检索结果-内蒙古大学图书馆

A two-stage detection method based on improved DP-TBD for marine weak extended targets

DIGITAL signal processing 2025年 160卷

作者： Han, Jiacheng Hou, Yuguan Tao, Chenchen Duan, Zhihui Harbin Inst Technol Sch Elect & Informat Engn Harbin 150001 Heilongjiang Peoples R China Beijing Inst Remote Sensing Equipment Beijing 100854 Peoples R China Beijing Res Inst Telemetry Beijing 100076 Peoples R China

Detecting weak extended targets on the sea surface under cluttered background is very challenging, and traditional detection methods are affected by clutter and low signal-to-clutter-and-noise ratio, resulting in poor detection performance. In order to improve the detection and discovery capabilities of weak extended targets on the sea surface, this paper proposes an improved method based on the dynamic programming track-before-detect (DP-TBD) algorithm, which is called DK-TBD. DK-TBD has two stages. One is the rough TBD detection based on virtual nodes, which solves the problem of target loss caused by environmental factors in classical DP-TBD. The other is the re-tracking based on kernel correlation filtering algorithm, which is used to correct the state of weak extended targets after rough detection. Compared with other DP-TBD-based methods, our method can better detect weak extended targets on the sea surface in simulated and real data with complex clutter as background.

关键词： Dynamic programming Kernel correlation filter Extended target Track-before-detect

来源：评论

学校读者我要写书评

暂无评论

Learning Source-Free Domain Adaptation for Infrared small Target Detection

引用

IEEE signal processing LETTERS 2025年 32卷 1121-1125页

作者： Jin, Hongxu Chen, Baiyang Lu, Qianwen Tao, Qingchuan Li, Yongxiang Sichuan Univ Coll Elect Informat Chengdu 610065 Peoples R China Sichuan Univ Coll Comp Sci Chengdu 610065 Peoples R China Univ Tokyo Grad Sch Interdisciplinary Informat Studies Tokyo 1638001 Japan

Existing infrared small target detection (IRSTD) methods mainly rely on the assumption that the training and testing data come from the same distribution, a premise that does not hold in many real-world scenarios. Additionally, the inability to access source domain data in numerous IRSTD tasks further complicates the domain adaptation process. To address these challenges, we propose a novel Source-Free Domain Adaptation (SFDA) framework for IRSTD, denoted as IRSTD-SFDA. This framework comprises two key components: Multi-expert Domain Adaptation (MDA) and Multi-scale Focused Learning (MFL). MDA leverages the source model to generate pseudo masks for the target domain, facilitating the transfer of knowledge from the source to the target domain. To account for the inherent diversity of small targets across domains, MDA refines these pseudo masks through a series of operations, including target localization, rolling guidance filtering, shape adaptation, and multi-expert decision, thereby mitigating morphological discrepancies between the source and target domains. Meanwhile, MFL employs a global-local fusion strategy to focus on critical regions, enhancing the model's ability to detect small infrared targets. Extensive experimental evaluations across various cross-domain scenarios demonstrate the effectiveness of the proposed framework.

关键词： Adaptation models Training Filtering Testing Shape Object detection Location awareness data models Clutter Image edge detection Infrared small target detection source-free domain adaptation multi-expert domain adaptation

来源：评论

学校读者我要写书评

暂无评论

Enhanced YOLOv8 for small-object detection in multiscale UAV imagery: Innovations in detection accuracy and efficiency

引用

DIGITAL signal processing 2025年 158卷

作者： Luo, Weixin Yuan, Sannan Shanghai Univ Elect Power Sch Elect & Informat Engn Shanghai 201306 Peoples R China

Due to the ability of drone technology to cover wide areas and reach difficult-to-access places, the detection of small targets by Unmanned Aerial Vehicles (UAVs) is crucial in various applications such as urban management planning and emergency response. This paper proposes innovations to the YOLOv8 architecture, significantly enhancing its performance in small target detection from multi-scale drone perspectives. We introduce Channel Priority Attention Dynamic Snake Convolution and Dynamic small Object Detection Head Layer (DyHead-SODL) to improve the model's ability to capture fine details and detection accuracy. Additionally, we implemented an enhanced loss function (MPDIoU) and a Deformable Attention Transformer (DAT) to optimize detection efficiency without increasing computational burden. Experimental results on the Visdrone and RSOD datasets demonstrate significant improvements, with the proposed method increasing mAP50 by 10.5 %, mAP95 by 6.9 %, reducing localization error Eloc by 1.27, and decreasing the miss rate by 2.42 on the Visdrone dataset, outperforming existing state-of-the-art detection models while maintaining low computational complexity. The proposed method has also been tested on four large public datasets: DOTA, HRSC, LEVIR, and CARPK, demonstrating the generalization capability of our model. These advances provide effective solutions for small target detection from drone perspectives and promote the development of single-stage object detection technology. These advancements provide an effective solution for small object detection from UAVs, advancing the field of one-stage object detection technology. (source code: https://***/GuccIceCream/yolov8/tree/master)

