测绘科学技术学报2025，Vol.41Issue(6)：604-611,8.DOI:10.3969/j.issn.1673-6338.2025.06.008
局部多尺度KD-tree与整体Pearson联合约束的异源影像误匹配剔除方法
Local Multi-scale KD-tree and Global Pearson Joint Constraint for Heterogeneous Image Mismatch Removal Method
摘要
Abstract
It is the key link of effective application of heterologous images to use the false match elimination method to automatically,quickly and accurately eliminate mismatches from heterogeneous image feature matching.Based on the existing research foundation of the false match elimination method in image feature matching,a local and global joint mismatch removal method is studied in this paper.Firstly,an initial feature matching set is constructed using the CMM-Net method.Then,a multi-scale KD-tree is used for local constraints to achieve rough purification.Finally,Pearson correlation coefficient is used to combine the distance and angle aspects with a trial removal strat-egy to achieve overall constraints for fine purification.The aim is to achieve false match elimination under the con-dition that the point rate in the initial feature matching set of heterologous images is very low.Through experimental verification and analysis,the performance of the proposed method is compared with the classical RANSAC algo-rithm and LPM algorithm in terms of the accuracy rate,recall rate and F-score of the balance function.The experi-mental results show that the proposed mismatch rejection method has better performance for mismatch rejection of heterologous images,with an accuracy rate about 55％higher than that of the LPM algorithm and about 40％higher than that of the RANSAC algorithm.关键词
异源影像/特征匹配/误匹配剔除/局部一致性/整体一致性Key words
heterogeneous imagery/feature matching/mismatch culling/local consistency/overall consistency分类
天文与地球科学引用本文复制引用
柳鑫,蓝朝桢,纪松,王岩..局部多尺度KD-tree与整体Pearson联合约束的异源影像误匹配剔除方法[J].测绘科学技术学报,2025,41(6):604-611,8.基金项目
国家自然科学基金项目(42371459). （42371459）
