检索结果-内蒙古大学图书馆

GCN-YOLO: YOLO Based on Graph Convolutional Network for SAR Vehicle Target Detection

IEEE GEOSCIENCE AND REMOTE SENSING LETTERS 2024年 21卷

作者： Chen, Peiyao Wang, Yinghua Liu, Hongwei Xidian Univ Natl Key Lab Radar Signal Proc Xian 710071 Peoples R China

Recently, deep convolutional neural networks have been widely applied in target detection of synthetic aperture radar (SAR) images. However, the regular convolution kernel cannot effectively establish dependency between features of SAR image with geometric distortion. Meanwhile, SAR images contain a small number of vehicle targets, and the imbalance problem between foreground-background class is serious during training. To solve these problems, we propose a you only look once (YOLO) detector based on graph convolutional network (GCN) called GCN-YOLO. First, a multilayer GCN model called vision GNN (ViG) is used as feature extractor to model the local area and build long-term dependencies between features. In addition, a convolutional block attention module (CBAM) is embedded into the last layer to enhance semantic features. Then, we introduce the VariFocal loss (VFL) as confidence loss to relief the imbalance problem between positive and negative samples. The experimental results on the miniSAR data demonstrate the effectiveness of the proposed method.

关键词： Feature extraction Synthetic aperture radar YOLO Training Radar polarimetry Convolution Detectors Graph convolutional network (GCN) synthetic aperture radar (SAR) target detection VariFocal loss (VFL) you only look once (YOLO)

来源：评论

学校读者我要写书评

暂无评论

small animal intrusion detection based on RAFT

Small animal intrusion detection based on RAFT

引用

2024 International Conference on Image, signal processing, and Pattern Recognition, ISPP 2024

作者： Li, Xiuguang He, Ninghui Sha, Weiyan Zhou, Xiu Ma, Bo Liu, Rui State Grid Ningxia Electric Power Co. Ltd. Electric Power Research Institute Ningxia Hui Autonomous Region Yinchuan City750001 China Beijing Smartchip Microelectronics Technology Company Limited BeiJing City100000 China

ISBN: (纸本)9781510680425

Due to the variety, size and shape of the invasive animals in substation, the general target detection and case segmentation tasks are often faced with the problems of fewer data and lack of *** flow estimation can accurately detect the boundary of the moving object through the change of optical flow during the moving process of the *** can effectively detect boundaries for small animals moving at low ***, RAFT for small animal intrusion detection can not only effectively eliminate the interference of non-invasive objects, but also cancel the range of intrusion targets, and improve the generalization ability of moving object intrusion *** experimental results show that the detection effect is good for small invasive animals with low speeds. © 2024 SPIE.

关键词： Invertebrates

来源：评论

学校读者我要写书评

暂无评论

AugTarget data Augmentation for Infrared small Target Detection 48

AugTarget Data Augmentation for Infrared Small Target Detect...

引用

48th IEEE International Conference on Acoustics, Speech and signal processing, ICASSP 2023

作者： Chen, Shengjia Zhu, Jiewen Ji, Luping Pan, Hongjun Xu, Yuhao University of Electronic Science and Technology of China School of Computer Science and Engineering Chengdu611731 China Sichuan University School of Computer Science and Engineering Chengdu610065 China

ISBN: (纸本)9781728163277

Sample shortage has always been a frequently-faced problem for the machine-learning models in infrared small target detection. As one of main limitations, it is hampering the further promotion of target detection performance. In this paper, we propose a simple and effective data augmentation scheme, AugTarget, to address this shortage issue of small target samples. Our scheme mainly consists of two crucial algorithms: target augmentation and batch augmentation. The former is designed to generate sufficient targets, by random target representation. The latter is devised to diversify training samples. Moreover, the initially-generated image samples of small targets are further enriched by randomly aggregating the feature representation of different images. The experiments on public datasets demonstrate that our AugTarget could bring an obvious improvement to mean intersection over union (mIoU). Cooperated by the augmentation of our AugTarget, the state-of-the-art (SOTA) method, AGPC could even achieve a distinct performance promotion by 3.03%, 2.17% and 2.63% on MDFA, SIRST-Aug and Merged datasets, respectively. In addition, the experimental results on three baseline models also show the universality & adaptivity of AugTarget to different dataset augmentation. Our source codes are available at https://***/UESTC-nnLab/AugTarget. © 2023 IEEE.

