航空学报 > 2026, Vol. 47 Issue (S1): 732819-732819   doi: 10.7527/S1000-6893.2025.32819

基于空天无人平台信息的伪装目标智能检测

耿航1,2, 伍圯煊1, 苟轩1, 李新建3,4(), 陈凯1   

  1. 1.电子科技大学 自动化工程学院,成都 611731
    2.电子科技大学 深圳高等研究院,深圳 518000
    3.广西产研院时空信息技术研究所有限公司,南宁 530201
    4.广西大学 电气工程学院,南宁 530004
  • 收稿日期:2025-09-24 修回日期:2025-09-28 接受日期:2025-10-10 出版日期:2025-11-11 发布日期:2025-11-10
  • 通讯作者: 李新建 E-mail:2329594059@st.gxu.edu.cn
  • 基金资助:
    国家自然科学基金(U2330206);国家自然科学基金(U2230206);国家自然科学基金(62173068);四川省科技计划项目(2025ZDZX0073);广西科技计划项目(AB24010157)

Intelligent camouflaged target detection based on information of aerospace unmanned platforms

Hang GENG1,2, Yixuan WU1, Xuan GOU1, Xinjian LI3,4(), Kai CHEN1   

  1. 1.School of Automation Engineering,University of Electronic Science and Technology,Chengdu 611731,China
    2.Shenzhen Institute of Advanced Study,University of Electronic Science and Technology,Shenzhen 518000,China
    3.Guangxi Institute of Industrial Technology for Space-Time Information Co. ,Ltd,Nanning 530201,China
    4.School of Electrical Engineering,Guangxi University,Nanning 530004,China
  • Received:2025-09-24 Revised:2025-09-28 Accepted:2025-10-10 Online:2025-11-11 Published:2025-11-10
  • Contact: Xinjian LI E-mail:2329594059@st.gxu.edu.cn
  • Supported by:
    National Natural Science Foundation of China(U2330206);Sichuan Science and Technology Program(2025ZDZX0073);Guangxi Science and Technology Program(AB24010157)

摘要:

伪装目标智能检测旨在利用有限特征信息,智能识别隐藏在伪装环境中的目标。现有伪装目标智能检测方法存在低显著性特征、类间差异小、数据集标注成本高、小像素点特征提取手段有限及伪装场景下目标关键特征提取性能不佳等问题。针对以上问题,提出了基于空天无人平台辅助信息的伪装目标智能检测新方法。首先,提出了一种改进的深度卷积生成对抗网络,通过设计新的损失函数和注意力机制模块提高图像生成质量,从而扩充数据集;然后,设计了一种改进型的Yolov11网络——Yolov11-Codattention,该网络的独特之处在于利用空天无人平台辅助信息,内嵌了上下文信息和感受野机制2类模块,以解决低显著性特征和类间差异小等问题。通过2类模块的引入,所设计的网络在目标特征提取能力大幅增强的同时,对模型参数量的需求大幅减小。基于公开的空天迷彩数据集,进行了伪装目标检测仿真实验,实验结果表明:提出的Yolov11-Codattention算法与传统Yolov11算法相比,在召回率和mAP@50性能指标上分别提高了7.0%和4.8%,且实时性达到40 FPS;通过与7种常用目标检测算法的对比,发现Yolov11-Codattention具有更高的伪装目标检测精度。外场嵌入式部署实验结果表明:Yolov11-Codattention的平均置信度达到0.63,实时性达到30 FPS,符合工程应用需求。实验结果充分验证了所设计的Yolov11-Codattention算法在伪装目标检测任务中的有效性。

关键词: 空天无人平台, 上下文信息, 感受野机制, 目标检测, 特征提取, 伪装场景

Abstract:

Intelligent camouflage target detection aims to identify targets hidden in camouflage environments using limited feature information. Existing camouflage target detection methods have problems such as low saliency features, small inter-class differences, high annotation costs for datasets, limited means of extracting small pixel features, and poor performance in extracting key features of targets in camouflage scenarios. To address these issues, this paper proposes a new intelligent camouflage target detection method based on an auxiliary information of unmanned aerospace platforms. Firstly, an improved deep convolutional generative adversarial network is proposed to expand the dataset, and enhance the image generation quality by designing a new loss function and embedding an attention mechanism module. Then, an improved Yolov11 network-Yolov11-Codattention is designed. The uniqueness of this network lies in its use of auxiliary information from space---unmanned aerospace platforms and the integration of two types of modules: context information modules and receptive field mechanism modules. These modules help solve problems such as low saliency features and small inter-class differences. By introducing these two types of modules, the designed network significantly enhances the target feature extraction capability while reducing the demand for model parameters. Based on the public camouflage dataset, camouflage target detection simulation experiments were conducted. The experimental results show that the proposed Yolov11-Codattention algorithm improves the recall rate and mAP@50 performance indicators by 7.0% and 4.8% respectively compared with the traditional Yolov11 algorithm, and the real-time performance reaches 40FPS. Through comparison with seven commonly used target detection algorithms, it is found that Yolov11-Codattention has higher camouflage target detection accuracy. The results of the field embedded deployment experiments show that the average confidence of Yolov11-Codattention reaches 0.63, and the real-time performance reaches 30FPS, meeting the engineering application requirements. The experimental results fully verify the effectiveness of the designed Yolov11-Codattention algorithm in the camouflage target detection task.

Key words: unmanned aerospace platform, context information, receptive field mechanism, target detection, feature extraction, camouflage environment

中图分类号: