摘要
首先采用内核投影算法Walsh-Hadamard降低了生成的SIFT特征描述子的特征维度;然后在基于距离相似度度量基础上加入方向约束,减少误配率;最后使用粒子群算法对搜索策略进行优化,减少算法耗时.实验结果表明,该改进算法有效提高了图像匹配准确率.
Firstly,the kernel projection algorithm Walsh-Hadamard is used to reduce the SIFT feature descriptor.Then based on the measure of distance similarity,the direction constraint is added to reduce the mismatch rate.Finally the PSO algorithm is used to optimize the search strategy to reduce the time-consuming of the algorithm.Experimental results show that the improved algorithm can effectively improve the accuracy of image matching.
作者
陈文华
岳雅
余本国
CHEN Wen-hua;YUE Ya;YU Ben-guo(School of Science, North University of China, Taiyuan 030051, China;Aerospace Power Test Technology Institute of Xi＇an,Xi＇an 710100,China)
出处
《云南师范大学学报（自然科学版）》
2018年第2期56-59,共4页
Journal of Yunnan Normal University:Natural Sciences Edition
基金
山西省教育厅教改计划资助项目(127/11011904)
关键词
图像匹配
SIFT
内核投影
粒子群
Image matching
Scale invariant feature transform（SIFT）
Kernel projection
Particleswarm optimization
