一种基于特征点匹配的图像拼接方法研究

针对图像间因具有旋转及光线强度差异等现象而导致的拼接效果不佳及拼接速度慢的问题,提出一种基于特征点的配准算法;在特征点提取阶段,尺度不变的特征变换方法 (SIFT)具有对图像尺度缩放、旋转、放射变换以及亮度变化保持不变的优点,文章采用了改进的SIFT特征点提取算法;在特征点匹配阶段,采用改进的RANSAC算法对特征点匹配对提纯;最后用加权平均法实现拼接图像的融合;实验证明,该算法有效提高了图像拼接的效率和准确性,拼接精度可以达到亚像素级。

An Image Mosaic Research Based on Feature Points Matching

In order to solve the problem of poor effect in image mosaic caused by the rotation,and the differences in light intensity,an image registration algorithm based on the feature points is proposed in this paper.Scale Invariant Feature Transform(SIFT) has advantages of scale invariant,rotating invariant,affine transform invariant and intensity invariant,an improved SIFT feature extraction algorithm was used in the feature extraction stage;During the matching phase,an improved RANSAC algorithm was used for feature point matching purification;Finally,the weighted average method to achieve fusion of the mosaic image.Experiments show that the algorithm effectively improves the efficiency and accuracy of the mosaic image,and its accuracy can reach sub-pixel level.