Small objects detection in UAV aerial images based on improved Faster R-CNN

In order to solve the problem of small objects detection in unmanned aerial vehicle(UAV)aerial images with complex background,a general detection method for multi-scale small objects based on Faster region-based convolutional neural network(Faster R-CNN)is proposed.The bird’s nest on the high-voltage tower is taken as the research object.Firstly,we use the improved convolutional neural network ResNet101 to extract object features,and then use multi-scale sliding windows to obtain the object region proposals on the convolution feature maps with different resolutions.Finally,a deconvolution operation is added to further enhance the selected feature map with higher resolution,and then it taken as a feature mapping layer of the region proposals passing to the object detection sub-network.The detection results of the bird’s nest in UAV aerial images show that the proposed method can precisely detect small objects in aerial images.

An Improved Non-Parametric Method for Multiple Moving Objects Detection in the Markov Random Field

Detecting moving objects in the stationary background is an important problem in visual surveillance systems.However,the traditional background subtraction method fails when the background is not completely stationary and involves certain dynamic changes.In this paper,according to the basic steps of the background subtraction method,a novel non-parametric moving object detection method is proposed based on an improved ant colony algorithm by using the Markov random field.Concretely,the contributions are as follows:1)A new nonparametric strategy is utilized to model the background,based on an improved kernel density estimation;this approach uses an adaptive bandwidth,and the fused features combine the colours,gradients and positions.2)A Markov random field method based on this adaptive background model via the constraint of the spatial context is proposed to extract objects.3)The posterior function is maximized efficiently by using an improved ant colony system algorithm.Extensive experiments show that the proposed method demonstrates a better performance than many existing state-of-the-art methods.

MULTIPLE KERNEL RELEVANCE VECTOR MACHINE FOR GEOSPATIAL OBJECTS DETECTION IN HIGH-RESOLUTION REMOTE SENSING IMAGES

Geospatial objects detection within complex environment is a challenging problem in remote sensing area. In this paper, we derive an extension of the Relevance Vector Machine (RVM) technique to multiple kernel version. The proposed method learns an optimal kernel combination and the associated classifier simultaneously. Two feature types are extracted from images, forming basis kernels. Then these basis kernels are weighted combined and resulted the composite kernel exploits interesting points and appearance information of objects simultaneously. Weights and the detection model are finally learnt by a new algorithm. Experimental results show that the proposed method improve detection accuracy to above 88%, yields good interpretation for the selected subset of features and appears sparser than traditional single-kernel RVMs.

Dot Size Thermal Objects Detection in Atmosphere of Infrared Range

The article deals with the experimental studies of atmosphere indistinct radiation structure. The information extraction background of dot size thermal object presence in atmosphere is reasonable. Indistinct generalization of experimental study regularities technique of space-time irregularity radiation structure in infrared wave range is offered. The approach to dot size thermal object detection in atmosphere is proved with a help of threshold method in the thermodynamic and turbulent process conditions, based on the indistinct statement return task solution.

A review of occluded objects detection in real complex scenarios for autonomous driving

Autonomous driving is a promising way to future safe,efficient,and low-carbon transportation.Real-time ac-curate target detection is an essential precondition for the generation of proper following decision and control signals.However,considering the complex practical scenarios,accurate recognition of occluded targets is a major challenge of target detection for autonomous driving with limited computational capability.To reveal the overlap and difference between various occluded object detection by sharing the same available sensors,this paper presents a review of detection methods for occluded objects in complex real-driving scenarios.Considering the rapid development of autonomous driving technologies,the research analyzed in this study is limited to the recent five years.The study of occluded object detection is divided into three parts,namely occluded vehicles,pedes-trians and traffic signs.This paper provided a detailed summary of the target detection methods used in these three parts according to the differences in detection methods and ideas,which is followed by the comparison of advantages and disadvantages of different detection methods for the same object.Finally,the shortcomings and limitations of the existing detection methods are summarized,and the challenges and future development prospects in this field are discussed.

Improvement of High-Speed Detection Algorithm for Nonwoven Material Defects Based on Machine Vision

Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.

