基于改进ORB算法的双目视觉定位测量方法  

Location and Measurement Method of Binocular Vision Based on Improved ORB Algorithm

作　　者：杨宇 许四祥[1] 方建中 蔡永祯 YANG Yu;XU Sixiang;FANG Jianzhong;CAI Yongzhen(School of Mechanical Engineering,Anhui University of Technology,Ma’anshan Anhui 243000,China)

机构地区：[1]安徽工业大学机械工程学院

出　　处：《传感技术学报》2019年第11期1694-1699,共6页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金项目(51374007)

摘　　要：针对基于传统特征点检测的双目视觉测量中匹配时间长、误匹配率高和测量精度低的问题,提出了基于改进ORB算法的双目视觉定位测量方法。首先对特征点邻域内的像素点灰度值进行加权,再采用灰度质心法提取特征点的主方向,最后以字符串描述子代替传统二进制描述子对特征点进行描述。在测量中采用二维二次函数精确拟合特征点的亚像素坐标。实验结果表明,该改进方法在特征点匹配正确率和测量精度上都有很大的提升,特征点匹配正确率提升了31.8%,测量最低误差达到0.42%,满足双目视觉测量的精度要求。Aiming at the problem of long matching time,high false matching rate and low measurement accuracy in binocular vision measurement based on traditional feature point detection,a binocular vision positioning measurement method based on improved ORB algorithm was proposed.Firstly,the gray value of the pixel in the neighborhood of the feature point was weighted,then the main direction of the feature points were extracted by the gray centroid method.Finally,the character description was replaced by the string descriptor instead of the traditional binary descriptor.The two-dimensional quadratic function was used to fit the feature points sub-pixel coordinates in the measurement.The experimental results show that the improved method has a great improvement in the matching rate and measurement accuracy of feature points.The correct matching rate of feature points is improved by 31.8%,and the minimum error of measurement is 0.42%,which meets the accuracy requirements of binocular vision measurement.

关 键 词：双目视觉 改进ORB算法 视觉测量 亚像素角点 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置] TP212.1[自动化与计算机技术—控制科学与工程]