一种基于雷达和相机数据融合网络的输电线路鸟类多目标识别方法

摘要
图/表
参考文献
相关文章 (0)

全文: PDF (11853 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要本文提出一种基于雷达与相机数据融合的输电线路鸟类多目标识别网络（RVFNet）。该网络结合雷达射频（RF）数据与视觉图像,实现监控范围内鸟类目标的高精度识别。针对多模态数据语义差异的问题,计算雷达射频信号与图像位置信息的对应关系,确保特征表达的一致性。在网络结构方面,设计鸟类姿态卷积（BPC）网络,有效融合多模态信息,提升对小目标特征的提取与细节保留能力;同时,采用特征融合模块（FFM）整合不同模态特征,在显著提升特征交互能力的同时保持较低计算成本。实验表明,RVFNet在多种天气条件下对鸟类的平均识别准确率达到80.18%,展现出良好的鲁棒性。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	范程涛
	高伟
	靳小喜

关键词 ：探鸟驱鸟, 视觉图像, 雷达射频图像, 传感器融合, 深度卷积神经网络

Abstract：This paper proposes a multi-target recgnition network for birds on power transmission lines, called RVFNet, based on the fusion of radar and camera data. The network achieves high-precision recgnition of bird targets within the monitoring range by integrating radar radio frequency (RF) data with visual images. To address the semantic differences between multimodal data, the correspondence between radar RF signals and image positional information is calculated to ensure consistency in feature representation. Structurally, the network incorporates a bird posture convolutional network (BPC) to effectively fuse multimodal information, enhancing the extraction of small-target features and preserving fine details. Additionally, a feature fusion module (FFM) is introduced to integrate multimodal features, significantly improving feature interaction while maintaining low computational costs. Experimental results demonstrate that RVFNet achieves an average bird recognition accuracy of 80.18% under various weather conditions, highlighting its robustness.

Key words： identify and repel birds visual images radar radio frequency images sensor fusion deep convolutional neural networks

收稿日期: 2025-01-09

作者简介: 范程涛（1999—）,男,福建省三明市人,硕士研究生,主要从事多模态信号输电线路探鸟驱鸟技术研究工作。

引用本文:

范程涛, 高伟, 靳小喜. 一种基于雷达和相机数据融合网络的输电线路鸟类多目标识别方法[J]. 电气技术, 2025, 26(6): 29-37. FAN Chengtao, GAO Wei, JIN Xiaoxi. A multi-target bird recognition method for transmission lines based on radar and camera data fusion. Electrical Engineering, 2025, 26(6): 29-37.

链接本文:

https://dqjs.cesmedia.cn/CN/Y2025/V26/I6/29

[1] 杨鑫, 周攀, 李斌, 等. 10kV配电线路鸟害防治措施研究[J]. 科技创新与应用, 2021, 11(18): 109-113.
[2] 戴宇辰, 叶青, 许安杰, 等. 配电线路鸟害故障预测模型研究[J]. 电气技术, 2018, 19(3): 100-102.
[3] 王翼虎, 徐浩. 330kV紧凑型铁塔鸟害故障研究[J]. 电工技术, 2023(22): 202-205.
[4] 付豪, 张东, 董新胜, 等. 输电线路涉鸟故障防范措施研究[J]. 电气技术, 2016, 17(3): 132-135.
[5] 李秋平. 关于20kV线路裸导线运行的防鸟害研究[J]. 电气技术, 2016, 17(8): 136-139.
[6] 张传民, 张东. 防鸟刺针刺长度选择方法[J]. 电气技术, 2023, 24(4): 48-51, 56.
[7] 刘希和, 刘佳俊, 柯贤友, 等. 220kV输电线路感应电压驱鸟试验和应用[J]. 电气技术, 2013, 14(9): 44-48.
[8] 周文平, 李永茂, 张君, 等. 电力线路驱鸟防治系统研究[J]. 电工技术, 2020(5): 58-59, 62.
[9] 刘传洋, 吴一全, 刘景景. 无人机航拍图像中绝缘子缺陷检测的深度学习方法研究进展[J]. 电工技术学报, 2025, 40(9): 2917-2930.
[10] 郑含博, 胡思佳, 梁炎燊, 等. 基于YOLO-2MCS的输电线路走廊隐患目标检测方法[J]. 电工技术学报, 2024, 39(13): 4164-4175.
[11] 仲林林, 胡霞, 刘柯妤. 基于改进生成对抗网络的无人机电力杆塔巡检图像异常检测[J]. 电工技术学报, 2022, 37(9): 2230-2240, 2262.
[12] 廖宁生, 曹天秀, 刘科言, 等. 复合特征与多尺度融合的无人机小目标检测算法[J]. 计算机工程与应用, 2025, 61(3): 111-120.
[13] 杨波, 曹雪虹, 焦良葆, 等. 改进实时目标检测算法的电力巡检鸟巢检测[J]. 电气技术, 2020, 21(5): 21-27, 32.
[14] LIANG Haijun, ZHANG Xiangwei, KONG Jianguo, et al.SMB-YOLOv5: a lightweight airport flying bird detection algorithm based on deep neural networks[J]. IEEE Access, 2024, 12: 84878-84892.
[15] WANG Yizhou, JIANG Zhongyu, LI Yudong, et al.RODNet: a real-time radar object detection network cross-supervised by camera-radar fused object 3D localization[J]. IEEE Journal of Selected Topics in Signal Processing, 2021, 15(4): 954-967.
[16] 吴洋铭, 洪翠, 高伟. 基于雷达点云与视觉图像融合的输电线路探鸟驱鸟技术[J]. 高电压技术, 2023, 49(8): 3446-3457.
[17] GAO Wei, WU Yangming, HONG Cui, et al.RCVNet: a bird damage identification network for power towers based on fusion of RF images and visual images[J]. Advanced Engineering Informatics, 2023, 57: 102104.
[18] LIANG Tianyi, LI Baopu, WANG Mengzhu, et al.A closer look at the joint training of object detection and re-identification in multi-object tracking[J]. IEEE Transactions on Image Processing, 2022, 32: 267-280.
[19] ZHENG Zhaohui, WANG Ping, LIU Wei, et al.Distance- IoU loss: faster and better learning for bounding box regression[C]//AAAI Conference on Artificial Intel- ligence, New York, USA, 2020: 12993-13000.
[20] REN Shaoqing, HE Kaiming, GIRSHICK R, et al.Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[21] DAI Jifeng, QI Haozhi, XIONG Yuwen, et al.Defor-mable convolutional networks[C]//2017 IEEE Inter- national Conference on Computer Vision (ICCV), Venice, Italy, 2017: 764-773.
[22] REDMON J, FARHADI A.YOLOv3: an incremental improvement[EB/OL]. https://arxiv.org/abs/1804.02767v1.
[23] KOWOL K, ROTTMANN M, BRACKE S, et al.YOdar: uncertainty-based sensor fusion for vehicle detection with camera and radar sensors[EB/OL]. https://arxiv.org/abs/2010.03320v2.