关键词： YOLOv8 small object detection Channel prioritized attention dynamic snake convolution Dynamic small object detection head MPDIoU Deformable attention transformer

来源：评论

学校读者我要写书评

暂无评论

IR-MPE: A Long-Term Optical Flow-Based Motion Pattern Extractor for Infrared small Dim targets

引用

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 2025年 74卷

作者： Liu, Xuyang Zhu, Wenming Yan, Pei Tan, Yihua Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Natl Key Lab Multispectral Informat Intelligent Pr Wuhan 430074 Peoples R China

In complex scenarios, the utilization of temporal motion information can improve the detection performance of infrared small and dim targets. However, existing multiframe methods only consider short-term motion information at each moment, which is difficult to capture reliable motion information for small and dim targets. In addition, existing multiframe data-driven methods generally utilize complex network structures which have longer inference times compared with the single-frame models of infrared target detection. Such a problem limits the applicability of the existing multiframe methods. In this article, we propose a long-term optical flow (OF)-based infrared small target motion pattern extractor (IR-MPE) to generate long-term OF energy (OFE) maps, which reflect the motion patterns of targets at the current moment. First, we design a long-term OF adaptive accumulation module (OFAAM) to adaptively control the update of current motion information and the retention of previous motion information. Second, we design an offset correction module (OCM), which is embedded in the OFAAM module to rectify the OFE from the previous frame. Meanwhile, the OCM also corrects the output of the previous frame to assist in the detection of the current frame. Embedding our IR-MPE module into the existing single-frame methods can easily extend them as multiframe methods. The only modification is adding an extra input channel of the first layer whose input is set as the concatenation of our OFE and the original infrared image. Such a simple structure can significantly improve the detection accuracy while maintaining fast inference speed. Extensive experiments on various public datasets show that our approach outperforms the state-of-the-art methods in challenging scenarios.

关键词： data mining Optical flow Object detection Feature extraction Deep learning Fuses Detectors Dams Accuracy signal to noise ratio Infrared small dim target detection long-term optical flow (OF) motion pattern extractor

来源：评论

学校读者我要写书评

暂无评论

Machine learning classifier for marine radar small targets based on normalization features 16

Machine learning classifier for marine radar small targets b...

引用

16th International Conference on signal processing Systems, ICSPS 2024

作者： Zhao, Zi-Jian Shui, Peng-Lang National Key Laboratory of Radar Signal Processing Xidian University China

ISBN: (纸本)9781510689251

targets on the sea surface are of great concern, especially the classification of small targets, which is crucial for marine safety. Therefore, we propose a machine learning classifier for marine radar small targets based on normalization features, which can effectively classify targets with small radar cross section (RCS) and signal-to-clutter ratio (SCR) on the sea surface. We standardize the radar parameters because different radars have different system parameters, resulting in limited generalization ability of the trained classifier. We extract 9 features from multiple dimensions of the signal such as time domain, frequency domain and time-frequency domain and then classify targets using decision tree (DT) classifier, K-nearest neighbors (KNN) classifier, bootstrap aggregation (Bagging) classifier and adaptive boosting (AdaBoost) classifier. Experimental validation is performed on publicly available IPIX, self-measured radar data and data provided by the South African Council for Scientific and Industrial Research (CSIR), which contains four types of targets: floating ball, floating boats, low-altitude unmanned aerial vehicles (UAVs) and moving speedboats, and the results of the experiments show that the proposed classifier achieves a correct classification accuracy of 0.8537. © 2025 SPIE.

