In view of the low efficiency of binocular visual stereo matching, an improvement has thus been made of SURF matching algorithm and camera calibration and calibration. Compared with the traditional visual matching algorithm, SURF matching algorithm is characterized with such advantages as high efficiency and strong anti-interference ability. After the calibration of the camera, the feature points become more accurate, which lays the foundation for an accurate matching.The polar line constraints have been used to improve the SURF algorithm matching rate, with relevant experiments carried out. Experimental results show that the algorithm not only reduces the matching time, but also improves the matching accuracy.