摘要
深度学习技术的快速发展推动了电力智能安防的自动化进程。电力场景中用于高压电力塔和接触网搭建的复杂钢结构往往成为铁路沿线鸟类筑巢之所,给电力系统安全运行带来了隐患。因此,使用深度学习技术及时发现并清理鸟巢具有重要的实际意义。提出了一种基于改进YOLOv5的鸟巢检测方法,该方法在YOLOv5的基础上考虑了鸟巢本身所独有的黑色属性和错综复杂的纹理特性,采用注意力机制强化鸟巢检测过程中对上述特征的学习。同时,根据电力场景中采集的实际鸟巢数据对该方法开展的验证性实验取得了良好的检测效果,算法检测性能达到88.6%,相比其他经典检测算法高1.5%以上。
The rapid development of deep learning technology promotes the automation of intelligent electrical safety protection.In electrical transmission line,complex steel structure of high-voltage power tower and overhead contact line is often nested by birds along the railway,which brings hidden dangers to the safe operation of power system.Therefore,it is of great practical significance to use deep learning technology to quickly find and help deal with bird nests.In this paper,a nest detection method based on improved YOLOv5 algorithm is proposed.The method takes into account the unique black properties and intricate texture characteristics of bird nest itself,and uses the attention mechanism to enhance the learning of the above features in nest detection process.Experiment conducted on the proposed method using practical bird nest data achieve satisfactory detection results,as the detection performance of the adopted algorithm reaches 88.6%,which is higher than other classical detection algorithms with the lead of at least 1.5%.
作者
刘香萍
罗彩艳
LIU Xiangping;LUO Caiyan(Shenzhen Tagen Pingshan Construction Engineering Co.,Ltd.,Shenzhen 518118,China;Asia Credit Technology(China)Co.,Ltd.,Beijing 100193,China)
出处
《电工技术》
2023年第18期22-25,共4页
Electric Engineering
