融合边缘感知与统计纹理知识的输电线路金具锈蚀检测  

Rust detection in transmission line fittings via fusion of edge perception and statistical texture knowledge

作　　者：赵振兵[1] 郭广学 王艺衡 赵文清[2] 翟永杰[2] ZHAO Zhenbing;GUO Guangxue;WANG Yiheng;ZHAO Wenqing;ZHAI Yongjie(School of Electrical and Electronic Engineering,North China Electric Power University,Baoding 071003,China;School of Control and Computer Engineering,North China Electric Power University,Baoding 071003,China)

机构地区：[1]华北电力大学电气与电子工程学院,河北保定071003 [2]华北电力大学控制与计算机工程学院,河北保定071003

出　　处：《智能系统学报》2024年第5期1228-1237,共10页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金项目(U21A20486,62373151,62371188,62303184);河北省自然科学基金项目(F2021502008,F2021502013);中央高校基本科研业务费专项资金项目(2023JC006).

摘　　要：针对输电线路金具目标小、背景环境复杂和锈蚀区域不规则等问题,提出了一种融合边缘感知与统计纹理知识的输电线路金具锈蚀检测算法。首先通过改进YOLOv7模型检测金具,然后利用改进Res-UNet模型对检测的金具进行锈蚀分割,加入SE(squeeze-excitation)注意力提高模型的稳定性,引入统计纹理知识模块(statistical texture knowledge module,STM)和边缘感知模块(edge-aware module,EAM),提出一种知识融合模块对边缘感知和统计纹理知识进行融合,提高对锈蚀分割精度。实验结果表明,检测和分割模型mAP分别提高了2.8百分点和7.7百分点。To address issues such as small target sizes,complex background environments,and irregular rust areas,a new rust detection algorithm combining edge sensing and statistical texture knowledge is proposed.First,the YOLOv7 model is improved for detecting fittings.Furthermore,the enhanced Res-UNet model is used for corrosion segmentation on the detected fittings.Additionally,squeeze excitation is incorporated to improve the stability of the model.The statistical texture knowledge module and edge-aware module are introduced,and a knowledge fusion module is proposed to integrate edge perception with statistical texture knowledge to enhance the precision of rust segmentation.Experimental results show that the detection and segmentation models increased by 2.8 percentage points and 7.7 percentage points,respectively.

关 键 词：目标检测 语义分割 输电线路 锈蚀检测 金具 注意力机制 统计纹理 边缘感知 知识融合 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]