关键词： Marine radar

来源：评论

学校读者我要写书评

暂无评论

Joint resource allocation and robust weighting in element-pulse coding MIMO radar for tracking multiple targets under deceptive jamming

引用

DIGITAL signal processing 2025年 161卷

作者： Zhang, Yi Tao, Haihong Guo, Jingjing Yan, Yingfei Xidian Univ Natl Key Lab Radar Signal Proc Xian 710071 Peoples R China

Given the practical challenges of implementing distributed multi-input multi-output (MIMO) radar systems, achieving enhanced tracking performance under constrained resources is essential for centralized MIMO radar in complex deceptive jamming environments. This paper proposes a resource-aware closed-loop signal processing framework for centralized element-pulse coding MIMO (EPC-MIMO) radar to enhance multi-target tracking (MTT) performance. Specifically, by leveraging a priori MTT information, signal preprocessing, and transmit-receive spatial frequency, true and false targets are effectively distinguished, and data-independent weight is employed to suppress deceptive jamming. To achieve robust deceptive jamming suppression and optimal MTT, a joint resource allocation and robust weighting (JRARW) strategy is developed for beam selection, power allocation, and robust weight optimization. Utilizing the posterior Cram & eacute;r-Rao lower bound (PCRLB) as a quantitative metric for tracking accuracy, the PCRLB for EPC-MIMO radar is derived and adopted as the objective function. Given the non-convexity of the JRARW, a three-stage cycle-based solution is proposed to address it. Simulation results demonstrate that the JRARW has outstanding capabilities in robust deceptive jamming suppression and MTT.

关键词： Resource allocation Deceptive jamming suppression EPC-MIMO radar Robust beampattern Multi-target tracking PCRLB

来源：评论

学校读者我要写书评

暂无评论

Cross-YOLO: an object detection algorithm for UAV based on improved YOLOv8 model

引用

signal IMAGE AND VIDEO processing 2025年第6期19卷 1-10页

作者： Dong, Ying Guo, Jiahao Xu, Fucheng Dalian Minzu Univ Coll Sci 31 Jinshi Rd Dalian 116000 Liaoning Peoples R China Dalian Minzu Univ Sch Preuniv 31 Jinshi Rd Dalian 116000 Liaoning Peoples R China

In this study, Cross-YOLO, an enhanced version of the YOLOv8 model, is specifically designed to address the challenge of detecting small objects in UAV target detection scenarios. The model refines the original YOLOv8 through several innovative improvements: Firstly, in order to improve the detection accuracy of small targets, we propose Cross-FPN to bolster the original FPN. Secondly, we have redesigned a lightweight detection head, DELDH, to solve the problem of network bloat caused by the introduction of small object detection heads. Thirdly, a new attention mechanism CMCA is designed, that unifies the Coordinate attention mechanism with the Multi-scale convolutional attention mechanism to further enhance the feature extraction of small targets. Finally, the WIoU loss function is introduced to improve the accuracy of bounding box regression and improve detection performance. Experimental data in the Visdrone dataset indicate that, under the condition of the selected model size n, Cross-YOLO achieves a substantial reduction of 35.4% in parameter count compared to YOLOv8n, with only a marginal increase of 7.8% in computational load, and a significant improvement of 5.3% in mAP0.5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$mAP_{0.5}$$\end{document}. Furthermore, its strong performance on the DOTA v1.5 and TinyPerson datasets confirms the model's generalization capabilities and practical applicability.

关键词： YOLOv8 small object Object detection UAV Attention

来源：评论

学校读者我要写书评

暂无评论

Enhanced adaptive detection of nearby and distant ships in fog: A real-time multi-scale target detection strategy

引用

DIGITAL signal processing 2025年 158卷

作者： Chen, Yantong Ren, Jianzhao Li, Jiabao Shi, Yuxin Dalian Maritime Univ Dept Informat Sci & Technol Dalian 116026 Peoples R China

Accurate real-time ship identification in fog is crucial for maritime safety. In foggy conditions, the contrast between ships and their background diminishes, leading to unclear edges. As distance increases, nearby ships appear larger and exhibit relatively clear details. Besides, distant ships appear smaller and have more obscure details due to the fog, which makes detection challenging. To address these challenges, this paper proposes a two-scheme enhanced adaptive ship detection network for foggy maritime environments, based on an improved YOLOv8s. In the first scheme, a codebook is designed that includes hyperparameter codebook, input-encoded features, and scale factors. The purpose is to highlight key features in feature maps and mitigate the impact of weather conditions on detection performance. In the second scheme, a multi-scale fusion module (MFBlock) fuses feature maps of uniform size. The feature injection module adjusts and combines multi-branch information, which is then input into the detection head for small targets. This process enables the model to accurately identify distant small targets. Our method significantly enhances various performance metrics, such as mAP50, mAP50-95, APL, APM, and APS. Additionally, it achieves the frames per second (FPS) of 121, maintaining a good balance. The overall object detection network achieves effective real-time detection of multi-scale ships in foggy conditions.

