基于BM算法的视觉匹配验证

双目立体匹配算法在非接触测量方面的优异表现，使其在机器人的图像识别领域得到了广泛的应用。文章采用局部立体匹配算法中的字符串搜索（BM）算法，将双目摄像头采集的图像对经过立体匹配、校正后得出深度图，由此得出被测物体与摄像头之间的距离，并与实际测量距离进行比较，以检验算法的准确性。在此基础上改变光照强度和测量距离来检测算法的适用性。实验结果表明，BM算法适用于距离小于1.5 m范围内的物体的测量，且该算法对光照强度要求不高，说明BM算法可以应用到双目视觉系统，用于对实际场景的识别。

Verification of Vision Matching Based on BM Algorithm

The excellent performance of Binocular Stereo Matching Algorithm in non-contact measurement brings about its wide use in the robot image recognition. In this paper, the Boyer Moore (BM) algorithm, which belongs to the local stereo matching algorithms, is used to obtain the depth map of the images collected by the binocular camera after stereo matching and correction,and the distance between the measured object and the camera is acquired. After the comparison between the calculated distance and the actual measured distance,the accuracy of this algorithm could be verified. On this basis,the applicability of the algorithm is further checked by changing the light intensity and the measure distance. The results show that the BM algorithm is suitable for measuring objects within a distance less than 1.5 m, and the algorithm is not so sensitive to light intensity,which means that BM algorithm could be applied to Binocular Vision System for actual scenes recognition.