检索结果-内蒙古大学图书馆

AFSar: An Anchor-Free SAR Target detection Algorithm Based on Multiscale Enhancement Representation Learning

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2022年 60卷 1页

作者： Wan, Huiyao Chen, Jie Huang, Zhixiang Xia, RunFan Wu, BoCai Sun, Long Yao, Baidong Liu, Xiaoping Xing, Mengdao Anhui Univ Sch Elect & Informat Engn Minist Educ Key Lab Intelligent Comp & Signal Proc Hefei 230601 Peoples R China China Elect Technol Grp Corp Res Inst 38 Hefei 230088 Peoples R China Xidian Univ Natl Lab Radar Signal Proc Xian 710071 Peoples R China Xidian Univ Collaborat Innovat Ctr Informat Sensing & Underst Xian 710071 Peoples R China

Unlike optical images, synthetic aperture radar (SAR) images have unique characteristics, such as few samples, strong scattering, sparseness, multiple scales, complex interference and background, and inconspicuous target edge contour information. Current SAR target detection algorithms have difficulty in balancing accuracy and speed, and the performance of these algorithms is relatively limited, thus making it difficult to deploy practical applications. To this end, this article proposes AFSar, an innovative anchor-free SAR target detection algorithm based on multiscale enhancement representation learning. First, we introduce the latest anchor-free architecture YOLOX as the basic framework. Second, to reduce the computational complexity of the model and to improve the ability of multiscale feature extraction, we redesigned the lightweight backbone, namely, MobileNetV2S. Furthermore, we propose an attention enhancement PAN module, called CSEMPAN, which highlights the unique strong scattering characteristics of SAR targets by integrating channel and spatial attention mechanisms. Finally, in view of the multiscale and strong sparse characteristics of SAR targets, we propose a new target detection head, namely, ESPHead. ESPHead extracts the features of targets with different scales by using dilated convolution with different dilated rates, so as to enhance the detection ability of the model for targets with different scales. The results of ablation experiments on the SSDD dataset show that the mAP of our algorithm reaches 0.977, while the Flops is only 9.86 G, achieving state of the art.

关键词： Feature extraction Object detection Radar polarimetry Synthetic aperture radar detection algorithms Optical imaging Prediction algorithms Anchor-free deep learning enhancement representation learning multiscale learning synthetic aperture radar (SAR) target detection

来源：评论

学校读者我要写书评

暂无评论

Change detection Based on Convolutional Neural Networks Using Stacks of Wavelength-Resolution Synthetic Aperture Radar Images

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2022年 60卷 1页

作者： Vinholi, Joao G. Palm, Bruna G. Silva, Danilo Machado, Renato Pettersson, Mats, I Aeronaut Inst Technol ITA Dept Telecommun BR-12228900 Sao Jose Dos Campos Brazil Blekinge Inst Technol BTH Dept Math & Nat Sci S-37179 Karlskrona Sweden Fed Univ Santa Catarina UFSC Dept Elect & Elect Engn BR-88040900 Florianopolis SC Brazil

This article presents two supervised change detection algorithms (CDAs) based on convolutional neural networks (CNNs) that use stacks of coregistered wavelength-resolution (WR) synthetic aperture radar (SAR) images to detect changes in an image under monitoring. The additional information of a scene of interest provided by SAR image stacks can be explored to enhance the performance of CDAs. In particular, stacks of images with similar statistics can be obtained for ultrawideband (UWB) very high-frequency (VHF) SAR systems, as they produce images highly stable in time. The proposed CDAs can be summed up into four stages: difference image (DI) formation, semantic segmentation, clustering, and change classification. The CNN-GSP algorithm is based on a ground scene prediction (GSP) image, which is used as a reference to form a DI. A CNN-based model then analyzes the DI. The CNN-MDI algorithm feeds multiple DIs with identical monitored images to a CNN-based model, which will concurrently analyze their features. Tests with CARABAS-II data show that the proposed CDAs can outperform other state-of-the-art algorithms that also use stacks of WR-SAR images. Beyond that, the proposed algorithms outperformed a CNN-based CDA that does not use image stacks, which shows that CNN-based algorithms can use the additional information provided by stacks of SAR images to reduce false alarm occurrences while increasing the probability of detection of changes.

