本文在大量调研国内外的相关算法的基础上,介绍了图像拼接的算法分类,并提出了一种基于灰度似然曲线的寻找匹配点的改进算法。其算法原理为首先进行匹配点的粗匹配,再进行改进后的似然曲线快速拼接,其不但保持了原始算法的精度,而且大大的加快了速度。通过仿真实验,在精度不变的情况下,对应于不同图像,速度较原始方法提高了约60%。

Base on investigate a great many of relevant arithmetic home and abroad.Therefore this paper classify image mosaic arithmetic and give a new arithmetic of image mosaic base on pixels grayscale curve base on pixel.This method do a cursory mosaic base on matching points at first,and then do a new arithmetic of image mosaic base on pixels grayscale curve.It is not only keep veracity similar to essence arithmetic,but also accelerate.In the end this arithmetic is tested in the experiment.The outcome indicate that in the condition that precision is similar to essence arithmetic,the speed increase to about160% than essence arithmetic when the testing pictures are different.