关键词： data augmentation Infrared small target detection Target & batch augmentation

来源：评论

学校读者我要写书评

暂无评论

Intelligent Reflecting Surface-Aided Electromagnetic Stealth Against Radar Detection

引用

IEEE TRANSACTIONS ON signal processing 2024年 72卷 3438-3452页

作者： Zheng, Beixiong Xiong, Xue Tang, Jie Zhang, Rui South China Univ Technol Sch Microelect Guangzhou 511442 Peoples R China Shenzhen Res Inst Big Data Shenzhen 518172 Peoples R China South China Univ Technol Sch Future Technol Guangzhou 511442 Peoples R China Frontier Res Ctr Peng Cheng Lab Shenzhen 518055 Peoples R China South China Univ Technol Sch Elect & Informat Engn Guangzhou 510640 Peoples R China Green Revolutionary Ind Power Commun Lab Guangzhou 510000 Peoples R China Chinese Univ Hong Kong Shenzhen Res Inst Big Data Sch Sci & Engn Shenzhen 518172 Guangdong Peoples R China Natl Univ Singapore Dept Elect & Comp Engn Singapore 117583 Singapore

While traditional electromagnetic stealth materials/metasurfaces can render a target virtually invisible to some extent, they lack flexibility and adaptability, and can only operate within a limited frequency and angle/direction range, making it challenging to ensure the expected stealth performance. In view of this, we propose in this paper a new intelligent reflecting surface (IRS)-aided electromagnetic stealth system mounted on targets to evade radar detection, by utilizing the tunable passive reflecting elements of IRS to achieve flexible and adaptive electromagnetic stealth in a cost-effective manner. Specifically, we optimize the IRS's reflection at the target to minimize the sum received signal power of all adversary radars. We first address the IRS's reflection optimization problem using the Lagrange multiplier method and derive a semi-closed-form optimal solution for the single-radar setup, which is then generalized to the multi-radar case. To meet real-time processing requirements, we further propose low-complexity closed-form solutions based on the reverse alignment/cancellation and minimum mean-square error (MMSE) criteria for the single-radar and multi-radar cases, respectively. Additionally, we propose practical low-complexity estimation schemes at the target to acquire angle-of-arrival (AoA) and/or path gain information via a small number of receive sensing devices. Simulation results validate the performance advantages of our proposed IRS-aided electromagnetic stealth system with the proposed IRS reflection designs.

关键词： Electromagnetics Radar Sensors Radar detection Electromagnetic scattering Wireless communication Real-time systems Electromagnetic stealth intelligent reflecting surface (IRS) reflection optimization anti-radar detection angle-of-arrival (AoA) estimation

来源：评论

学校读者我要写书评

暂无评论

A small target detection algorithm for aerial image with improved YOLOv5 3

A small target detection algorithm for aerial image with imp...

引用

3rd International Conference on Electronic Information Engineering, Big data, and Computer Technology, EIBDCT 2024

作者： Xue, Xianli Jiao, Weidong Zhang, Siyuan Qu, Wenkang Tianjin Key Laboratory of Intelligent Signal and Lmage Processing Civil Aviation University of China TIANJIN300300 China

ISBN: (纸本)9781510680449

The problem of complex backgrounds, dramatic changes in the scale of objects and the presence of a large number of dense small targets in the UAV aerial images. In this study, a detection algorithm for small targets is proposed based on the YOLOv5s algorithm. Firstly, a new BottleNeck: C3LK module is designed for the C3 module using depthwise separable convolution to construct a large convolutional kernel on the YOLOv5 backbone network, which improves the ability to capture global information and compensates for the lack of local information for detecting small targets. Secondly, a small target detection layer and two feature fusion layers are added. At the same time, In order to better detect objects of different scales in complex backgrounds, a module for multi-scale feature extraction is designed by combining dilated convolution and attention mechanism. The experimental results show that the proposed algorithm improved 5.4% over YOLOv5s in terms of mAP@0.5 and over YOLOv7-tiny. The method meet the requirement of real-time detection in aerial image small target detection, and improve the detection performance in small target scenes. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Sixteenth International Conference on signal processing Systems, ICSPS 2024

Sixteenth International Conference on Signal Processing Syst...