Automatic detection of small bowel lesions with different bleeding risks based on deep learning models

BACKGROUND Deep learning provides an efficient automatic image recognition method for small bowel(SB)capsule endoscopy(CE)that can assist physicians in diagnosis.However,the existing deep learning models present some unresolved challenges.AIM To propose a novel and effective classification and detection model to automatically identify various SB lesions and their bleeding risks,and label the lesions accurately so as to enhance the diagnostic efficiency of physicians and the ability to identify high-risk bleeding groups.METHODS The proposed model represents a two-stage method that combined image classification with object detection.First,we utilized the improved ResNet-50 classification model to classify endoscopic images into SB lesion images,normal SB mucosa images,and invalid images.Then,the improved YOLO-V5 detection model was utilized to detect the type of lesion and its risk of bleeding,and the location of the lesion was marked.We constructed training and testing sets and compared model-assisted reading with physician reading.RESULTS The accuracy of the model constructed in this study reached 98.96%,which was higher than the accuracy of other systems using only a single module.The sensitivity,specificity,and accuracy of the model-assisted reading detection of all images were 99.17%,99.92%,and 99.86%,which were significantly higher than those of the endoscopists’diagnoses.The image processing time of the model was 48 ms/image,and the image processing time of the physicians was 0.40±0.24 s/image(P<0.001).CONCLUSION The deep learning model of image classification combined with object detection exhibits a satisfactory diagnostic effect on a variety of SB lesions and their bleeding risks in CE images,which enhances the diagnostic efficiency of physicians and improves the ability of physicians to identify high-risk bleeding groups.

An Underwater Target Detection Algorithm Based on Attention Mechanism and Improved YOLOv7

For underwater robots in the process of performing target detection tasks,the color distortion and the uneven quality of underwater images lead to great difficulties in the feature extraction process of the model,which is prone to issues like error detection,omission detection,and poor accuracy.Therefore,this paper proposed the CER-YOLOv7(CBAM-EIOU-RepVGG-YOLOv7)underwater target detection algorithm.To improve the algorithm’s capability to retain valid features from both spatial and channel perspectives during the feature extraction phase,we have added a Convolutional Block Attention Module(CBAM)to the backbone network.The Reparameterization Visual Geometry Group(RepVGG)module is inserted into the backbone to improve the training and inference capabilities.The Efficient Intersection over Union(EIoU)loss is also used as the localization loss function,which reduces the error detection rate and missed detection rate of the algorithm.The experimental results of the CER-YOLOv7 algorithm on the UPRC(Underwater Robot Prototype Competition)dataset show that the mAP(mean Average Precision)score of the algorithm is 86.1%,which is a 2.2%improvement compared to the YOLOv7.The feasibility and validity of the CER-YOLOv7 are proved through ablation and comparison experiments,and it is more suitable for underwater target detection.

Enhancing Dense Small Object Detection in UAV Images Based on Hybrid Transformer

Transformer-based models have facilitated significant advances in object detection.However,their extensive computational consumption and suboptimal detection of dense small objects curtail their applicability in unmanned aerial vehicle(UAV)imagery.Addressing these limitations,we propose a hybrid transformer-based detector,H-DETR,and enhance it for dense small objects,leading to an accurate and efficient model.Firstly,we introduce a hybrid transformer encoder,which integrates a convolutional neural network-based cross-scale fusion module with the original encoder to handle multi-scale feature sequences more efficiently.Furthermore,we propose two novel strategies to enhance detection performance without incurring additional inference computation.Query filter is designed to cope with the dense clustering inherent in drone-captured images by counteracting similar queries with a training-aware non-maximum suppression.Adversarial denoising learning is a novel enhancement method inspired by adversarial learning,which improves the detection of numerous small targets by counteracting the effects of artificial spatial and semantic noise.Extensive experiments on the VisDrone and UAVDT datasets substantiate the effectiveness of our approach,achieving a significant improvement in accuracy with a reduction in computational complexity.Our method achieves 31.9%and 21.1%AP on the VisDrone and UAVDT datasets,respectively,and has a faster inference speed,making it a competitive model in UAV image object detection.

Human intrusion detection for high-speed railway perimeter under all-weather condition

Purpose – The paper aims to solve the problem of personnel intrusion identification within the limits of highspeed railways. It adopts the fusion method of millimeter wave radar and camera to improve the accuracy ofobject recognition in dark and harsh weather conditions.Design/methodology/approach – This paper adopts the fusion strategy of radar and camera linkage toachieve focus amplification of long-distance targets and solves the problem of low illumination by laser lightfilling of the focus point. In order to improve the recognition effect, this paper adopts the YOLOv8 algorithm formulti-scale target recognition. In addition, for the image distortion caused by bad weather, this paper proposesa linkage and tracking fusion strategy to output the correct alarm results.Findings – Simulated intrusion tests show that the proposed method can effectively detect human intrusionwithin 0–200 m during the day and night in sunny weather and can achieve more than 80% recognitionaccuracy for extreme severe weather conditions.Originality/value – (1) The authors propose a personnel intrusion monitoring scheme based on the fusion ofmillimeter wave radar and camera, achieving all-weather intrusion monitoring;(2) The authors propose a newmulti-level fusion algorithm based on linkage and tracking to achieve intrusion target monitoring underadverse weather conditions;(3) The authors have conducted a large number of innovative simulationexperiments to verify the effectiveness of the method proposed in this article.

