Target Detection Algorithm in Foggy Scenes Based on Dual Subnets

Under the influence of air humidity,dust,aerosols,etc.,in real scenes,haze presents an uneven state.In this way,the image quality and contrast will decrease.In this case,It is difficult to detect the target in the image by the universal detection network.Thus,a dual subnet based on multi-task collaborative training(DSMCT)is proposed in this paper.Firstly,in the training phase,the Gated Context Aggregation Network(GCANet)is used as the supervisory network of YOLOX to promote the extraction of clean information in foggy scenes.In the test phase,only the YOLOX branch needs to be activated to ensure the detection speed of the model.Secondly,the deformable convolution module is used to improve GCANet to enhance the model’s ability to capture details of non-homogeneous fog.Finally,the Coordinate Attention mechanism is introduced into the Vision Transformer and the backbone network of YOLOX is redesigned.In this way,the feature extraction ability of the network for deep-level information can be enhanced.The experimental results on artificial fog data set FOG_VOC and real fog data set RTTS show that the map value of DSMCT reached 86.56%and 62.39%,respectively,which was 2.27%and 4.41%higher than the current most advanced detection model.The DSMCT network has high practicality and effectiveness for target detection in real foggy scenes.

A Normalizing Flow-Based Bidirectional Mapping Residual Network for Unsupervised Defect Detection

Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately evaluate sample distributions,mapping normal features to the normal distribution and anomalous features outside it.Consequently,this paper proposes a Normalizing Flow-based Bidirectional Mapping Residual Network(NF-BMR).It utilizes pre-trained Convolutional Neural Networks(CNN)and normalizing flows to construct discriminative source and target domain feature spaces.Additionally,to better learn feature information in both domain spaces,we propose the Bidirectional Mapping Residual Network(BMR),which maps sample features to these two spaces for anomaly detection.The two detection spaces effectively complement each other’s deficiencies and provide a comprehensive feature evaluation from two perspectives,which leads to the improvement of detection performance.Comparative experimental results on the MVTec AD and DAGM datasets against the Bidirectional Pre-trained Feature Mapping Network(B-PFM)and other state-of-the-art methods demonstrate that the proposed approach achieves superior performance.On the MVTec AD dataset,NF-BMR achieves an average AUROC of 98.7%for all 15 categories.Especially,it achieves 100%optimal detection performance in five categories.On the DAGM dataset,the average AUROC across ten categories is 98.7%,which is very close to supervised methods.

A CNN-Based Single-Stage Occlusion Real-Time Target Detection Method

Aiming at the problem of low accuracy of traditional target detection methods for target detection in endoscopes in substation environments, a CNN-based real-time detection method for masked targets is proposed. The method adopts the overall design of backbone network, detection network and algorithmic parameter optimisation method, completes the model training on the self-constructed occlusion target dataset, and adopts the multi-scale perception method for target detection. The HNM algorithm is used to screen positive and negative samples during the training process, and the NMS algorithm is used to post-process the prediction results during the detection process to improve the detection efficiency. After experimental validation, the obtained model has the multi-class average predicted value (mAP) of the dataset. It has general advantages over traditional target detection methods. The detection time of a single target on FDDB dataset is 39 ms, which can meet the need of real-time target detection. In addition, the project team has successfully deployed the method into substations and put it into use in many places in Beijing, which is important for achieving the anomaly of occlusion target detection.

Intelligent Passive Detection of Aerial Target in Space-Air-Ground Integrated Networks

Passive detection of moving target is an important part of intelligent surveillance. Satellite has the potential to play a key role in many applications of space-air-ground integrated networks(SAGIN). In this paper, we propose a novel intelligent passive detection method for aerial target based on reservoir computing networks. Specifically, delayed feedback networks are utilized to refine the direct signals from the satellite in the reference channels. In addition, the satellite direct wave interference in the monitoring channels adopts adaptive interference suppression using the minimum mean square error filter. Furthermore, we employ decoupling echo state networks to predict the clutter interference in the monitoring channels and construct the detection statistics accordingly. Finally, a multilayer perceptron is adopted to detect the echo signal after interference suppression. Extensive simulations is conducted to evaluate the performance of our proposed method. Results show that the detection probability is almost 100% when the signal-to-interference ratio of echo signal is-36 dB, which demonstrates that our proposed method achieves efficient passive detection for aerial targets in typical SAGIN scenarios.

