快速ACE算法及其在图像拼接中的应用

A Fast ACE Algorithm and Its Application in Image Stitching

为了提高图像拼接过程中图像配准的准确性以及图像融合的拼接缝的平滑性,在图像质量一般的情况下,图像增强是必要的.为此本文提出了在图像预处理阶段进行快速ACE处理,然后通过传统的SURF和RANSAC算法进行图像配准,通过加权平滑方法实现图像无缝拼接的方法.实验结果表明,快速ACE算法可以提高检测到的特征点的数量和提升特征点匹配的准确性,图像拼接结果整体观感更好,图像细节得到保留.

Image enhancement is necessary under general image quality in improving the accuracy of image registration and smoothness of stitching joints in image stitching. In this paper a fast ACE （Automatic Color Enhancement） processing is proposed in the image preprocessing stage. Then, image registration is clone with traditional SURF and RANSAC algorithm. Finally seamless stitching is achieved by weighted smoothing. Experimental results show that the fast ACE algorithm can improve the number of detected fea- ture points and enhance the accuracy of feature points matching, and as a result a better overall impression of the image stitching is achieved and the image details are preserved.