基于SIFT特征和改进融合策略的图像拼接算法

Image Mosaic Algorithms Based on SIFT Features and Improved Fusion Strategy

为了消除图像拼接中的缝隙,本文提出一种改进的图像拼接算法。该算法首先利用经典的SIFT算子提取特征点,接着建立两幅图像的映射关系,利用RANSAC算法进行特征匹配点提纯,最后重叠区域的像素融合采用改进的加权平均算法。实验表明,该算法可以很好地消除重叠区域的拼接缝隙。

In order to eliminate the gap in image mosaic,we propose an improved image mosaic algorithm. The algorithm first uses classical SIFT operator to extract feature points and then set up the mapping relationship between the two images. The match points are purified by using RANSAC algorithm. Finally the pixel of overlapping region is fused by using the improved weighted average algorithm. Experiments show that the algorithm can eliminate the gap of the overlapping area.