基于鲁棒方法的图像整合技术

ROBUST BASED IMAGE CONFORMITY

  • 摘要: 本文讨论了基于圆序列的整合图像问题.由于相机旋转拍摄图片,可避免光源变化、遮挡和外观变化等出现在一般匹配中的问题,因此自动生成匹配点要容易得多.我们采用一种鲁棒的方法,只要输入少量点,就可以自动寻找匹配点对并对投影变换进行求解.整合参数的变化,可得到不同视点下的整合图像.该方法不需要事先知道摄象机的内、外参数.

     

    Abstract: This paper discusses circle images based image conformity. Because of finding matched points between images taken from the same point,it may avoid the problems of lighting changes, occlusion, and aspect changes,than the general point matching problem. A robust way is used in the paper. A few of matching points are used to automatically find more matching points. Projective transformations are solved with the robust way and optimized. When the conformity coefficient is changed, the conformity image corresponding to different view can be obtain.The camera needn't to be calibrated.

     

/

返回文章
返回