摘要
针对SIFT算法在图像融合中耗时长,维度高的问题,论文设计了一种基于SURF、FLANN和RANSAC三者结合的拼接方法。首先利用SURF算法鲁棒性强、算法复杂度低的优势来进行特征点的检测,凭借FLANN算法可以调整参数来进行精确度的提升的优点来进行特征点的匹配,并与常见的BF算法匹配进行比较;针对其中错误匹配对的存在,采用RANSAC算法对存在匹配错误的点进行剔除并进行单应矩阵的计算,来找到最好的模型匹配对,降低误差;最终采用加权平均法进行图像的融合。通过实验验证,算法提高了匹配效率,拼接效果良好。
In view of the problems of SIFT algorithm in image fusion with long time consuming and high dimension,this paper designs a splicing method based on SURF,FLANN and RANSAC.First,SURF algorithm is used for feature point detection with strong robustness and low algorithm complexity.FLANN algorithm can adjust parameters to improve the accuracy of feature point matching,and compare with the common BF algorithm matching.RANSAC algorithm is used to eliminate the points with wrong matching pairs and calculate homography matrix to find the best model matching pairs and reduce the error.Finally,the weighted av⁃erage method is used for image fusion.The experimental results show that the algorithm improves the matching efficiency and achieves good stitching effect.
作者
原伟杰
文中华
彭擎宇
YUAN Weijie;WEN Zhonghua;PENG Qingyu(School of Computer and Communication,Hu'nan Institute of Engineering,Xiangtan 411104;School of Computer Science&School of Cyberspace Science,Xiangtan University,Xiangtan 411105)
出处
《计算机与数字工程》
2022年第1期169-173,185,共6页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号:61272295,61070232)资助。