家电科技 ›› 2022, Vol. 0 ›› Issue (4): 92-95.doi: 10.19784/j.cnki.issn1672-0172.2022.04.016

• 论文 • 上一篇    下一篇

基于双目视觉的测距系统研究

王浩1, 张善房1, 朱建高1, 张凤生2   

  1. 1.海信家电集团股份有限公司 山东青岛 266000;
    2.青岛大学 山东青岛 266000
  • 出版日期:2022-08-01 发布日期:2022-08-18
  • 作者简介:王浩,硕士。研究方向:测试技术。地址:山东省青岛市崂山区松岭路399号。E-mail:wanghao40@hisense.com。

Research on ranging system based on binocular vision

WANG Hao1, ZHANG Shanfang1, ZHU Jiangao1, ZHANG Fengsheng2   

  1. 1. Hisense Research and Development Center Qingdao 266000;
    2. Qingdao University Qingdao 266000
  • Online:2022-08-01 Published:2022-08-18

摘要: 基于双目视差原理,对双目视觉测距原理进行研究分析,设计双目测距系统并搭建双目测距系统平台。采用张正友相机标定法对系统内的相机进行单目标定获得相机实际焦距及相机畸变系数,进一步对系统进行双目标定获得两相机的相对位置关系。基于标定结果,对采集的图像进行畸变矫正得到更加精确的图像,后提出通过RANSAC法与极线约束法结合的方式对左右相机采集的对应图像进行SURF特征匹配,最终得出被测对象与测距系统的相对位置关系。实验结果表明,在有效测量范围内,设计的测距系统具有较高的测量精度,且在有效测量区域内,中心位置区域测量精度最高。

关键词: 双目视觉, 相机标定, 畸变矫正, 特征匹配

Abstract: Based on the principle of binocular parallax, the principle of binocular visual ranging is studied and analyzed, the binocular ranging system is designed and the binocular ranging system platform is built. Zhang Zhengyou camera calibration method is used to determine the camera in the system with a single target to obtain the actual focal length and distortion coefficient of the camera. Further, the system itself is determined with a double target to obtain the relative position relationship between the two cameras. Based on the calibration results, the distortion of the collected image is corrected to obtain a more accurate image. Then, it is proposed to carry out surf feature matching for the corresponding images collected by the left and right cameras through the combination of RANSAC method and epipolar constraint method. Finally, the relative position relationship between the experimental object and the ranging system is obtained. The experimental results show that the measurement accuracy of the designed ranging system meets the requirements in the effective measurement range, and the measurement accuracy of the central position area is the highest in the effective measurement area.

Key words: Binocular vision, Camera calibration, Distortion correction, Feature matching

中图分类号: