中文摘要：

针对摄像机自标定方法大多需要求解一个多元高维的非线性方程组,其求解过程非常困难,为此提出一种基于遗传算法的摄像机自标定方法,该方法通过二幅或多幅图像的多个对应匹配特征点,利用二个视图之间的极几何关系,建立某些相关约束,最后得到一个代价函数;然后,通过使用Matlab遗传算法工具箱来求解该函数的最小值,进而一次性求出摄像机所有内、外参数。因为该方法不必用到摄像机的外部信息,所以可推广到序列图像中的变参数情况下的摄像机自标定。实验结果表明,该方法简单、有效、快速,可用于摄像机自标定。

英文摘要：

Generally speaking, the process of camera self-calibration is very difficult because camera self-calibration often requires solving non-linearity of the multi-dimension equations. Therefore, a new self-calibration method for camera is introduced based on genetic algorithm. The target function is gotten by building relative constraints, thus the all internal and external parameters of the camera is acquired which used matlab, by solving the minimum value of the target function from several characteristic points of two or more images. Because the method needs not the external information of the camera, it can be used in sequences of images. Experimental results show that the proposed method is simple, effective and fast, so the method can become a tool for camera calibration.