微型电脑应用2025，Vol.41Issue(4)：69-72,4.
基于递归神经网络的变电站巡检机器人避障路径自动识别方法
Automatic Recognition Method for Obstacle Avoidance Path of Substation Inspection Robot Based on Recursive Neural Network
摘要
Abstract
To address the issues of inaccurate automatic obstacle avoidance path recognition and long recognition time for ro-bots,this study proposes an automatic obstacle avoidance path recognition method for substation inspection robots based on a recursive neural network.The method captures real-time images of the inspection path via a camera and processes them using a biphasic standard deviation filtering method combined with cascaded morphological filtering.The time-based backpropagation algorithm is employed to handle the temporal sequence of the network,optimizing the recursive neural network by deriving hi-erarchical feedforward local gradients.The distance information of obstacles detected by sensors is input into the recursive neu-ral network to achieve automatic obstacle avoidance path recognition for substation inspection robots.Test results demonstrate that the proposed method significantly improves both the recognition accuracy and efficiency of automatic obstacle avoidance path recognition for substation inspection robots.关键词
递归神经网络/变电站巡检机器人/避障路径/自动识别Key words
recursive neural network/substation inspection robot/obstacle avoidance path/automatic identification分类
信息技术与安全科学引用本文复制引用
黄炜昭,谢欢欢,辛拓,张宏钊,陈龙,何维..基于递归神经网络的变电站巡检机器人避障路径自动识别方法[J].微型电脑应用,2025,41(4):69-72,4.基金项目
深圳供电局有限公司(090000KK52200150) （090000KK52200150）
