机器人视觉下的视差图效果优化研究  

Research on the Effect Optimization of Parallax Map in Robot Vision

作　　者：张京军 刘松松[2] 方智立 高瑞贞 ZHANG Jing-jun;LIU Song-song;FANG Zhi-li;GAO Rui-zhen(School of Mechanical and Equipment Engineering,Hebei University of Engineering,Hebei Handan 056000,China;School of Information and Electrical Engineering,Hebei University of Engineering,Hebei Handan 056000,China)

机构地区：[1]河北工程大学机械与装备工程学院,河北邯郸056000 [2]河北工程大学信息与电气工程学院,河北邯郸056000

出　　处：《机械设计与制造》2023年第12期258-262,共5页Machinery Design & Manufacture

基　　金：河北省高校科技攻关项目(ZD2018207)。

摘　　要：立体匹配(Stereo Matching)算法的效果在机器人应用中至关重要。为了提高立体匹配算法效果,首先利用基于一阶差分的Sobel检测、Scharr检测和Canny检测对图像进行梯度处理,获取图像边缘,其次利用图像的边缘信息作为初始匹配代价进行立体匹配,大幅减少不必要的计算量,然后利用所提算法分别对Middlebury数据集中的图像和自采集的图像进行实验,获得立体匹配的误匹配率和匹配时间。经过实验数据对比,Sobel检测算法和BM(Boyer-Moore)算法结合的方法所得视差图速度快且精度高,效果最佳,实验证明所提方法具有明显优势,可提高立体匹配效果。The effect of the stereo matching algorithm is very important in robot applications.In order to improve the effect of the stereo matching algorithm.First,use Sobel detection based on first-order difference,Scharr detection and Canny detection to perform gradient processing on the image to obtain the edge of the image.Secondly,use the edge information of the image as the initial matching cost for stereo matching,which greatly reduces the amount of calculation necessary.Then use the method proposed in this paper to conduct experiments on the images in the Middlebury data set and self-collected images to obtain the mismatch rate and the calculation time of stereo matching.Through the comparison of experimental data,the combination of Sobel detection algorithm and BM(Boyer-Moore)algorithm can get the best disparity map with high speed and accuracy.The experimental results that the proposed method has obvious advantages and can improve the stereo matching effect.

关 键 词：机器视觉 边缘信息 双目视觉 误匹配率 视差图 立体匹配 

分 类 号：TH16[机械工程—机械制造及自动化] TP242[自动化与计算机技术—检测技术与自动化装置]