化工矿物与加工2023，Vol.52Issue(12)：71-75,5.DOI:10.16283/j.cnki.hgkwyjg.2023.12.011
基于FREAK特征的多视角影像匹配算法研究
Research on multi-view image matching algorithm based on FREAK feature
摘要
Abstract
The image registration algorithm based on features is widely used for its stability and efficiency.The extraction of feature points is affected by factors such as different lighting conditions,scale changes and rotation angles.When the accuracy rate of the same feature points is low,it will directly affect the normal operation of the sub-sequent image mosaic,target recognition and so on.Based on the above description,a new image matching algorithm based on FREAK was proposed to solve the problem of image matching effect with different rotation angles.First,the SIFT algorithm was used to detect the point feature.Then,we used the FREAK descriptor to descript the detected points.And then used the nearest neighbor and the second nearest neighbor distance ratio of the obtained feature points for rough registration.Finally,the RANSAC algorithm was used in precise registration to get the best matching result.In order to verify the effectiveness of the proposed algorithm,the SURF algorithm,the SURF-FREAK algo-rithm and the operation results of the algorithm mentioned in this paper are compared based on the image data of two different rotation angles.The experimental results show that the algorithm has high accuracy and higher precise fea-tured points,the image matching with different rotation angles has high stability.关键词
影像匹配/旋转影像/FREAK描述子/RANSAC算法/SURF算法Key words
image registration/rotated image/FREAK descriptor/RANSAC algorithm/SURF algorithm分类
天文与地球科学引用本文复制引用
杨阳,王瑞,宋伟..基于FREAK特征的多视角影像匹配算法研究[J].化工矿物与加工,2023,52(12):71-75,5.基金项目
国家自然科学基金项目(52364018). （52364018）
