistle seed using the Detectron2 deep learning algorithm. One hundred manually labeled images were used to train the initial elaiosome detection model. This model was used to predict elaiosome from new datasets, and the precise predictions were manually selected and used as new labeled images for retraining the model. Such semi-automatic image labeling, i.e., using the prediction results of the previous stage for retraining the model, allowed the production of sufficient labeled data for retraining. Finally, a total of 6,000 labeled images were used to train Detectron2 for elaiosome detection and attained a promising result. The results demonstrate the effectiveness of Detectron2 in detecting milk thistle seed elaiosomes with an accuracy of 99.9%. The proposed method automatically detects and segments elaiosome from the milk thistle seed. The predicted mask images of elaiosome were used to analyze its area as one of the seed phenotypic traits along with other seed morphological traits by image-based highthroughput phenotyping in ImageJ. Enabling high-throughput phenotyping of elaiosome and other seed morphological traits will be useful for breeding milk thistle cultivars with desirable traits. [ABSTRACT FROM AUTHOR]

As the pillar industry of coastal areas, aquaculture needs artificial intelligence technology to promote economic development. To realize the automation of the aquaculture industry, this paper proposes a new underwater object detection network: WBi-YOLOSF. It realizes the automatic classification and detection of aquatic products, improves the production efficiency of the aquaculture industry, and promotes its economic development. This paper creates an image dataset containing 15 aquatic products to lay the data foundation for model training. In the data preprocessing part, an underwater image enhancement algorithm is proposed to improve the quality of the data set effectively. Aiming at the problem of high false detection rate and missed detection rate of underwater dense small targets, a new data enhancement method was proposed to improve the training set's data quality comprehensively. Inspired by the weighted bidirectional feature pyramid network, this paper proposes a new feature extraction network: AU-BiFPN, which solves the gradient problem caused by the network hierarchy's deepening on enhancing the network's multi-scale feature fusion. The AU-BiFPN network structure is embedded into the YOLO series network framework, significantly improving the basic network's feature extraction and feature fusion ability and dramatically improving the prediction accuracy without affecting the network inference speed. Here, the swarm intelligence algorithm is introduced to optimize the model hyperparameters, accelerating the convergence speed of model training and significantly reducing the computational cost. At the same time, the model's accuracy is improved by a cliff. In addition, the Funnel Activation is introduced in the network's backbone, and the simple, parameter-free attention module is integrated, effectively improving the accuracy and speed of the model prediction. Ablation and comparison experiments show the effectiveness and superiority of the proposed model. Verified by the mean average precision and frame rate evaluation indicators, the experimental results of the WBi-YOLOSF target detection network can reach 0.982 and 203 frames per second, which are 1.4% and five frames per second higher than the network with the second score. In summary, this method can quickly and accurately identify aquatic products, realize real-time target detection of aquatic products, and lay the foundation for developing an aquaculture automation system. [ABSTRACT FROM AUTHOR]

This document introduces the RadIOCD, which is a dataset that contains sparse point cloud representations of indoor objects, collected by subjects wearing a commercial off-the-shelf mmWave radar. In particular, RadIOCD includes the recordings of 10 volunteers moving towards 5 different objects (i.e., backpack, chair, desk, human, and wall), placed in 3 different environments. RadIOCD includes sparse 3D point cloud data, together with their doppler velocity and intensity provided by the mmWave radar. A total of 5,776 files are available, with each one having an approximate duration of 8s. The scope of RadIOCD is the availability of data for the recognition of objects solely recorded by the mmWave radar, to be used in applications were the vision-based classification is cumbersome though critical (e.g., in search and rescue operation where there is smoke inside a building). Furthermore, we showcase that this dataset after being segmented into 76,821 samples contains enough data to apply Machine Learning-based techniques, ensuring that they could generalize in different environments and "unseen" subjects. [ABSTRACT FROM AUTHOR]

