一般位置上的双摄像头立体视觉测距

Distance Measurement from Double Webcams Stereo Vision in General Positions

本文基于一般位置上的双普通摄像头通过立体视觉的方法进行测距研究。为了减少错误匹配点对的个数,先采用Harris角点检测法对双摄像头捕获的两幅图像进行角点检测,然后将检测出的角点作为候选特征点,结合SIFT立体匹配算法实现对应点的匹配,从而进一步进行测距,该方法提高了测距的精度。

With two webcams in general positions, the distance is measured based on stereo vision technology. To avoid too many wrong matching points, Harris corner detection algorithm is used firstly, providing candidate matching points for SIFT algorithm. In such a way the accuracy of the distance measurement is improved obviously.