基于视觉引导的机器人自主爬楼

Self-determination Climbing Stairs of Mobile Robot Based on Visual Navigation

采用了一系列方法从机器人视觉图像中提取出可供电机调速的信号。先对视觉图像关键部位进行图像分割;随后用灰度直方图中谷底信号作阈值,区分出台阶水平面与竖直面;再用形态学方法消除小斑点;然后采用选取中间值的方法得到台阶斜率;最后用提取出的台阶棱线斜率来控制机器人左右轮的速度,达到调整移动机器人姿态的目的。该方法具有重复性好、实时性强、易使用的特点。

Adjusting-velocity signal of electromotor is picked up from robot visual picture with a serial of methods. Firstly the key place of visual picture is divided up and valley signal in gray straight-square picture is regarded as threshold, in order to differentiate horizontal and upright surface. Secondly small blobs are cleared up with morphological method and the nearest step slope is selected on basis of middle value. Finally according to step slope, velocities of robot left and right wheels are controlled, for adjusting mobile robot pose. The method is characteristic of good repetition, real-time and easy use.