摘要
给出了基于度量误差模型的摄像机定标方案,即由最小二乘法计算投影矩阵,从投影矩阵中分解出参数作为进一步计算的初始值,然后考虑投影矩阵中约束条件及图像坐标的误差结构,由度量误差模型优化参数。结果表明,该方案不仅保证了旋转矩阵的正交性,同时提高了定标精度。
The camera-calibrating plan based on error-in-measure models is provided in this paper. The projective matrix is calculated with the Least Square Algorithm. The projective matrix is decomposed and the results become the initial value for further calculating. Considering the restricted condition in projective matrix and the error construct of the image coordinates, the parameters are optimized by error-in-measure models. The results prove that this plan not only ensures the rotating matrix 's orthogonality but also improves the calibrating precision.
出处
《计算机应用研究》
CSCD
北大核心
2004年第5期12-14,共3页
Application Research of Computers
基金
国家自然科学基金(30271079)
中国林科院基金(32-9901)
关键词
摄像机定标
立体视觉
度量误差模型
Camera Calibration
Stereo Vision
Error- in- measure Models
