四川轻化工大学学报（自然科学版）2025，Vol.38Issue(3)：54-64,11.DOI:10.11863/j.suse.2025.03.07
基于改进YOLOv8的变电站典型缺陷检测算法
Typical Defect Detection Algorithm for Substations Based on Improved YOLOv8
摘要
Abstract
As an indispensable part of the power system,the normal operation of the substations play a crucial role in daily life.This paper proposes an improved YOLOv8 substation typical defect detection algorithm to address the issues of increased computational costs and low efficiency caused by using separate models for detecting each type of defect.Firstly,the algorithm first utilizes the Coordinate Attention(CA)mechanism to design a CCA attention module,allowing the network to focus on the location information of defect areas;secondly,Content Aware ReAssembly of Features(CARAFE)upsampling is adopted,which uses the content of the input features themselves to guide upsampling and achieve higher accuracy;in addition,in order to solve the problem of missed and false detections caused by unclear target features,a multi-scale attention mechanism module EMP is designed to effectively extract the contour features of the target;finally,the non maximum suppression algorithm is combined with Inner_MPDIOU to address the limitations of traditional loss functions in detecting irregular objects and objects with large size changes,providing a more accurate measurement method.The improved algorithm in this paper is compared with YOLOv8n,mAP@0.5 is improved by 6.6 percentage points,recall rate is increased by 6.0 percentage points,and algorithm detection speed reaches 89.9 FPS,meeting the requirements of real-time detection.Compared to other models,the improved YOLOv8 has certain advantages in detecting typical defect in substations.关键词
YOLOv8/变电站典型缺陷检测/多尺度注意力机制/Inner_MPDIoUKey words
YOLOv8/typical defect detection in substations/multi-scale attention mechanism/Inner_MPDIoU分类
信息技术与安全科学引用本文复制引用
吴宇浩,朱文忠,罗缘,袁倩文,王文..基于改进YOLOv8的变电站典型缺陷检测算法[J].四川轻化工大学学报（自然科学版）,2025,38(3):54-64,11.基金项目
四川省科技研发重点项目(2023YFS0371) （2023YFS0371）
四川省科技创新(苗子工程)培育项目(2022049) （苗子工程）
企业信息化与物联网测控技术四川省高校重点实验室基金项目(2022WYY03) （2022WYY03）
四川轻化工大学研究生创新基金项目(Y2024120) （Y2024120）
