基于卷积神经网络立体匹配的变电站三维信息感知方法分析  

Analysis of 3D Information Perception Method for Substation Based on CNN Stereo Matching

作　　者：刘金柱 刘军 王青 闫海洋 李智恒 李凌云 LIU Jinzhu;LIU Jun;WANG Qing;YAN Haiyang;LI Zhiheng;LI Lingyun(Henan Power Transmission and Transformation Construction Co.,Ltd.,Henan 450051,China;Nanjing Steen Smart Energy Technology Co.,Ltd.,Jiangsu 211899,China)

机构地区：[1]河南送变电建设有限公司,河南450051 [2]南京斯泰恩智慧能源技术有限公司,江苏211899

出　　处：《电子技术（上海）》2024年第6期88-89,共2页Electronic Technology

摘　　要：阐述针对变电站复杂多变的环境,提出基于卷积神经网络立体匹配的变电站三维信息感知方法,包括构建变电站三维信息实时感知框架,提出基于双重注意力机制的立体匹配模型。This paper describes a three-dimensional information perception method for substations based on convolutional neural network stereo matching,targeting the complex and ever-changing environment of substations.The method includes constructing a real-time perception framework for substation three-dimensional information and proposing a stereo matching model based on dual attention mechanism.

关 键 词：双目立体匹配 PSMNet 三维信息感知 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP391[自动化与计算机技术—控制科学与工程]