A spatio-temporal multi-scale fusion algorithm for pine wood nematode disease tree detection

Pine wood nematode infection is a devastating disease.Unmanned aerial vehicle(UAV)remote sensing enables timely and precise monitoring.However,UAV aerial images are challenged by small target size and complex sur-face backgrounds which hinder their effectiveness in moni-toring.To address these challenges,based on the analysis and optimization of UAV remote sensing images,this study developed a spatio-temporal multi-scale fusion algorithm for disease detection.The multi-head,self-attention mechanism is incorporated to address the issue of excessive features generated by complex surface backgrounds in UAV images.This enables adaptive feature control to suppress redundant information and boost the model’s feature extraction capa-bilities.The SPD-Conv module was introduced to address the problem of loss of small target feature information dur-ing feature extraction,enhancing the preservation of key features.Additionally,the gather-and-distribute mechanism was implemented to augment the model’s multi-scale feature fusion capacity,preventing the loss of local details during fusion and enriching small target feature information.This study established a dataset of pine wood nematode disease in the Huangshan area using DJI(DJ-Innovations)UAVs.The results show that the accuracy of the proposed model with spatio-temporal multi-scale fusion reached 78.5%,6.6%higher than that of the benchmark model.Building upon the timeliness and flexibility of UAV remote sensing,the pro-posed model effectively addressed the challenges of detect-ing small and medium-size targets in complex backgrounds,thereby enhancing the detection efficiency for pine wood nematode disease.This facilitates early preemptive preser-vation of diseased trees,augments the overall monitoring proficiency of pine wood nematode diseases,and supplies technical aid for proficient monitoring.

Confusing Object Detection:A Survey

Confusing object detection(COD),such as glass,mirrors,and camouflaged objects,represents a burgeoning visual detection task centered on pinpointing and distinguishing concealed targets within intricate backgrounds,leveraging deep learning methodologies.Despite garnering increasing attention in computer vision,the focus of most existing works leans toward formulating task-specific solutions rather than delving into in-depth analyses of methodological structures.As of now,there is a notable absence of a comprehensive systematic review that focuses on recently proposed deep learning-based models for these specific tasks.To fill this gap,our study presents a pioneering review that covers both themodels and the publicly available benchmark datasets,while also identifying potential directions for future research in this field.The current dataset primarily focuses on single confusing object detection at the image level,with some studies extending to video-level data.We conduct an in-depth analysis of deep learning architectures,revealing that the current state-of-the-art(SOTA)COD methods demonstrate promising performance in single object detection.We also compile and provide detailed descriptions ofwidely used datasets relevant to these detection tasks.Our endeavor extends to discussing the limitations observed in current methodologies,alongside proposed solutions aimed at enhancing detection accuracy.Additionally,we deliberate on relevant applications and outline future research trajectories,aiming to catalyze advancements in the field of glass,mirror,and camouflaged object detection.

YOLO-MFD:Remote Sensing Image Object Detection with Multi-Scale Fusion Dynamic Head

Remote sensing imagery,due to its high altitude,presents inherent challenges characterized by multiple scales,limited target areas,and intricate backgrounds.These inherent traits often lead to increased miss and false detection rates when applying object recognition algorithms tailored for remote sensing imagery.Additionally,these complexities contribute to inaccuracies in target localization and hinder precise target categorization.This paper addresses these challenges by proposing a solution:The YOLO-MFD model(YOLO-MFD:Remote Sensing Image Object Detection withMulti-scale Fusion Dynamic Head).Before presenting our method,we delve into the prevalent issues faced in remote sensing imagery analysis.Specifically,we emphasize the struggles of existing object recognition algorithms in comprehensively capturing critical image features amidst varying scales and complex backgrounds.To resolve these issues,we introduce a novel approach.First,we propose the implementation of a lightweight multi-scale module called CEF.This module significantly improves the model’s ability to comprehensively capture important image features by merging multi-scale feature information.It effectively addresses the issues of missed detection and mistaken alarms that are common in remote sensing imagery.Second,an additional layer of small target detection heads is added,and a residual link is established with the higher-level feature extraction module in the backbone section.This allows the model to incorporate shallower information,significantly improving the accuracy of target localization in remotely sensed images.Finally,a dynamic head attentionmechanism is introduced.This allows themodel to exhibit greater flexibility and accuracy in recognizing shapes and targets of different sizes.Consequently,the precision of object detection is significantly improved.The trial results show that the YOLO-MFD model shows improvements of 6.3%,3.5%,and 2.5%over the original YOLOv8 model in Precision,map@0.5 and map@0.5:0.95,separately.These results illustrate the clear advantages of the method.

