摘要
给出了一种基于遗传优化算法的虚拟立体校准方法。利用红外发光二极管随三坐标测量机的测头做定间距移动构造了一个立体校准模板,通过奇异值分解算法初始估计校准参数,最后,通过遗传算法优化求解所有校准参数。实验中校准点的空间位置反推误差小于0.04 mm,表明该方法能够解决像机内参数精确校准问题,可以满足大尺寸视觉坐标测量技术对摄像机内参数精确校准的要求。
A dummy solid camera calibration method based on genetic algorithm was presented. Using the infrared LE moved with the CMM probe according to some distance,the solid calibration template was formed to estimate the initialization calibration parameters by the Singular Value Decomposition （SVD） algorithm. Finally, the genetic algorithm was used to optimize and determine all calibration parameters. The experimental results show that the inverse result error of calibration point is less than 0.04 mm, and the calibration method can resolve the problem of high-precision camera calibration, and also can meet the requirement of high-precision camera calibration in the field of large scale vision coordinate measurement.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2006年第3期485-489,共5页
Optics and Precision Engineering
基金
国家自然科学基金资助项目(NO.50275040)
哈尔滨工业大学校科研基金资助项目(HIT.2002.17)
关键词
大尺寸视觉测量
像机校准
畸变修正
遗传算法
立体校准
large scale vision measurement
camera calibration
distortions correction
genetic algorithm
solid calibration
