基于特征点的图像配准与拼接技术研究

图像配准与拼接成功的关键在于特征点提取和匹配。针对传统的sift算法提取特征点计算量大、耗时长的问题,文章采用改进后的surf算法先对两幅图像的重叠部分提取特征点,使用bbf（best-bin-first）查找初始匹配对后采用ransac算法剔除误匹配并计算两幅图的变换关系,最后采用无缝拼接技术实现了图像的平滑镶嵌。

Study on Image Registration and Mosaic Technologhy Based on Surf Feature

The key points of the success of image registration and mosaic are feature extraction and matching.As the traditional sift algorithm used to extract feature points need large computation and is time-consuming,an improved surf algorithm was used to extract feature points in the overlap region between the registered images and bbf algorithm was used to find initial matching pairs.Then,the transforming matrix H was computed with RANSAC algorithm after wrong pairs were removed.In the end,smooth mosaic was reali...