The existing ship object detection algorithms mostly rely on optimized improvements based on traditional object detection algorithms, without considering the scale and aspect ratio characteristics of ships, leading to issues such as missed detections and false detections in multi-scale object detection. To address this, the paper proposed a scale-adaptive receptive field ship detection method (SAF-YOLOX) based on YOLOXs. Firstly, it extracted different feature layers by the backbone network, which were fused by constructing a bidirectional feature pyramid, improving feature representation at various scales. Simultaneously, it designed an adaptive feature enhancement module to suppress redundant information introduced by the fusion of features at different scales, thereby attenuating background information. During the prediction phase, it employed a multi-branch parallel receptive field detection head, utilizing receptive fields adapted to target sizes and proportions for extracting scale-adaptive feature information. Additionally, it implemented a convergence-aware strategy, dynamically selecting and allocating samples based on the network s convergence state. This strategy ensured improved detection accuracy while maintaining detection speed. Experimental results demonstrate that the proposed method achieves an average detection accuracy of 93. 21% on the SeaShips dataset and 92.34% on the MCShips dataset. When compared to traditional YOLOXs, the method exhibits an improvement of 1.01% and 1.09%, respectively. The experimental results confirm that the proposed method, utilizing scale-adaptive receptive fields, can achieve high-precision detection of multi-scale ship targets. [ABSTRACT FROM AUTHOR]

The occurrence of the “structural-loss” defect during single-crystal silicon growth (SCSG) is a significant issue in semiconductor manufacturing. When structural-loss occurs, it signifies a deviation from the desired quality of single-crystal formation, leading to the need to halt the growth process. Currently, there is a lack of scholarly literature addressing the determination of an optimal stopping time to promptly halt the process upon the occurrence of the defect on-line. Our research makes a substantial contribution by addressing this gap in the SCSG process, specifically focusing on orientations <100> and <111>. The study utilizes advanced AI with YOLO-v7 and innovative approaches. These include precise annotation of crystal misorientation features through a comprehensive definition of structural-loss and novel labeling techniques, identification of optimal hyper-parameters through a robust design, and the implementation of effective stopping rule mechanisms. Significant progress has been achieved in decision-making through the implementation of the stoping time shift to terminate the SCSG process within an average of less than 3 minutes for <100> orientations (with a standard error of 0.3 minutes) and less than 5 minutes for <111> orientations (with a standard error of 0.5 minutes). The promising results indicate that the proposed approaches have the capability to substitute manual inspections, opening up possibilities for new perspectives in this particular field. [ABSTRACT FROM AUTHOR]

Gait recognition is to recognise different individuals based on their faint differences of gait characteristics, which is different from and more challengeable than the recognition of human activities based on relatively bigger differences between different motions. Existing millimetre‐wave Multiple Input Multiple Output radar point cloud data contains time‐varying three‐dimensional spatial positions, velocity, and intensity information. How to enhance the accuracy of gait recognition by effectively utilising the available radar point cloud data has become an attractive research topic in recent years. A velocity‐depth‐time (VDT) based point cloud construction method for millimetre‐wave Multiple Input Multiple Output radar is proposed for gait recognition application, which can not only alleviate the sparsity problem of mmWave point cloud but also make the constructed point cloud to exhibit temporal structural features of micro‐motions, and therefore enable the successful application of PointNet++ to mmWave‐MIMO point cloud gait recognition. New point clouds are constructed by the proposed method using public gait recognition datasets of 10 and 20 individuals from mmWave‐MIMO radar, which are used to conduct gait recognition experiments using PointNet++. The results show that the point clouds constructed based on VDT are more conducive to the gait recognition task. Even using the classic PointNet++ model, which is not specially designed for radar point clouds, high recognition accuracy can be achieved for gait recognition tasks. The recognition accuracies are improved by 11% and 12% in this work for datasets of 10 and 20 individuals, respectively, compared with the 84% and 80% achieved by the traditional method using the same dataset and the same PointNet++ model, while the accuracies are improved by 5% and 12%, respectively, compared with the 90% and 80% achieved by the original dataset thesis method corresponding to 10‐individual and 20‐individual datasets. [ABSTRACT FROM AUTHOR]

Radar High‐Resolution Range Profile (HRRP) has great potential for target recognition because it can provide target structural information. Existing work commonly applies deep learning to extract deep features from HRRPs and achieve impressive recognition performance. However, most approaches are unable to distinguish between the target and non‐target regions in the feature extraction process and do not fully consider the impact of background noise, which is harmful to recognition, especially at low signal‐to‐noise ratios (SNR). To tackle these problems, the authors propose a radar HRRP target recognition framework termed Adaptive Soft Threshold Transformer (ASTT), which is composed of a patch embedding (PE) layer, ASTT blocks, and Discrete Wavelet Patch Merging (DWPM) layers. Given the limited semantic information of individual range cells, the PE layer integrates nearby isolated range cells into semantically explicit target structure patches. Thanks to its convolutional layer and attention mechanism, the ASTT blocks assign a weight to each patch to locate the target areas in the HRRP while capturing local features and constructing sequence correlations. Moreover, the ASTT block efficiently filters noise features in combination with a soft threshold function to further enhance the recognition performance at low SNR, where the threshold is adaptively determined. Utilising the reversibility of the discrete wavelet transform, the DWPM layer efficiently eliminates the loss of valuable information during the pooling process. Experiments based on simulated and measured datasets show that the proposed method has excellent target recognition performance, noise robustness, and small‐scale range shift robustness. [ABSTRACT FROM AUTHOR]

