基于颜色不变量和SURF的彩色图像拼接  

Color image stitching based on color invariant and SURF

作　　者：高盛丰[1] 王正勇[1] 臧媛媛[1] 何小海[1] 

机构地区：[1]四川大学电子信息学院,成都610064

出　　处：《计算机工程与应用》2013年第12期113-115,128,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.11176018);四川省科技支撑计划项目资助项目(No.2010GZ0167)

摘　　要：提出了一种基于颜色不变量和SURF算法相结合的彩色图像拼接方法。该方法利用图像彩色信息计算得到的颜色不变量信息代替灰度信息作为输入,提取图像SURF特征点并进行特征点匹配,根据相似性变换原理对误匹配点进行过滤,提高变换矩阵计算的准确率,采用亮度渐变原则对重叠区域进行像素平滑过渡,实现图像无缝拼接。实验结果表明,该方法在保持算法的快速性和准确性的同时,获得的配准点多而且准确,采用亮度渐变原则拼接增强了对光照变化的鲁棒性,且有效地消除了拼接痕迹,在图像拼接和地质分析领域有一定的实用价值。A method of color image stitching based on color invariant and SURF （Speeded Up Robust Feature） is proposed. The SURF feature points are extracted by the color invariable information calculated from color image in stead of gray information, and then to improve the accuracy of the transformation matrix, the mismatch points are removed in accordance with similarity transformation principle during the match process. To get a realistic seamless image, pixels of the image within the overlap region are smoothly transited by the brightness adjustment principle. Experimental results show that many more match points can be extracted in this method while retaining fast computational speed and high accuracy rather than traditional SURF algorithm. Stitching trace can be eliminated effectively, and also illumination robust has improved with the method mentioned above. It＇ s proved that the algorithm proposed in this paper has certain practical value in the fields of images stitching and geological analysis.

关 键 词：快速鲁棒特征(SURF) 颜色不变量 相似性变换 图像拼接 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]