【摘要】 为解决设备自主获取目标距离信息并控制测距时间在毫秒级别的问题,提出一种基于视差的几何模型立体视觉测距方法.首先,通过对标定过程的改进,利用MATLAB获得了较高精度的摄像机参数;然后,利用Bouguet算法和BM算法进行立体校正和立体匹配;最后,利用视差几何模型去除干扰区域获得清晰完整的视差图以实现对目标的标记和测距.实验表明:该方法测距时间为181. 59 ms,且距离小于1 800 mm时,误差最大不超过2%,满足了精度和实时性要求.更多还原

【Abstract】 To achieve autonomous distance measurement and control the distance measurement time in milliseconds,a method of stereo vision ranging based on disparity is proposed. First,through the improvement of the calibration process,the higher precision camera parameters is obtained by MATLAB. Then,the Bouguet algorithm and the BM algorithm are used for stereo correction and stereo matching respectively. Finally,the disparity geometry model is used to remove the interference area to obtain clear and complete disparity map and achieve the target mark and ranging. The experiments show that the ranging time of the method is 181. 59 ms,and the error is less than 2% when the distance is less than 1800 mm,it meets the requirements of precision and real-time.更多还原