根据上海城轨交通接触网的现状及特点，阐述应用机器视觉理论研究开发专用非接触式接触网检测系统的关键技术，主要包括弓网关系、国内外检测技术现状、采用的技术路线、方案具体实施等，内容涉及高速数字成像、图像识别模型，图像边缘处理、几何光学布置设计、里程定位设计、多任务多进程软件设计。结果表明，接触网监测产品作为轨道交通接触网的专业检测装置，可适应国内轨道交通的接触网现状，同时利用自主研发的专业技术，可摆脱对国外进口设备的依赖，并能节约成本。

The paper elaborates the application of machine vision theory for the development of the key technologies of non-contact overhead contact line detection system based on the present situation of Shanghai urban rail transit overhead contact line and its characteristics． The key technologies include pantograph -net relationship，domestic and foreign inspection technology status， technical route and plan implementation，etc． ，which are related to high-speed digital imaging，image recognition model，image edge processing，geometric optics，odometer positioning design，and design of multi-task and multi-process software． Application practice indicated overhead contact net detection system is adaptable to the present situation of rail transit，self-reliant and can save cost．