Semi-supervised object detection helps to monitor and manage maritime transportation effectively, saving labeling costs. Currently, many semi-supervised object detection methods use a combination of data augmentation and pseudo-label to improve model performance. However, these methods may get into trouble in complex maritime scenes, including occlusion, scale variations and lighting variations, leading to distribution bias between labeled data and unlabeled data and pseudo-label bias. To address these problems, we propose a semi-supervised object detection method in complex maritime scenes based on adaptive adversarial self-training, which provides a teacher–student detection framework to use a robust pseudo-label with data augmentation. The proposed method contains two modules called adversarial distribution discriminator and label adaptive assigner. The adversarial distribution discriminator is proposed to match the distribution between augmented data generated from different data augmentations, while the label adaptive assigner is proposed to reduce the labeling bias for unlabeled data so that the pseudo-label of unlabeled data contributes to the detection performance effectively. Experimental results show that the proposed method achieves a better mean average precision of 91.4%, with only 5% of the labeled samples compared with other semi-supervised object detection methods, and its detection speed is 11.1 frames per second. Experiments also demonstrate that the proposed method improves the detection performance compared with fully supervised detectors. [ABSTRACT FROM AUTHOR]

High-complexity network models are challenging to execute on agricultural robots with limited computing capabilities in a large-scale pineapple planting environment in real time. Traditional module replacement often struggles to reduce model complexity while maintaining stable network accuracy effectively. This paper investigates a pineapple detection framework with a YOLOv7-tiny model improved via pruning and a lightweight backbone sub-network (the RGDP-YOLOv7-tiny model). The ReXNet network is designed to significantly reduce the number of parameters in the YOLOv7-tiny backbone network layer during the group-level pruning process. Meanwhile, to enhance the efficacy of the lightweight network, a GSConv network has been developed and integrated into the neck network, to further diminish the number of parameters. In addition, the detection network incorporates a decoupled head network aimed at separating the tasks of classification and localization, which can enhance the model's convergence speed. The experimental results indicate that the network before pruning optimization achieved an improvement of 3.0% and 2.2%, in terms of mean average precision and F1 score, respectively. After pruning optimization, the RGDP-YOLOv7-tiny network was compressed to just 2.27 M in parameter count, 4.5 × 10 9 in computational complexity, and 5.0MB in model size, which were 37.8%, 34.1%, and 40.7% of the original YOLOv7-tiny network, respectively. Concurrently, the mean average precision and F1 score reached 87.9% and 87.4%, respectively, with increases of 0.8% and 1.3%. Ultimately, the model's generalization performance was validated through heatmap visualization experiments. Overall, the proposed pineapple object detection framework can effectively enhance detection accuracy. In a large-scale fruit cultivation environment, especially under the constraints of hardware limitations and limited computational power in the real-time detection processes of agricultural robots, it facilitates the practical application of artificial intelligence algorithms in agricultural engineering. [ABSTRACT FROM AUTHOR]

For the accurate and automatic mapping of forest stands based on very-high-resolution satellite imagery and digital orthophotos, precise object detection at the individual tree level is necessary. Currently, supervised deep learning models are primarily applied for this task. To train a reliable model, it is crucial to have an accurate tree crown annotation dataset. The current method of generating these training datasets still relies on manual annotation and labeling. Because of the intricate contours of tree crowns, vegetation density in natural forests and the insufficient ground sampling distance of the imagery, manually generated annotations are error-prone. It is unlikely that the manually delineated tree crowns represent the true conditions on the ground. If these error-prone annotations are used as training data for deep learning models, this may lead to inaccurate mapping results for the models. This study critically validates manual tree crown annotations on two study sites: a forest-like plantation on a cemetery and a natural city forest. The validation is based on tree reference data in the form of an official tree register and tree segments extracted from UAV laser scanning (ULS) data for the quality assessment of a training dataset. The validation results reveal that the manual annotations detect only 37% of the tree crowns in the forest-like plantation area and 10% of the tree crowns in the natural forest correctly. Furthermore, it is frequent for multiple trees to be interpreted in the annotation as a single tree at both study sites. [ABSTRACT FROM AUTHOR]

