快速的圆投影图像匹配算法  

Fast image matching algorithm based on circular projection

作　　者：曹田 李勃[1] 任福继[2,3] 董蓉[4] CAO Tian;LI Bo;REN Fuji;DONG Rong(School of Electronic Science and Engineering,Nanjing University,Nanjing 210046,China;School of Computer and Information Science,Hefei University of Technology,Hefei 230009,China;Department of Information Science and Intelligent Systems,Tokushima University,Tokushima 7708500,Japan;School of Electronics and Information,Nantong University,Nantong 226019,China)

机构地区：[1]南京大学电子科学与工程学院,江苏南京210046 [2]合肥工业大学计算机与信息学院,安徽合肥230009 [3]德岛大学智能信息工学部,日本德岛7708500 [4]南通大学电子与信息学院,江苏南通226019

出　　处：《智能系统学报》2020年第1期84-91,共8页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金深圳联合基金重点项目(U1613217).

摘　　要：针对现有圆投影匹配方法计算复杂度高、对同质区域无法识别的缺点,提出了一种新的图像匹配算法。该算法基于混合圆投影向量,结合塔式分解和角度直方图估计,不仅可以识别出模板在待匹配图像中的准确位置,还可以通过角度估计策略得到模板的旋转角度。通过图像金字塔策略,结合混合圆投影向量快速找到候选点;然后在诸多候选点中精确定位,确定位置;最后通过角度直方图计算出准确的角度。实验结果证明该算法识别率高,且匹配速度快。In view of the high computational complexity and the incapability to recognize homogeneous regions of existing circular projection matching algorithms,a new image matching algorithm is proposed in this study.On the basis of hybrid circular projection and combined with pyramid decomposition and angle histogram estimation,the proposed algorithm can not only identify the exact position of the template in the image to be matched but also obtain the rotation angle of the template through the angle estimation strategy.First,candidate points are identified by the image pyramid strategy combined with the hybrid circular projection vector.Then,the positions of the candidate points are precisely located and determined.Finally,the exact angle is calculated by the angle histogram algorithm.The experimental results show the high recognition rate and fast matching speed of the proposed algorithm.

关 键 词：圆投影 模板匹配 图像匹配 图像金字塔 角度直方图 混合圆投影 顶层局部聚类 非极大值抑制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]