Deep convolutional neural network for meteorology target detection in airborne weather radar images

Considering the problem that the scattering echo images of airborne Doppler weather radar are often reduced by ground clutters,the accuracy and confidence of meteorology target detection are reduced.In this paper,a deep convolutional neural network(DCNN)is proposed for meteorology target detection and ground clutter suppression with a large collection of airborne weather radar images as network input.For each weather radar image,the corresponding digital elevation model(DEM)image is extracted on basis of the radar antenna scan-ning parameters and plane position,and is further fed to the net-work as a supplement for ground clutter suppression.The fea-tures of actual meteorology targets are learned in each bottle-neck module of the proposed network and convolved into deeper iterations in the forward propagation process.Then the network parameters are updated by the back propagation itera-tion of the training error.Experimental results on the real mea-sured images show that our proposed DCNN outperforms the counterparts in terms of six evaluation factors.Meanwhile,the network outputs are in good agreement with the expected mete-orology detection results(labels).It is demonstrated that the pro-posed network would have a promising meteorology observa-tion application with minimal effort on network variables or parameter changes.

Target Detection in Three-Dimension Sensor Networks Based on Clifford Algebra

The three-dimensional sensor networks are supposed to be deployed for many applications. So it is signifi-cant to do research on the problems of coverage and target detection in three-dimensional sensor networks. In this paper, we introduced Clifford algebra in 3D Euclidean space, developed the coverage model of 3D sensor networks based on Clifford algebra, and proposed a method for detecting target moving. With Clif-ford Spinor, calculating the target moving formulation is easier than traditional methods in sensor node’s coverage area.

Sonar Image Target Detection for Underwater Communication System Based on Deep Neural Network

Target signal acquisition and detection based on sonar images is a challenging task due to the complex underwater environment.In order to solve the problem that some semantic information in sonar images is lost and model detection performance is degraded due to the complex imaging environment,we proposed a more effective and robust target detection framework based on deep learning,which can make full use of the acoustic shadow information in the forward-looking sonar images to assist underwater target detection.Firstly,the weighted box fusion method is adopted to generate a fusion box by weighted fusion of prediction boxes with high confidence,so as to obtain accurate acoustic shadow boxes.Further,the acoustic shadow box is cut down to get the feature map containing the acoustic shadow information,and then the acoustic shadow feature map and the target information feature map are adaptively fused to make full use of the acoustic shadow feature information.In addition,we introduce a threshold processing module to improve the attention of the model to important feature information.Through the underwater sonar dataset provided by Pengcheng Laboratory,the proposed method improved the average accuracy by 3.14%at the IoU threshold of 0.7,which is better than the current traditional target detection model.

An infrared target intrusion detection method based on feature fusion and enhancement

Infrared target intrusion detection has significant applications in the fields of military defence and intelligent warning.In view of the characteristics of intrusion targets as well as inspection difficulties,an infrared target intrusion detection algorithm based on feature fusion and enhancement was proposed.This algorithm combines static target mode analysis and dynamic multi-frame correlation detection to extract infrared target features at different levels.Among them,LBP texture analysis can be used to effectively identify the posterior feature patterns which have been contained in the target library,while motion frame difference method can detect the moving regions of the image,improve the integrity of target regions such as camouflage,sheltering and deformation.In order to integrate the advantages of the two methods,the enhanced convolutional neural network was designed and the feature images obtained by the two methods were fused and enhanced.The enhancement module of the network strengthened and screened the targets,and realized the background suppression of infrared images.Based on the experiments,the effect of the proposed method and the comparison method on the background suppression and detection performance was evaluated,and the results showed that the SCRG and BSF values of the method in this paper had a better performance in multiple data sets,and it’s detection performance was far better than the comparison algorithm.The experiment results indicated that,compared with traditional infrared target detection methods,the proposed method could detect the infrared invasion target more accurately,and suppress the background noise more effectively.

Arbitrary-oriented target detection in large scene sar images

Target detection in the field of synthetic aperture radar(SAR) has attracted considerable attention of researchers in national defense technology worldwide,owing to its unique advantages like high resolution and large scene image acquisition capabilities of SAR.However,due to strong speckle noise and low signal-to-noise ratio,it is difficult to extract representative features of target from SAR images,which greatly inhibits the effectiveness of traditional methods.In order to address the above problems,a framework called contextual rotation region-based convolutional neural network(RCNN) with multilayer fusion is proposed in this paper.Specifically,aimed to enable RCNN to perform target detection in large scene SAR images efficiently,maximum sliding strategy is applied to crop the large scene image into a series of sub-images before RCNN.Instead of using the highest-layer output for proposal generation and target detection,fusion feature maps with high resolution and rich semantic information are constructed by multilayer fusion strategy.Then,we put forwards rotation anchors to predict the minimum circumscribed rectangle of targets to reduce redundant detection region.Furthermore,shadow areas serve as contextual features to provide extraneous information for the detector identify and locate targets accurately.Experimental results on the simulated large scene SAR image dataset show that the proposed method achieves a satisfactory performance in large scene SAR target detection.