关键词： Synthetic aperture radar Radar polarimetry Convolutional neural networks Monitoring detection algorithms Complexity theory Ultra wideband technology CARABAS-II classification convolutional neural network (CNN) deep learning remote sensing semantic segmentation

来源：评论

学校读者我要写书评

暂无评论

An Improved Support Vector Machine Attack detection Algorithm for Industry Controls System

An Improved Support Vector Machine Attack Detection Algorith...

引用

Power Electronics, Computer Applications (ICPECA), IEEE International Conference on

作者： Shangqing Cui Dongqin Feng College of Control Science and Engineering Zhejiang University Hangzhou China

With the application of Internet technologies in the industrial field, industrial control systems (ICS) have the possibility of suffering malicious network attacks. Because ICS is widely used in critical infrastructure, designing attack detection algorithms for it can effectively reduce losses. Many papers have built physical models for cyber-attacks against ICS and the corresponding detection algorithms have been proposed. However, the detection performance can be further improved. This paper establishes the simulation attack on the Tennessee Eastman (TE) process and proposes an improved support vector machine (SVM) method to detect integrity attacks. This algorithm uses contribution plots and recursive feature elimination to select features. Compared with the principal component analysis (PCA) feature extraction, the detection accuracy is improved. After adding system dynamics to the algorithm, the time to detect attacks is advanced.

关键词： Support vector machines System dynamics Industrial control Heuristic algorithms Feature extraction detection algorithms Integrated circuit modeling

来源：评论

学校读者我要写书评

暂无评论

Change Point detection with Neural Online Density-Ratio Estimator

Change Point Detection with Neural Online Density-Ratio Esti...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Xiuheng Wang Ricardo Augusto Borsoi Cédric Richard Jie Chen Université Côte d’Azur CNRS OCA Nice France Université de Lorraine CNRS CRAN Vandoeuvre-lès-Nancy France School of Marine Science and Technology Northwestern Polytechnical University Xi’an China

Detecting change points in streaming time series data is a long standing problem in signal processing. A plethora of methods have been proposed to address it, depending on the hypotheses at hand. Non-parametric approaches are particularly interesting as they do not make any assumption on the distribution of data or on the nature of changes. Nevertheless, leveraging recent advances in deep learning to detect change points in time series data is still challenging. In this paper, we propose a change point detection method using an online approach based on neural networks to directly estimate the density-ratio between current and reference windows of the data stream. A variational continual learning framework is employed to train the neural network in an online manner while retaining information learned from past data. This leads to a statistically-principled fully nonparametric framework to detect change points from streaming data. Experimental results with synthetic and real data illustrate the effectiveness of the proposed approach.

关键词： Deep learning Neural networks Time series analysis Signal processing Acoustics Speech processing detection algorithms

来源：评论

学校读者我要写书评

暂无评论

A Density-based Clustering Approach to detect Colluding SSDF Attackers in Cognitive Radio Networks 37

A Density-based Clustering Approach to detect Colluding SSDF...

引用

37th International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC)

作者： Taggu, Amar Marchang, Ningrinla North Eastern Reg Inst Sci & Technol NERIST Dept Comp Sci & Engn Nirjuli India

ISBN: (数字)9781665485593

ISBN: (纸本)9781665485593

Among the many attacks that Cognitive Radio Networks (CRN) are vulnerable to, Spectrum Sensing Data Falsification (SSDF) attack is one of the most common attacks. The current work focusses on an advanced form of SSDF attacks, called Colluding SSDF attack. The colluding attack is a collaborative well-planned and well-timed attack where a leader instructs the other attackers when and how to attack a CRN during sensing period. Needless to say, Colluding SSDF attacks are harder to detect and thus, they are more harmful. To detect this attack, we have proposed a technique based on Density-Based Spatial Clustering of Applications with Noise (DBSCAN) [1]. The density-based clustering technique separates the high-density clusters (normal SUs) from low-density clusters (attackers). Simulation results demonstrate the effectiveness of the proposed when the percentage of attackers is less than 50%. Additionally, a performance comparison of the proposed method has been done with an existing work, agglomerative clustering detection algorithm (ACD)[2].