引用

16th International Conference on signal processing Systems, ICSPS 2024

ISBN: (纸本)9781510689251

The proceedings contain 197 papers. The topics discussed include: variational modal decomposition based on whale optimization algorithm for vital signs detection of FMCW radar;sea surface small target detection using tri-median normalized doppler spectrum feature;a physical cluster-based passive scatterer detection method in ISAC systems;on detection of moving weak targets using MIMO continuous active sonar and Doppler filtering;design of optical testing turntable data acquisition system based on LabVIEW;AOA /TDOA location technology of weak satellite signal based on satellite-ground cooperation;bearing-crossing localization technology for UAV target based on signal characteristics;and separation and localization of multiple bat signals using frequency domain beamforming and dual-array-localization method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Convolutional Neural Network Algorithm and Application Method for Real-Time Beam Steering in RF System

引用

IEEE ACCESS 2024年 12卷 134498-134509页

作者： Byun, Sung-June Ann, Da-Yeong Jo, Jong-Wan Lee, Heejeong Jasmine Jung, Yeon-Jae Kim, Seok-Kee Pu, Young-Gun Lee, Kang-Yoon Sungkyunkwan Univ Dept Elect & Comp Engn Suwon 16419 South Korea SKAIChips Suwon 16571 South Korea Sungkyunkwan Univ Coll Informat & Commun Engn Suwon 16419 South Korea

This paper presents a novel artificial intelligence (AI)-based phase shift system in a beamforming system implemented with field programmable gate array (FPGA)-based hardware by integrating a conventional convolutional neural network (CNN) algorithm. The position of the target can be determined through a phase shifter in a beamforming system using artificial intelligence. In a system that emits a beam from a radio frequency (RF) transmitter and receives a beam from an RF receiver, artificial intelligence can control the phase. It controls the phase of the transmitter for beam scanning and the phase to optimize the signal-to-noise ratio (SNR) of the receiver. The position of the target was detected by learning the signal input data from the receiver. targets were detected through two-beam scanning processes in a 3D space. The first is a coarse process of detecting the approximate position of the target in the entire space, and the second is a fine process of detecting the area in detail after detecting the first approximate position. The phases of the individual antennae should be controlled for optimal beamforming based on the 5x 5 antenna, and the phase is detected at high speed by holding the phase large in the first coarse tuning. The second scan entails a narrow range scan with a small phase to detect it at a high speed accurately. This study shows that with FPGA, AI beamforming can be implemented through two scanning methods without image sensors. Based on the receiver's 5x5 antenna, the CNN input feature consisted of 35x35 classifies the class with high accuracy.

关键词： Artificial intelligence Array signal processing Field programmable gate arrays Antennas Radio frequency Receiving antennas Transceivers Convolutional neural networks Convolutional neural network artificial intelligence beamforming beamforming algorithm RF system

来源：评论

学校读者我要写书评

暂无评论

Improved pedestrian detection algorithm of Yolov5 in dense scenes

Improved pedestrian detection algorithm of Yolov5 in dense s...

引用

2024 International Conference on Image, signal processing, and Pattern Recognition, ISPP 2024

作者： Lin, Chuanwen Liang, Jiaxing School of Artificial Intelligence and Big Data Hefei University Hefei China Institute of Dataspace Hefei China

ISBN: (纸本)9781510680425

Occluded, small-scale pedestrians are easy to occur in dense scenes to solve this problem, an improved Yolov5 is proposed detection algorithm. First of all, feature extraction for small-size pedestrians is insufficient the problem uses SPD-Conv convolution to enhance the complex background small target feature extraction ability. Second, fuse an efficient multi-foot the degree of attention module further enhances the visual area of the target pedestrian feature extraction. Finally, focal eiou loss was used as the loss letter number, so that the regression more attention to high-quality samples, improve the accuracy of the model and robustness. Training and testing were performed on the dataset CrowdHuman in the test, the improved Yolov5 algorithm AP50 can reach 83.6 %, phase for the original algorithm, AP50 and AP50−95 are improved by 2.5 %, respectively. Average accuracy improved by 2.2%. The experimental results validate the proposed method effectiveness in crowded scenarios. © 2024 SPIE.

