基于改进SURF算法的双目视觉测量系统研究

Research on Binocular Stereo Measurement System Based on Improved SURF Algorithm

摘要:

针对双目立体测量中所使用的图像匹配算法精度不高的问题,提出了一种基于SURF算法的改进图像匹配算法并运用到双目立体测量系统中。首先运用SURF算法检测和描述图像对特征点、特征向量；其次采用双向特征向量匹配策略对匹配点集进行初始过滤；最后采用PROSAC算法根据极线约束几何模型对初始过滤后的匹配点集进行二次筛选,得到最终的优质匹配集合。实验表明,改进的SURF算法具有更高的匹配准确性和较好的算法实时性,运用改进SURF匹配算法的双目测量系统可实现更准确的定位测距。

Abstract:

Aiming at the problem that the SURF image matching algorithm used in binocular stereo measurement is not accurate enough, an improved image matching algorithm based on SURF algorithm is proposed and applied to the binocular distance measurement system. Firstly, SURF algorithm is used to detect and describe feature points and feature vectors. Secondly, reverse feature vectors matching strategy is used to initialize the matching points set. Finally, the PROSAC algorithm is used to select the initial matching point set according to the polar geometric model, and get the final good match set. Experiments show that the improved SURF algorithm has higher matching accuracy and good real-time performance, and the binocular measuring system with improved SURF matching algorithm can achieve More accurate positioning。