计算机视觉
人工智能
计算机立体视觉
立体视觉
双眼视觉
计算机科学
立体摄像机
立体摄像机
摄像机切除
匹配(统计)
校准
双眼视差
机器视觉
数学
统计
作者
Wenyong Lu,Yan Shi,Jin Zhan,Weijie Luo,Jing Chen,Jiajie He
摘要
Binocular stereo vision technology has good application prospects and practical value in industrial production, autonomous driving, quality inspection, and many other aspects, and is an important research direction in computer vision. Strain measurement based on binocular stereo vision is different from traditional strain measurement methods, and has advantages such as convenient operation, high accuracy, and global dynamic measurement. This article focuses on the strain measurement technology of binocular stereo vision, and studies the camera calibration, stereo matching, 3D reconstruction, and SGBM algorithm in stereo matching of binocular cameras. By constructing a binocular camera strain testing system, the distance between the corresponding point AB before and after the strain of the target object is measured, thereby achieving the measurement of the strain of the target object. The strain of the target object obtained from the experimental test results is 47.53%, the theoretical tensile strain is 48.54%, and the relative error is controlled within ± 2%.
科研通智能强力驱动
Strongly Powered by AbleSci AI