Deep neural networks (DNNs) have been applied in many pattern recognition or object detection applications. DNNs generally consist of millions or even billions of parameters. These demanding computational storage and requirements impede deployments of DNNs in resource-limited devices, such as mobile devices, micro-controllers. Simplification techniques such as pruning have commonly been used to slim DNN sizes. Pruning approaches generally quantify the importance of each component such as network weight. Weight values or weight gradients in training are commonly used as the importance metric. Small weights are pruned and large weights are kept. However, small weights are possible to be connected with significant weights which have impact to DNN outputs. DNN accuracy can be degraded significantly after the pruning process. This paper proposes a roulette wheel-like pruning algorithm, in order to simplify a trained DNN while keeping the DNN accuracy. The proposed algorithm generates a branch of pruned DNNs which are generated by a roulette wheel operator. Similar to the roulette wheel selection in genetic algorithms, small weights are more likely to be pruned but they can be kept; large weights are more likely to be kept but they can be pruned. The slimmest DNN with the best accuracy is selected from the branch. The performance of the proposed pruning algorithm is evaluated by two deterministic datasets and four non-deterministic datasets. Experimental results show that the proposed pruning algorithm generates simpler DNNs while DNN accuracy can be kept, compared to several existing pruning approaches. [ABSTRACT FROM AUTHOR]

Pedestrian trajectory prediction is crucial for developing collision avoidance algorithms in autonomous driving systems, aiming to predict the future movement of the detected pedestrians based on their past trajectories. The traditional methods for pedestrian trajectory prediction involve a sequence of tasks, including detection and tracking to gather the historical movement of the observed pedestrians. Consequently, the accuracy of trajectory prediction heavily relies on the accuracy of the detection and tracking models, making it susceptible to their performance. The prior research in trajectory prediction has mainly assessed the model performance using public datasets, which often overlook the errors originating from detection and tracking models. This oversight fails to capture the real-world scenario of inevitable detection and tracking inaccuracies. In this study, we investigate the cumulative effect of errors within integrated detection, tracking, and trajectory prediction pipelines. Through empirical analysis, we examine the errors introduced at each stage of the pipeline and assess their collective impact on the trajectory prediction accuracy. We evaluate these models across various custom datasets collected in Taiwan to provide a comprehensive assessment. Our analysis of the results derived from these integrated pipelines illuminates the significant influence of detection and tracking errors on downstream tasks, such as trajectory prediction and distance estimation. [ABSTRACT FROM AUTHOR]

To address the issues of decreased detection accuracy, false detections, and missed detections caused by scale differences between near and distant targets and environmental factors (such as lighting and water waves) in surface target detection tasks for uncrewed vessels, the YOLOv8-MSS algorithm is proposed to be used to optimize the detection of water surface targets. By adding a small target detection head, the model becomes more sensitive and accurate in recognizing small targets. To reduce noise interference caused by complex water surface environments during the downsampling process in the backbone network, C2f_MLCA is used to enhance the robustness and stability of the model. The lightweight model SENetV2 is employed in the neck component to improve the model's performance in detecting small targets and its anti-interference capability. The SIoU loss function enhances detection accuracy and bounding box regression precision through shape awareness and geometric information integration. Experiments on the publicly available dataset FloW-Img show that the improved algorithm achieves an mAP@0.5 of 87.9% and an mAP@0.5:0.95 of 47.6%, which are improvements of 5% and 2.6%, respectively, compared to the original model. [ABSTRACT FROM AUTHOR]

