三角形约束下的图像特征点匹配方法  

Image Feature Points Matching Approach with Triangle Constraint

作　　者：吴飞[1] 蔡胜渊[1] 郭同强[1] 庄越挺[1] 

机构地区：[1]浙江大学计算机学院人工智能研究所,杭州310027

出　　处：《计算机辅助设计与图形学学报》2010年第3期503-510,共8页Journal of Computer-Aided Design & Computer Graphics

基　　金：高等学校科技创新工程重大项目培育资金项目(706033);国家自然科学基金重点项目(60533090;60603096);长江学者和创新团队发展计划(IRT0652)

摘　　要：为了在特征点匹配过程中控制错误传播,提出一种基于三角形约束来进行图像特征点匹配的方法.首先检测可能误匹配或者位于变化区域的三角形,然后在后续三角形匹配细分过程中将这些三角形赋予较低优先级,从而可有效地控制错误在匹配过程中传播.实验结果表明,采用该方法能有效地控制匹配过程中的错误传播,提高匹配精度.This paper presents an image matching approach based on triangle constraint which can control the error propagation during the matching process. Firstly, this approach detects the triangles that may have been mismatched or located in changed regions. In the following process, those triangles are assigned with a lower priority; therefore, it is effective to control the error propagation during the matching process. Experiments show that error propagation in the matching process is effectively controlled by this approach and the matching precision is improved.

关 键 词：图像匹配 误差扩散 三角细分 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]