Abstract: When the metro train is operating under FAO mode, obstruction detection system is an effective auxiliary device . This paper proposes a method of balancing among scale-invariant, rotate-invariant and system real-time, so as to ensure machine vision system to efficiently complete image grayscale matching, which can be applied to the image analysis of obstruction detection device.

Key words: FAO, grayscale matching, obstruction detection, object detection, machine vision

摘要： 当地铁FAO系统运行时，障碍物检测系统是有效的辅助设备。提出一种在尺度不变性、旋转不变性和实时性之间权衡，以确保机器视觉系统有效进行图像灰度匹配的方法，可以应用于障碍物检测装置图像分析。

关键词: 全自动运行, 灰度匹配, 障碍物检测, 目标检测, 机器视觉