Abstract: Aiming at the narrow positioning range and poor flexibility of AGV, a solution is proposed and verified experimentally. A red single ball is used as a marker of the AGV car body, and a plurality of cameras are constructed based on the global vision positioning to guide the autotransport vehicle. Red color single ball recognition, multi - visual field space expansion and coordinate position mapping are realized by using OpenCV and QT platform, and the real - time coordinate information is transmitted to lower - level AGV through WiFi. Automatic transport vehicles to calculate the transport path, and the use of time-sharing dual feedback closed-loop algorithm to solve the problem of blocking the transport process markers. The experimental results show that the positioning scheme based on the red single-ball marker has the advantages of large positioning range and strong applicability.

Key words: global vision positioning, automatic transport vehicle, multi-visual coordination, double feedback closed loop

摘要： 针对自动运输车（AGV）定位范围窄、灵活性差问题提出解决方案并实验验证。采用红色单球作为AGV车体的标志物，基于全局视觉定位技术构建起多个摄像头协同引导自动运输车。利用OpenCV和QT平台实现红色单球识别、多目视场空间扩展、坐标位置映射计算，并将实时坐标信息通过WiFi发送给下位机AGV。自动运输车根据坐标信息计算出运输路径，并且采用分时双反馈闭环算法解决运输过程中标志物被遮挡的问题。实验结果表明：基于红色单球标志物的多目视觉导引自动运输车定位方案具有定位范围大、灵活性好、适用性强的优势。

关键词: 全局视觉定位, 自动运输车, 多目协同, 双反馈闭环