条纹投射测量标定中摄像机畸变图像的修正  

Distorted image rectification of camera in calibration of fringe projection measurement

作　　者：陆敏杰[1] 郑鹏[2] 陈明仪[1] 

机构地区：[1]上海大学机电工程与自动化学院,上海200072 [2]郑州大学机械工程学院,河南郑州450001

出　　处：《光学仪器》2008年第4期50-54,共5页Optical Instruments

基　　金：国家自然科学基金资助项目(60678036)

摘　　要：在条纹投射测量系统的标定过程中对摄像机畸变图像进行修正,能有效地提高系统的标定精度。采用改进的Harris算子检测标定板上特征点的亚像素坐标,通过预标定获得特征点空间坐标,并应用BP神经网络技术校正畸变图像。仿真实验表明该方法能简单有效地提高了标定系统的精度。Correcting the distorted image of camera in calibration process of fringe projection measurement system that can effectively improve the calibration accuracy of system. In this paper,an improved Harris detector is applied to obtain subpixel coordinates of earmark on the calibration gauge,can obtain the feature space coordinates through pre-calibration, and use BP neural network technology for correcting distorted image. The experimental results show that this simple and effective method can improve the calibration accuracy of system.

关 键 词：条纹投射 摄像机 畸变图像 HARRIS算子 BP神经网络 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]