(1) Background: Small objects in Unmanned Aerial Vehicle (UAV) images are often scattered throughout various regions of the image, such as the corners, and may be blocked by larger objects, as well as susceptible to image noise. Moreover, due to their small size, these objects occupy a limited area in the image, resulting in a scarcity of effective features for detection. (2) Methods: To address the detection of small objects in UAV imagery, we introduce a novel algorithm called High-Resolution Feature Pyramid Network Mamba-Based YOLO (HRMamba-YOLO). This algorithm leverages the strengths of a High-Resolution Network (HRNet), EfficientVMamba, and YOLOv8, integrating a Double Spatial Pyramid Pooling (Double SPP) module, an Efficient Mamba Module (EMM), and a Fusion Mamba Module (FMM) to enhance feature extraction and capture contextual information. Additionally, a new Multi-Scale Feature Fusion Network, High-Resolution Feature Pyramid Network (HRFPN), and FMM improved feature interactions and enhanced the performance of small object detection. (3) Results: For the VisDroneDET dataset, the proposed algorithm achieved a 4.4% higher Mean Average Precision (mAP) compared to YOLOv8-m. The experimental results showed that HRMamba achieved a mAP of 37.1%, surpassing YOLOv8-m by 3.8% (Dota1.5 dataset). For the UCAS_AOD dataset and the DIOR dataset, our model had a mAP 1.5% and 0.3% higher than the YOLOv8-m model, respectively. To be fair, all the models were trained without a pre-trained model. (4) Conclusions: This study not only highlights the exceptional performance and efficiency of HRMamba-YOLO in small object detection tasks but also provides innovative solutions and valuable insights for future research. [ABSTRACT FROM AUTHOR]

Falls are a major concern for people with Parkinson's disease (PwPD), but accurately assessing real-world fall risk beyond the clinic is challenging. Contemporary technologies could enable the capture of objective and high-resolution data to better inform fall risk through measurement of everyday factors (e.g., obstacles) that contribute to falls. Wearable inertial measurement units (IMUs) capture objective high-resolution walking/gait data in all environments but are limited by not providing absolute clarity on contextual information (i.e., obstacles) that could greatly influence how gait is interpreted. Video-based data could compliment IMU-based data for a comprehensive free-living fall risk assessment. The objective of this study was twofold. First, pilot work was conducted to propose a novel artificial intelligence (AI) algorithm for use with wearable video-based eye-tracking glasses to compliment IMU gait data in order to better inform free-living fall risk in PwPD. The suggested approach (based on a fine-tuned You Only Look Once version 8 (YOLOv8) object detection algorithm) can accurately detect and contextualize objects (mAP50 = 0.81) in the environment while also providing insights into where the PwPD is looking, which could better inform fall risk. Second, we investigated the perceptions of PwPD via a focus group discussion regarding the adoption of video technologies and AI during their everyday lives to better inform their own fall risk. This second aspect of the study is important as, traditionally, there may be clinical and patient apprehension due to ethical and privacy concerns on the use of wearable cameras to capture real-world video. Thematic content analysis was used to analyse transcripts and develop core themes and categories. Here, PwPD agreed on ergonomically designed wearable video-based glasses as an optimal mode of video data capture, ensuring discreteness and negating any public stigma on the use of research-style equipment. PwPD also emphasized the need for control in AI-assisted data processing to uphold privacy, which could overcome concerns with the adoption of video to better inform IMU-based gait and free-living fall risk. Contemporary technologies (wearable video glasses and AI) can provide a holistic approach to fall risk that PwPD recognise as helpful and safe to use. [ABSTRACT FROM AUTHOR]

YOLOv8, as an efficient object detection method, can swiftly and precisely identify objects within images. However, traditional algorithms encounter difficulties when detecting small objects in remote sensing images, such as missing information, background noise, and interactions among multiple objects in complex scenes, which may affect performance. To tackle these challenges, we propose an enhanced algorithm optimized for detecting small objects in remote sensing images, named HP-YOLOv8. Firstly, we design the C2f-D-Mixer (C2f-DM) module as a replacement for the original C2f module. This module integrates both local and global information, significantly improving the ability to detect features of small objects. Secondly, we introduce a feature fusion technique based on attention mechanisms, named Bi-Level Routing Attention in Gated Feature Pyramid Network (BGFPN). This technique utilizes an efficient feature aggregation network and reparameterization technology to optimize information interaction between different scale feature maps, and through the Bi-Level Routing Attention (BRA) mechanism, it effectively captures critical feature information of small objects. Finally, we propose the Shape Mean Perpendicular Distance Intersection over Union (SMPDIoU) loss function. The method comprehensively considers the shape and size of detection boxes, enhances the model's focus on the attributes of detection boxes, and provides a more accurate bounding box regression loss calculation method. To demonstrate our approach's efficacy, we conducted comprehensive experiments across the RSOD, NWPU VHR-10, and VisDrone2019 datasets. The experimental results show that the HP-YOLOv8 achieves 95.11%, 93.05%, and 53.49% in the mAP@0.5 metric, and 72.03%, 65.37%, and 38.91% in the more stringent mAP@0.5:0.95 metric, respectively. [ABSTRACT FROM AUTHOR]

