救灾机器人目标导航中的视觉识别技术研究

Research on Visual Recognition Technology in Target Navigation of Disaster Rescue Robot

笔者介绍了某型救灾机器人工作过程中视觉导航的要求,提出了采用张正友标定方法进行摄像机参数标定和目标环境下圆形旋转手轮的识别方法,详细介绍了利用MATLAB标定工具箱进行摄像机内参数和外参数的标定过程,最后采用改进的椭圆识别算法对手轮识别方法进行了验证,实验证明了该方法的有效性。

This paper introduces the requirements of visual navigation for a type of disaster rescue robot,puts forward the method of using Zhang Zhengyou’s calibration to calibrate camera parameters and the method of recognizing circular revolving handwheel in the target environment,introduces in detail the calibration process of camera internal and external parameters by using MATLAB calibration toolbox,and finally uses the improved ellipse recognition algorithm to carry out handwheel recognition method.The validity of the method is verified by experiments.