关键词： Computers Simulation Collaboration Clustering algorithms Sensors Cognitive radio detection algorithms

来源：评论

学校读者我要写书评

暂无评论

Research on Automatic detection Algorithm for Lumbar Disc Herniation Based on FCOS

Research on Automatic Detection Algorithm for Lumbar Disc He...

引用

Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI), IEEE International Conference on

作者： Tingqiang Guan Ziming Wang Wei Hu School of Computer Science and Information Security Guilin University of Electronic Technology Guilin China Department of Spine and Osteopathy Ward Liuzhou Peoples' Hospital Guilin China

ISBN: (数字)9798350372052

ISBN: (纸本)9798350372069

Lumbar disc herniation is a common spinal disease, and its diagnosis and treatment usually rely on medical imaging technology. Traditional medical image diagnosis methods require high labor and time costs. This article proposes an object detection network based on FCOS to achieve automated detection of lumbar disc herniation. This network introduces the idea of group convolution and incorporates attention mechanisms. In addition, the Soft NMS algorithm was used in the inference stage to avoid the problem of target boxes being violently deleted. The method proposed in this article achieved a mAP of 0.661 on the MRI test set, which has higher accuracy compared to other networks. This method provides a feasible and effective solution for detecting lumbar disc herniation, which is expected to assist doctors in clinical diagnosis.

关键词： Computer science Costs Convolution Magnetic resonance imaging Object detection Inference algorithms detection algorithms

来源：评论

学校读者我要写书评

暂无评论

YOLO-LF: An Ultra-Lightweight and Fast Forest Fire detection Algorithm for Edge Devices

YOLO-LF: An Ultra-Lightweight and Fast Forest Fire Detection...

引用

Networking Systems of AI (INSAI), International Conference on

作者： Zhuoyue Ding Run Ye Xiaojia Zhuo Qilin Li Lei Huang Bin Yan School of Automation Engineering University of Electronic Science and Technology of China Chengdu China Metering Center State Grid Sichuan Electric Power Company Chengdu China School of Resources and Environment University of Electronic Science and Technology of China Chengdu China

ISBN: (数字)9798350344660

ISBN: (纸本)9798350344677

In recent years, there has been signification progress in deep learning-based forest fire detection techniques for natural environment monitoring. However, existing object detection algorithms designed for forest fire scenarios often suffer from computational intensity and storage space consumption, rendering them unsuitable for deployment on edge devices. Moreover, forest fire scenarios have high real-time performance, but current edge device deployments of object detection algorithms result in substantial latency and often miss critical rescue opportunities. To address these challenges, this paper proposes a novel lightweight forest fire detection algorithm, YOLO-LF (LF means both light and fast), which leverages a ShuffleNetv2 backbone network, Group Shuffle Convolutions (GSConv) as the base feature extraction block, and lightweight C3_Faster modules as the feature fusion module, and the lightweight operator Content-Aware ReAssembly of Features (CARAFE) as the up-sampling method. The algorithm ensures real-time detection by edge devices while being able to guarantee the accuracy of forest fires. Ultimately, the results on the forest fire dataset show that YOLO-LF mAP@ 0.5 (average accuracy) and FPS (frames per second) reach 71.2% and 110f/s on the PC side, respectively. On the Raspberry Pi 4B platform, the FPS of YOLO-LF is 5.88 times higher than that of the YOLOv5s, which achieves a high accuracy real-time detection with 20+FPS. The model is not only applicable to forest fires, but also a generalized lightweight detection model that can meet most of the requirements for real-time detection deployed on edge devices, and thus has great theoretical significance and application value.

关键词： YOLO Accuracy Image edge detection Computational modeling Forestry Feature extraction Rendering (computer graphics) Real-time systems Complexity theory detection algorithms

来源：评论

学校读者我要写书评

暂无评论

SLAM Loop Closure detection Algorithm Based on 3D Bag-of-Words Model

SLAM Loop Closure Detection Algorithm Based on 3D Bag-of-Wor...