关键词： Detection and tracking algorithms Object detection Convolution Performance modeling Feature extraction small targets data modeling Evolutionary algorithms Target detection Statistical modeling

来源：评论

学校读者我要写书评

暂无评论

Bioradar signal processing Algorithms based on Machine Learning 2

Bioradar Signal Processing Algorithms based on Machine Learn...

引用

2nd IEEE International Conference on Integrated Intelligence and Communication Systems, ICIICS 2024

作者： Guo, Huanyi Wuhan University of Bioengineering Hubei Wuhan China

ISBN: (纸本)9798331504960

Since human physiological signals (e.g., respiration and heart rate) generate very small Doppler shifts, they are highly susceptible to being masked by strong clutter backgrounds. The main research focus of this project is to accurately extract weak targets (i.e., vital signs) from complex environments. The signals received by BioRadar contain vital signs signals and clutter signals. Based on the analysis of the collected data, it is known that vital signals have low signal-to-noise ratio and are characterized by quasi-periodicity and multiple harmonics;whereas clutter can be manifested as Gaussian colored noise. For this reason, this study translates the bio-radar signal processing into the problems of harmonic modeling, clutter suppression, and signal-to-noise ratio enhancement in the presence of Gaussian colored noise. The research objective is to extract the weak vital sign signals of human body from the complex background by machine learning techniques. In order to effectively extract physiological signals such as respiration and heart rate, this paper employ an ultra-wideband bio-radar system and advanced signal processing algorithms, and the experimental results show that the respiratory signal of participant 1 at 0 s is 2.1 mV, the heartbeat signal is 1.0 mV, and the suppressed clutter signal is 0.1 mV, which effectively suppresses the clutter interference with high accuracy and reliability. © 2024 IEEE.

关键词： White noise

来源：评论

学校读者我要写书评

暂无评论

An Ionospheric Pollution Correction Method Based on Parametric Time-Frequency Analysis of PCT 2

An Ionospheric Pollution Correction Method Based on Parametr...

引用

2nd IEEE International Conference on signal, Information and data processing, ICSIDP 2024

作者： Zhao, Suqiang Li, Yajun Xiong, Huai Yao, Tianni Xu, Lin Ming, Xinyue East China Normal University Aerospace System Engineering Shanghai School of Communication and Electronic Engineering Shanghai China East China Normal University School of Communication and Electronic Engineering Shanghai China Shanghai Dianji University Micro-nano Satellite Intelligent Research Institute Shanghai China Aerospace System Engineering Shanghai Shanghai China

ISBN: (纸本)9798331515669

The ionosphere serves as the primary propagation medium for High-Frequency Sky-Surface Wave Radar (HFSSWR). However, its non-stationary and time-varying characteristics cause severe broadening of the first-order sea clutter spectrum in the combined sky-emitting and ground-receiving mode, hindering the detection of slow ship targets. To address this issue, this paper presents an ionospheric contamination correction method for HFSSWR based on the Polynomial Chirplet Transform (PCT) algorithm. Unlike the traditional Chirplet Transform (CT), which uses a sinusoidal kernel, this method employs a polynomial kernel for time-frequency analysis of the broadened echo spectrum. It extracts the phase contamination function by estimating the instantaneous frequency, thereby correcting the ionospheric phase path perturbation in the echo spectrum. This approach effectively suppresses sea clutter broadening caused by ionospheric perturbations. Simulation and experimental results demonstrate that, compared to conventional ionospheric decontamination methods, this method achieves superior suppression of broadened sea clutter spectra caused by both small- and large-amplitude ionospheric phase perturbations, improving the signal-to-noise ratio. This method effectively mitigates sea clutter broadening in HFSSWR. © 2024 IEEE.

关键词： signal to noise ratio

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：