A Novel Tensor Decomposition-Based Efficient Detector for Low-Altitude Aerial Objects With Knowledge Distillation Scheme

Unmanned aerial vehicles(UAVs) have gained significant attention in practical applications, especially the low-altitude aerial(LAA) object detection imposes stringent requirements on recognition accuracy and computational resources. In this paper, the LAA images-oriented tensor decomposition and knowledge distillation-based network(TDKD-Net) is proposed,where the TT-format TD(tensor decomposition) and equalweighted response-based KD(knowledge distillation) methods are designed to minimize redundant parameters while ensuring comparable performance. Moreover, some robust network structures are developed, including the small object detection head and the dual-domain attention mechanism, which enable the model to leverage the learned knowledge from small-scale targets and selectively focus on salient features. Considering the imbalance of bounding box regression samples and the inaccuracy of regression geometric factors, the focal and efficient IoU(intersection of union) loss with optimal transport assignment(F-EIoU-OTA)mechanism is proposed to improve the detection accuracy. The proposed TDKD-Net is comprehensively evaluated through extensive experiments, and the results have demonstrated the effectiveness and superiority of the developed methods in comparison to other advanced detection algorithms, which also present high generalization and strong robustness. As a resource-efficient precise network, the complex detection of small and occluded LAA objects is also well addressed by TDKD-Net, which provides useful insights on handling imbalanced issues and realizing domain adaptation.

Enhanced Object Detection and Classification via Multi-Method Fusion

Advances in machine vision systems have revolutionized applications such as autonomous driving,robotic navigation,and augmented reality.Despite substantial progress,challenges persist,including dynamic backgrounds,occlusion,and limited labeled data.To address these challenges,we introduce a comprehensive methodology toenhance image classification and object detection accuracy.The proposed approach involves the integration ofmultiple methods in a complementary way.The process commences with the application of Gaussian filters tomitigate the impact of noise interference.These images are then processed for segmentation using Fuzzy C-Meanssegmentation in parallel with saliency mapping techniques to find the most prominent regions.The Binary RobustIndependent Elementary Features(BRIEF)characteristics are then extracted fromdata derived fromsaliency mapsand segmented images.For precise object separation,Oriented FAST and Rotated BRIEF(ORB)algorithms areemployed.Genetic Algorithms(GAs)are used to optimize Random Forest classifier parameters which lead toimproved performance.Our method stands out due to its comprehensive approach,adeptly addressing challengessuch as changing backdrops,occlusion,and limited labeled data concurrently.A significant enhancement hasbeen achieved by integrating Genetic Algorithms(GAs)to precisely optimize parameters.This minor adjustmentnot only boosts the uniqueness of our system but also amplifies its overall efficacy.The proposed methodologyhas demonstrated notable classification accuracies of 90.9%and 89.0%on the challenging Corel-1k and MSRCdatasets,respectively.Furthermore,detection accuracies of 87.2%and 86.6%have been attained.Although ourmethod performed well in both datasets it may face difficulties in real-world data especially where datasets havehighly complex backgrounds.Despite these limitations,GAintegration for parameter optimization shows a notablestrength in enhancing the overall adaptability and performance of our system.

Robust and Discriminative Feature Learning via Mutual Information Maximization for Object Detection in Aerial Images

Object detection in unmanned aerial vehicle(UAV)aerial images has become increasingly important in military and civil applications.General object detection models are not robust enough against interclass similarity and intraclass variability of small objects,and UAV-specific nuisances such as uncontrolledweather conditions.Unlike previous approaches focusing on high-level semantic information,we report the importance of underlying features to improve detection accuracy and robustness fromthe information-theoretic perspective.Specifically,we propose a robust and discriminative feature learning approach through mutual information maximization(RD-MIM),which can be integrated into numerous object detection methods for aerial images.Firstly,we present the rank sample mining method to reduce underlying feature differences between the natural image domain and the aerial image domain.Then,we design a momentum contrast learning strategy to make object features similar to the same category and dissimilar to different categories.Finally,we construct a transformer-based global attention mechanism to boost object location semantics by leveraging the high interrelation of different receptive fields.We conduct extensive experiments on the VisDrone and Unmanned Aerial Vehicle Benchmark Object Detection and Tracking(UAVDT)datasets to prove the effectiveness of the proposed method.The experimental results show that our approach brings considerable robustness gains to basic detectors and advanced detection methods,achieving relative growth rates of 51.0%and 39.4%in corruption robustness,respectively.Our code is available at https://github.com/cq100/RD-MIM(accessed on 2 August 2024).

