基于改进特征点检测的快速图像配准算法

摘要

In view of slow efficiency and false detection on feature point of traditional feature point registration algorithm,propose an algorithm for image registration which is based on feature point detection.Detection method of feature points is improved,use gray relationship between pixels and surrounding pixels to filter non-feature points.Next,sets up a set of feature points,after detect remained points accurately.Iterative closest point(ICP) algorithm is used to match set of feature points. The experimental results show that the proposed algorithm significantly improves detecting accuracy and detecting time,and has good registration effect.%针对传统特征点配准算法效率过慢、对特征点存在误检的现象,提出了一种基于特征点检测的图像配准算法.对特征点检测方法进行了改进,利用像素点与周围像素点的灰度关系滤除非特征点;对剩余的点使用提出的菱形模版进行精确检测,建立了特征点集合;利用迭代最近点(ICP)算法对特征点集合进行配准.实验结果表明:改进算法在特征点检测准确性和检测时间上明显提高,并且具有良好配准效果.