关键词： Foggy Multi-scale ship Adaptive image weather weakening Codebook Distant small object Information injection

来源：评论

学校读者我要写书评

暂无评论

An NN Framework for Target Number Detection in FMCW Radar Systems via Hankelization: A Low-Dimensional data Representation Perspective

引用

IEEE ACCESS 2025年 13卷 58823-58834页

作者： Kim, Jung-Hwan Kim, Gyung-Eun Lee, Woong-Hee Dongguk Univ Seoul Div Elect & Elect Engn Seoul 04620 South Korea

In this paper, we introduce a novel neural network (NN)-based algorithm that significantly improves the target number detection in frequency modulated continuous wave (FMCW) radar systems. By integrating the mathematical processes of Hankelization and singular value extraction, we can perform input data manipulation for effective target number detection, resulting in constructing an efficient NN framework. This is based on the following mathematical properties: 1) A sequence obtained by uniform sampling of the superposition of K radio waves can be represented as a superposition of K geometric sequences;2) A Hankelized matrix formed by the superposition of K geometric sequences exhibits low-rank characteristics;and 3) In an FMCW radar system with K targets, if the received signal, which is represented as a matrix, is ideal, the vectors obtained by extracting this matrix in row, column, diagonal, and anti-diagonal patterns can all be modeled as a superposition of K geometric sequences. The proposed NN framework showcases remarkable improvements in accuracy and efficiency for target number detection, leveraging a small sized dataset and a compact NN design to achieve unprecedented performance levels. Numerical results validate the superiority of our method across various scenarios, establishing a new benchmark for low-dimensional data representation in radar systems.

关键词： Radar Estimation Accuracy Heuristic algorithms Radar detection Radar antennas Artificial neural networks Vectors signal processing algorithms data visualization Hankelization singular value extraction neural networks radar signal processing

来源：评论

学校读者我要写书评

暂无评论

small object detection in UAV imagery based on channel-spatial fusion cross attention

引用

signal IMAGE AND VIDEO processing 2025年第4期19卷 1-14页

作者： Li, Jianlong Zheng, Chunhou Chen, Peng Zhang, Jun Wang, Bing Anhui Univ Inst Phys Sci & Informat Technol Natl Engn Res Ctr Agroecol Big Data Anal & Applica Sch InternetInformat Mat & Intelligent Sensing La Hefei 230601 Peoples R China Anhui Univ Finance & Econ Sch Management Sci & Engn Bengbu 233030 Peoples R China Anhui Rocvis Intelligent Technol Co Ltd Hefei 230001 Peoples R China

Object detection in unmanned aerial vehicle (UAV) images has become an important research area in computer vision due to its unique value and challenges. UAV images are characterized by densely distributed small targets, significant changes in target scale, and background noise, which affect the accuracy and reliability of detection. To address these issues, we propose an small target detection network based on Enhanced Scale Sequence Fusion and channel space fusion cross-attention mechanism, called *** tackle the high proportion of small targets and scale variation in UAV images, we employ Enhanced Scale Sequence Fusion, integrating fine-grained information from shallow feature maps and semantic information from deep feature maps. Additionally, we incorporate an tiny target detection head to enhance the network's ability to extract fine-grained information features for small targets. To address the issue of background noise, we propose a channel space fusion cross-attention mechanism, which first performs attention calculation on local patch block feature maps, and then performs attention calculation global patch blocks. This captures both long-range dependencies and detailed information. The method for calculating attention combines spatial description information and channel description *** experiments were conducted to validate the effectiveness of the model on the VisDrone benchmark dataset, UAVDT dataset and our self-made UAV power inspection dataset PIDrone. In comparison to the YOLOv8s model, the CSFCANet demonstrated an improvement in mAP of 7% on the PIDrone, 2.4% on the VisDrone, and 3.6% on the UAVDT.

关键词： small object detection Attention mechanism Multi-scale feature fusion UAV Aerial Imagery

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：