The intelligent harvesting technology for jujube leaf branches presents a novel avenue for enhancing both the quantity and quality of jujube leaf tea, whereas the precise detection technology for jujube leaf branches emerges as a pivotal factor constraining its development. The precise identification and localization of jujube leaf branches using real-time object detection technology are crucial steps toward achieving intelligent harvesting. When integrated into real-world scenarios, issues such as the background noise introduced by tags, occlusions, and variations in jujube leaf morphology constrain the accuracy of detection and the precision of localization. To address these issues, we describe a jujube leaf branch object detection network based on YOLOv7. First, the Polarized Self-Attention module is embedded into the convolutional layer, and the Gather-Excite module is embedded into the concat layer to incorporate spatial information, thus achieving the suppression of irrelevant information such as background noise. Second, we incorporate implicit knowledge into the Efficient Decoupled Head and replace the original detection head, enhancing the network's capability to extract deep features. Third, to address the issue of imbalanced jujube leaf samples, we employ Focal-EIoU as the bounding box loss function to expedite the regression prediction and enhance the localization accuracy of the model's bounding boxes. Experiments show that the precision of our model is 85%, which is increased by 3.5% compared to that of YOLOv7-tiny. The mAP@0.5 value is 83.7%. Our model's recognition rate, recall and mean average precision are superior to those of other models. Our method could provide technical support for yield estimation in the intelligent management of jujube orchards. [ABSTRACT FROM AUTHOR]

Robust object detection in complex environments, poor visual conditions, and open scenarios presents significant technical challenges in autonomous driving. These challenges necessitate the development of advanced fusion methods for millimeter-wave (mmWave) radar point cloud data and visual images. To address these issues, this paper proposes a radar–camera robust fusion network (RCRFNet), which leverages self-supervised learning and open-set recognition to effectively utilise the complementary information from both sensors. Specifically, the network uses matched radar–camera data through a frustum association approach to generate self-supervised signals, enhancing network training. The integration of global and local depth consistencies between radar point clouds and visual images, along with image features, helps construct object class confidence levels for detecting unknown targets. Additionally, these techniques are combined with a multi-layer feature extraction backbone and a multimodal feature detection head to achieve robust object detection. Experiments on the nuScenes public dataset demonstrate that RCRFNet outperforms state-of-the-art (SOTA) methods, particularly in conditions of low visual visibility and when detecting unknown class objects. [ABSTRACT FROM AUTHOR]

A lightweight aircraft visual navigation algorithm that fuses neural networks is proposed to address the limited computing power issue during the offline operation of aircraft edge computing platforms in satellite-denied environments with complex working scenarios. This algorithm utilizes object detection algorithms to label dynamic objects within complex scenes and performs dynamic feature point elimination to enhance the feature point extraction quality, thereby improving navigation accuracy. The algorithm was validated using an aircraft edge computing platform, and comparisons were made with existing methods through experiments conducted on the TUM public dataset and physical flight experiments. The experimental results show that the proposed algorithm not only improves the navigation accuracy but also has high robustness compared with the monocular ORB-SLAM2 method under the premise of satisfying the real-time operation of the system. [ABSTRACT FROM AUTHOR]

Monitoring nocturnal animals in the field is an important task in ecological research and wildlife conservation, but the complexity of nocturnal images and low light conditions make it difficult to cope with traditional image processing methods. To address this problem, researchers have introduced infrared cameras to improve the accuracy of nocturnal animal behaviour observations. Object detection in nighttime images captured by infrared cameras faces several challenges, including low image quality, animal scale variations, occlusion, and pose changes. This study proposes the YOLOv8‐night model, which effectively overcomes these challenges by introducing a channel attention mechanism in YOLOv8. The model is more focused on capturing animal‐related features by dynamically adjusting the channel weights, which improves the saliency of key features and increases the accuracy rate in complex backgrounds. The main contribution of this study is the introduction of the channel attention mechanism into the YOLOv8 framework to create a YOLOv8‐night model suitable for object detection in nighttime images. When tested on nighttime images, the model performs well with a significantly higher mAP (0.854) than YOLOv8 (0.831), and YOLOv8‐night scores 0.856 on mAP_l, which is obviously better than YOLOv8 (0.833) in terms of processing large objects. The study provides a reliable technical tool for ecological research, wildlife conservation and environmental monitoring, and offers new methods and insights for the study of nocturnal animal behaviour. [ABSTRACT FROM AUTHOR]