校准
计算机科学
计算机视觉
人工智能
摄像机切除
光学
帧(网络)
投影(关系代数)
物理
数学
算法
统计
电信
作者
Zhuoyi Yin,Xiangyun Ren,Yifang Du,Fang Yuan,Xiaoyuan He,Fujun Yang
出处
期刊:Applied Optics
[Optica Publishing Group]
日期:2022-01-26
卷期号:61 (6): 1475-1475
被引量:14
摘要
Camera calibration is crucial to the application of binocular measurement. However, many of the calibrations are carried out manually, so shaking of the calibration plate cannot be avoided. When the cameras are not synchronized or the exposure time is inconsistent, the image acquisition time of different cameras does not correspond strictly, which will introduce inevitable calibration errors. In this paper, two consecutive frames are collected instead of one frame, and the calibration data are corrected by the image acquisition time value returned by the camera. Experiments show that the proposed method can greatly reduce the re-projection error of calibration, and achieve higher calibration accuracy and less image quantity dependence.
科研通智能强力驱动
Strongly Powered by AbleSci AI