现代电子技术2025，Vol.48Issue(22)：81-87,7.DOI:10.16652/j.issn.1004-373x.2025.22.015
基于改进YOLOv8s的螺栓缺销目标检测方法
Method of pin-losing bolt object detection based on improved YOLOv8s
摘要
Abstract
In allusion to problem of low accuracy in pin-losing bolt fault detection caused by the small target proportion and complex background in the bolt images collected by drones,a transmission line pin-losing bolt fault detection algorithm based on improved YOLOv8s is proposed.The target detection layer is improved based on YOLOv8s,and the model's ability to capture small target features is improved by increasing the upsampling multiple.The collaborative attention module is constructed to emphasize the scale information of bolt targets while diminishing background information,thereby further enhancing the capability of target detection in complex backgrounds.The C2f_DASI module is designed to enrich the semantic information of target features and enhance the network's ability to perform nonlinear mapping of complex features,as well as improve the overall feature representation capability.The deconvolution module is introduced to adjust the size of the convolutional kernels,so as to reduce the computational complexity of the model.The experimental results indicate that in comparison with the original model,the average accuracy of the improved model is improved by 5.8%,and the average accuracy,recall rate and accuracy ofpin-losing bolt fault detection are improved by 11.6%,6.0%and 4.1%respectively,which can effectively detect the pin-losing bolt fault of transmission line.关键词
输电线路/螺栓缺销/故障检测/YOLOv8s/协同注意力/C2f_DASI/逆卷积Key words
transmission line/pin-losing bolt/fault detection/YOLOv8s/collaborative attention/C2f_DASI/deconvolution分类
电子信息工程引用本文复制引用
邵罗,吐松江·卡日,依马木·艾山,周远翔,马小晶..基于改进YOLOv8s的螺栓缺销目标检测方法[J].现代电子技术,2025,48(22):81-87,7.基金项目
国家自然科学基金项目(52067021) （52067021）
新疆维吾尔自治区自然科学基金面上项目(2022D01C35) （2022D01C35）
新疆维吾尔自治区优秀青年科技人才培养项目(2019Q012) （2019Q012）