Research on fast detection method of infrared small targets under resourceconstrained conditions

Infrared small target detection is a common task in infrared image processing.Under limited computa⁃tional resources.Traditional methods for infrared small target detection face a trade-off between the detection rate and the accuracy.A fast infrared small target detection method tailored for resource-constrained conditions is pro⁃posed for the YOLOv5s model.This method introduces an additional small target detection head and replaces the original Intersection over Union(IoU)metric with Normalized Wasserstein Distance(NWD),while considering both the detection accuracy and the detection speed of infrared small targets.Experimental results demonstrate that the proposed algorithm achieves a maximum effective detection speed of 95 FPS on a 15 W TPU,while reach⁃ing a maximum effective detection accuracy of 91.9 AP@0.5,effectively improving the efficiency of infrared small target detection under resource-constrained conditions.

Probability-Enhanced Anchor-Free Detector for Remote-Sensing Object Detection

Anchor-free object-detection methods achieve a significant advancement in field of computer vision,particularly in the realm of real-time inferences.However,in remote sensing object detection,anchor-free methods often lack of capability in separating the foreground and background.This paper proposes an anchor-free method named probability-enhanced anchor-free detector(ProEnDet)for remote sensing object detection.First,a weighted bidirectional feature pyramid is used for feature extraction.Second,we introduce probability enhancement to strengthen the classification of the object’s foreground and background.The detector uses the logarithm likelihood as the final score to improve the classification of the foreground and background of the object.ProEnDet is verified using the DIOR and NWPU-VHR-10 datasets.The experiment achieved mean average precisions of 61.4 and 69.0 on the DIOR dataset and NWPU-VHR-10 dataset,respectively.ProEnDet achieves a speed of 32.4 FPS on the DIOR dataset,which satisfies the real-time requirements for remote-sensing object detection.

HWD-YOLO:A New Vision-Based Helmet Wearing Detection Method

It is crucial to ensure workers wear safety helmets when working at a workplace with a high risk of safety accidents,such as construction sites and mine tunnels.Although existing methods can achieve helmet detection in images,their accuracy and speed still need improvements since complex,cluttered,and large-scale scenes of real workplaces cause server occlusion,illumination change,scale variation,and perspective distortion.So,a new safety helmet-wearing detection method based on deep learning is proposed.Firstly,a new multi-scale contextual aggregation module is proposed to aggregate multi-scale feature information globally and highlight the details of concerned objects in the backbone part of the deep neural network.Secondly,a new detection block combining the dilate convolution and attention mechanism is proposed and introduced into the prediction part.This block can effectively extract deep featureswhile retaining information on fine-grained details,such as edges and small objects.Moreover,some newly emerged modules are incorporated into the proposed network to improve safety helmetwearing detection performance further.Extensive experiments on open dataset validate the proposed method.It reaches better performance on helmet-wearing detection and even outperforms the state-of-the-art method.To be more specific,the mAP increases by 3.4%,and the speed increases from17 to 33 fps in comparison with the baseline,You Only Look Once(YOLO)version 5X,and themean average precision increases by 1.0%and the speed increases by 7 fps in comparison with the YOLO version 7.The generalization ability and portability experiment results show that the proposed improvements could serve as a springboard for deep neural network design to improve object detection performance in complex scenarios.

Rail-Pillar Net:A 3D Detection Network for Railway Foreign Object Based on LiDAR

Aiming at the limitations of the existing railway foreign object detection methods based on two-dimensional(2D)images,such as short detection distance,strong influence of environment and lack of distance information,we propose Rail-PillarNet,a three-dimensional(3D)LIDAR(Light Detection and Ranging)railway foreign object detection method based on the improvement of PointPillars.Firstly,the parallel attention pillar encoder(PAPE)is designed to fully extract the features of the pillars and alleviate the problem of local fine-grained information loss in PointPillars pillars encoder.Secondly,a fine backbone network is designed to improve the feature extraction capability of the network by combining the coding characteristics of LIDAR point cloud feature and residual structure.Finally,the initial weight parameters of the model were optimised by the transfer learning training method to further improve accuracy.The experimental results on the OSDaR23 dataset show that the average accuracy of Rail-PillarNet reaches 58.51%,which is higher than most mainstream models,and the number of parameters is 5.49 M.Compared with PointPillars,the accuracy of each target is improved by 10.94%,3.53%,16.96%and 19.90%,respectively,and the number of parameters only increases by 0.64M,which achieves a balance between the number of parameters and accuracy.