摘要
图像配准中易受到外界干扰发生变形或者部分特征缺失,为此提出融合特征点匹配和互信息的连续形变图像配准方法。提取连续形变图像特征,对提取到的特征实施匹配处理,结合互信息算法统计两图像之间的相似程度,得出最佳配准指标,即像素灰度值为最大值,结合特征点匹配结果以及互信息算法生成图像配准刚性指标,实现连续形变图像配准。实验结果表明,所提方法配准后图像信息完整,配准效果较好,且匹配耗时较短,说明其能够降低图像配准中干扰因素的影响,有效提升图像配准质量。
At present,image registration is easy to be deformed by external interference.Therefore,a method to register continuous deformation images based on feature point matching and mutual information was proposed.Firstly,we extracted the features of continuous deformation images and then matched these features.Secondly,we calculated the similarity between the two images by using the mutual information algorithm,thus obtaining the best registration index,that is,the pixel gray value is the maximum value.Finally,the registration result was combined with the mutual information algorithm to generate a rigidity index of image registration.Thus,the continuous deformation image registration was achieved.Experimental results show that the proposed method has complete image information after registration,and shorter matching time,indicating that it can reduce the influence of interference factors and effectively improve the quality of image registration.
作者
陆杨
杨茂云
LU Yang;YANG Mao-yun(School of Information and Electrical Engineering,Xuzhou University of Technology,Xuzhou Jiangsu 221018,China;School of Wisdom Education,Jiangsu Normal University,Xuzhou Jiangsu 221116,China)
出处
《计算机仿真》
北大核心
2023年第8期182-185,共4页
Computer Simulation
关键词
特征点匹配
互信息
图像配准
自动尺度选择方法
相似度
Feature point matching
Mutual information
Image registration
Automatic scale selection
Similarity