M-FCN based sea-surface weak target detection

This paper focuses on the sea-surface weak target detection based on memory-fully convolutional network(M-FCN)in strong sea clutter.Firstly,the constant false alarm rate(CFAR)detection method utilizes a low threshold with high probability of false alarm to detect sea-surface weak targets after non-coherent integration.Reducing the detection threshold can generate a large number of false alarms while increasing the detection rate,and how to suppress a large number of false alarms is the key to improve the performance of weak target detection.Then,the detection result of the low threshold is operated to construct the target matrix suitable for the size of fully convolutional networks and the convolution operator form.Finally,the M-FCN architecture is designed to learn the different accumulation characteristics of the target and the sea clutter between different frames.For improving the detection performance,the historical multi-frame information is memorized by the network,and the end-to-end structure is established to detect sea-surface weak target automatically.Experimental results on measured data demonstrate that the M-FCN method outperforms the traditional track before detection(TBD)method and reduces false alarm tracks by 35.1%,which greatly improves the track quality.

LF-CNN:Deep Learning-Guided Small Sample Target Detection for Remote Sensing Classification

Target detection of small samples with a complex background is always difficult in the classification of remote sensing images.We propose a new small sample target detection method combining local features and a convolutional neural network(LF-CNN)with the aim of detecting small numbers of unevenly distributed ground object targets in remote sensing images.The k-nearest neighbor method is used to construct the local neighborhood of each point and the local neighborhoods of the features are extracted one by one from the convolution layer.All the local features are aggregated by maximum pooling to obtain global feature representation.The classification probability of each category is then calculated and classified using the scaled expected linear units function and the full connection layer.The experimental results show that the proposed LF-CNN method has a high accuracy of target detection and classification for hyperspectral imager remote sensing data under the condition of small samples.Despite drawbacks in both time and complexity,the proposed LF-CNN method can more effectively integrate the local features of ground object samples and improve the accuracy of target identification and detection in small samples of remote sensing images than traditional target detection methods.

A Novel SAR Image Ship Small Targets Detection Method

To satisfy practical requirements of high real-time accuracy and low computational complexity of synthetic aperture radar (SAR) image ship small target detection, this paper proposes a small ship target detection method based on the improved You Only Look Once Version 3 (YOLOv3). The main contributions of this study are threefold. First, the feature extraction network of the original YOLOV3 algorithm is replaced with the VGG16 network convolution layer. Second, general convolution is transformed into depthwise separable convolution, thereby reducing the computational cost of the algorithm. Third, a residual network structure is introduced into the feature extraction network to reuse the shallow target feature information, which enhances the detailed features of the target and ensures the improvement in accuracy of small target detection performance. To evaluate the performance of the proposed method, many experiments are conducted on public SAR image datasets. For ship targets with complex backgrounds and small ship targets in the SAR image, the effectiveness of the proposed algorithm is verified. Results show that the accuracy and recall rate improved by 5.31% and 2.77%, respectively, compared with the original YOLOV3. Furthermore, the proposed model not only significantly reduces the computational effort, but also improves the detection accuracy of ship small target.

多尺度特征和极化自注意力的Faster-RCNN水漂垃圾识别

针对小目标水漂垃圾形态多变、分辨率低且信息有限,导致检测效果不理想的问题,提出一种改进的Faster-RCNN(Faster Regions with Convolutional Neural Network)水漂垃圾检测算法MP-Faster-RCNN(Faster-RCNN with Multi-scale feature and Polarized self-attention)。首先,建立黄河兰州段小目标水漂垃圾数据集,将空洞卷积结合ResNet-50代替原来的VGG-16(Visual Geometry Group 16)作为主干特征提取网络,扩大感受野以提取更多小目标特征;其次,在区域生成网络(RPN)利用多尺度特征,设置3×3和1×1的两层卷积,补偿单一滑动窗口造成的特征丢失;最后,在RPN前加入极化自注意力,进一步利用多尺度和通道特征提取更细粒度的多尺度空间信息和通道间依赖关系,生成具有全局特征的特征图,实现更精确的目标框定位。实验结果表明,MP-Faster-RCNN能有效提高水漂垃圾检测精度,与原始Faster-RCNN相比,平均精度均值(mAP)提高了6.37个百分点,模型大小从521 MB降到了108 MB,且在同一训练批次下收敛更快。

Using deep learning to detect small targets in infrared oversampling images

According to the oversampling imaging characteristics, an infrared small target detection method based on deep learning is proposed. A 7-layer deep convolutional neural network(CNN) is designed to automatically extract small target features and suppress clutters in an end-to-end manner. The input of CNN is an original oversampling image while the output is a cluttersuppressed feature map. The CNN contains only convolution and non-linear operations, and the resolution of the output feature map is the same as that of the input image. The L1-norm loss function is used, and a mass of training data is generated to train the network effectively. Results show that compared with several baseline methods, the proposed method improves the signal clutter ratio gain and background suppression factor by 3–4 orders of magnitude, and has more powerful target detection performance.