引用

Computer, Internet of Things and Smart City (CIoTSC), International Conference on

作者： Wu Liangdong Zou Yaping Wang Dong Chen Ling School of Computer Engineering and Design Hunan Normal University Changsha China

ISBN: (数字)9798350306569

ISBN: (纸本)9798350306576

In outdoor environments, the loop closure detection in laser-based SLAM faces challenges such as complex lighting variations, low-texture regions, occlusions, and dynamic objects, which limit the accuracy and efficiency of loop closure detection. To address this issue, a novel laser-based SLAM loop closure detection algorithm is proposed, specifically designed for outdoor environments, based on three-dimensional global descriptor features. The algorithm utilizes a BoWЗD model constructed from the three-dimensional laser features for precise point-to-point matching and loop closure detection. The proposed algorithm is integrated into the LeGO-LOAM system and evaluated on the publicly available KITTI dataset. Experimental results demonstrate a 64% improvement in absolute trajectory accuracy and reduced computational time compared to the baseline algorithm.

关键词： Solid modeling Simultaneous localization and mapping Accuracy Three-dimensional displays Smart cities Feature extraction Laser modes Approximation algorithms Trajectory detection algorithms

来源：评论

学校读者我要写书评

暂无评论

A DDoS attack detection algorithm based on improved grid search to optimize SVM in SDN environment

A DDoS attack detection algorithm based on improved grid sea...

引用

Electrical Engineering, Big Data and algorithms (EEBDA), IEEE International Conference on

作者： Meigen Huang Boqi Zhao Chongqing University of Posts and Telecommunications Chongqing China

Software defined network has proposed a new architecture idea, but the controller is vulnerable to distributed denial of service attacks, resulting in a single point of failure. This paper proposes a DDoS attack detection algorithm based on improved grid search optimization SVM in SDN environment. Firstly, according to the characteristics of DDoS attacks and information of flow table items, the most relevant six-dimensional features are selected, then use the improved grid search algorithm to calculate the optimal penalty coefficient and kernel function parameters of SVM model, finally use SVM algorithm to detect DDoS attacks and improve the performance of intrusion detection algorithm. The experimental results in the same scene show that the improved SVM has higher detection rate and recall rate than the SVM algorithm.

关键词： Support vector machines Electrical engineering Intrusion detection Denial-of-service attack Feature extraction detection algorithms Software defined networking

来源：评论

学校读者我要写书评

暂无评论

A Low Coupling and Lightweight Algorithm for Ship detection in Optical Remote Sensing Images

引用

IEEE GEOSCIENCE AND REMOTE SENSING LETTERS 2022年 19卷 1页

作者： Deng, Guochao Wang, Qin Jiang, Jianfei Hong, Qirun Jing, Naifeng Sheng, Weiguang Mao, Zhigang Shanghai Jiao Tong Univ Dept Micro Nano Elect Sch Elect Informat & Elect Engn Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ Natl Key Lab Sci & Technol Micro Nano Fabricat Shanghai 200240 Peoples R China

In recent years, many ship detection algorithms based on convolutional neural networks (CNNs) have been proposed to improve the performance of ship detection. However, with the increase in model complexity and size, it is challenging to deploy these models to resource-constrained edge platforms. In this letter, a low coupling algorithm that belongs to anchor-free methods is proposed for ship detection to reduce the model complexity and still obtain a competitive performance. The proposed low coupling network (LCNet) is easy to deploy and contributes to speeding up the inference and improving memory utilization. In addition, we propose a model compression process consisting of the quantization-aware training (QAT) method and a structural pruning method based on Taylor expansion, which can effectively reduce the model size according to hardware resource constraints. Comparative experimental results demonstrate that LCNet outperforms the state-of-the-art ship detection and natural object detection algorithms, with a 95.27 mAP and 88.91 F1 score on the HRSC2016 dataset. Our proposed model compression method also achieves a compression ratio of at least 80 with a negligible loss of performance.

关键词： Marine vehicles Training Couplings Convolution Quantization (signal) detection algorithms Deconvolution Low coupling model compression optical remote sensing images ship detection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：