四川大学学报（自然科学版）2011，Vol.48Issue(4)：822-826,5.DOI:10.3969/j.issn.0490-6756.2011.04.018
一种抗旋转、抗尺度变化的图像匹配算法
An algorithm of rotation and scale invariance for image matching
摘要
Abstract
Based on the study of circular projection,parametric template,and histogram invariant moments ,an image matching algorithm with combination of circular projection vector and circular projection transformation and parametric template was proposed. First of all,the input images were projected with the method of circular projection vector and circular projection transformation which has less computation amount and the character of fixedness to rotation. Then the images were projected into the vector field,then the method of parametric template was used to match images in variable scales to find several potential alternative points ,finally used the image' s histogram invariant moments to do fine match. The method solved the problems of rotation and scale changes between the template images and sensed images,meanwhile used Euclidean distance to match images,avoided the large amount of calculation caused by the traditional method resulted from calculating the correlation function. The experiment shown that the algorithm not only improved the speed of matching,but also achieved good matching results under the circumstance of rotation and scale changes of aimed images.关键词
图像匹配/圆投影矢量和/参数化模板/直方图Key words
image matching/vector summation of circular projection/parameterized model/histogram invariant moments分类
信息技术与安全科学引用本文复制引用
张少蕾,吴晓红,何小海..一种抗旋转、抗尺度变化的图像匹配算法[J].四川大学学报（自然科学版）,2011,48(4):822-826,5.基金项目
中央高校基本科研业务费资助(2009SCU11009) （2009SCU11009）