MOVING TARGETS PATTERN RECOGNITION BASED ON THE WAVELET NEURAL NETWORK

Based on pattern recognition theory and neural network technology, moving objects automatic detection and classification method integrating advanced wavelet analysis are discussed in detail. An algorithm of moving targets pattern recognition on the combination of inter-frame difference and wavelet neural network is presented. The experimental results indicate that the designed BP wavelet network using this algorithm can recognize and classify moving targets rapidly and effectively.

Fusion network for small target detection based on YOLO and attention mechanism

Target detection is an important task in computer vision research, and such an anomaly detection and the topic of small target detection task is more concerned. However, there are still some problems in this kind of researches, such as small target detection in complex environments is susceptible to background interference and poor detection results. To solve these issues, this study proposes a method which introduces the attention mechanism into the you only look once(YOLO) network. In addition, the amateur-produced mask dataset was created and experiments were conducted. The results showed that the detection effect of the proposed mothed is much better.

Mask Text Detector:一种检测自然场景下任意形状的文本分割网络

近年来场景文本检测技术飞速发展,提出一种可适用于任意形状文本检测的新颖算法Mask Text Detector.该算法在Mask R-CNN的基础上,用anchor-free的方法替代了原本的RPN层生成建议框,减少了超参、模型参数和计算量.还提出LQCS(Localization Quality and Classification Score)joint regression,能够将坐标质量和类别分数关联到一起,消除预测阶段不一致的问题.为了让网络区分复杂样本,结合传统的边缘检测算法提出Socle-Mask分支生成分割掩码.该模块在水平和垂直方向上分区别提取纹理特征,并加入通道自注意力机制,让网络自主选择通道特征.我们在三个具有挑战性的数据集(Total-Text、CTW1500和ICDAR2015)中进行了广泛的实验,验证了该算法具有很好的文本检测性能.

基于改进RPN的Faster-RCNN网络SAR图像车辆目标检测方法

针对传统Faster-RCNN方法中候选区域生成网络(RPN)模块在进行目标检测时对目标特征提取不够充分的问题,提出一种基于改进RPN的Faster-RCNN网络SAR图像车辆目标检测方法.首先基于VGG-16网络提取出图片的多层特征,然后利用卷积核对最深的3个特征层作进一步的特征提取和正则化处理,最后对处理后的3个特征层进行信息融合.利用MSTAR数据集中车辆目标SAR图像和自然背景SAR图像,通过图像分割和贴图的方式制作了SAR场景数据集,对所改进网络进行训练和测试.实验结果表明,在SAR图像车辆目标检测中,与传统RPN相比,改进RPN收敛速度更快,不仅将检测结果的查准率从97.7%提高到了99.7%,虚警率明显降低,而且泛化性能更强,针对训练范围以外的目标,能将查准率由98.0%提高到99.0%.

基于Faster-RCNN的结核杆菌自动检测方法研究与应用

染色处理可使结核杆菌在显微镜拍摄的医学图像中显现,医生通过检测图像中的结核杆菌辅助诊断结核病。近年来卷积神经网络(CNN)在目标检测上取得了突破性进展,但结核杆菌存在图像上尺度小,构造标注数据难,不适用迁移学习等问题,使得基于CNN的目标检测方法在结核杆菌检测方面尚存在一定的困难。为此,以Faster-RCNN目标检测算法为基础,研究在医学图像上的结核杆菌检测问题。针对结核杆菌尺度小,提出重叠子图划分策略;针对标注数据构造难,提出分块、迭代标注策略。实践证明,该方法有较高的准确度以及可接受的速度,已构建了13 261个结核杆菌的训练数据,应用于合作单位的医疗检测产品,能满足实际应用需求。

Morphological neural networks for automatic target detection by simulated annealing learning algorithm

A practical neural network model for morphological filtering and a simulated annealing optimal algorithm for the network parameters training are proposed in this paper. It is pointed out that the opti- mal designing process of the morphological filtering network in fact is the optimal learning process of adjusting network parameters (structuring element, or SE for short) to accommodate image environment. Then the network structure may possess the characteristics of image targets, and so give specific infor- mation to the SE. Morphological filters formed in this way become certainly intelligent and can provide good filtering results and robust adaptability to complex changing image. For application to motional image target detection, dynamic training algorithm is applied to the designing process using asymptotic shrinking error and appropriate network weights adjusting. Experimental results show that the algorithm has invariant property with respect to shift, scale and rotation of moving target in continuing detection of moving targets.