苏州科技学院学报（自然科学版）Issue(4)：72-76,5.
基于薄板样条的遥感影像非刚性配准方法
Non-rigid registration of remote sensing images based on thin-plate splines
摘要
Abstract
This paper has proposed a non-rigid registration method for remote sensing images based on Thin-plate Splines. Firstly, feature points were extracted from reference images and images to be registered by SIFT algorithm. Then, initial registration was carried out by intensity correlation algorithm and error matches were e-liminated by using two-step random sample consensus (RANSAC) from coarse level to fine level. Finally, via building a non-rigid registration model with matching points, registration image was transformed. Experimental results show that the algorithm has good robustness and high practicability under the condition of temporal differ-ence and geometry deformation.关键词
薄板样条/SIFT算法/非刚性配准/随机采样一致性Key words
thin-plate spline/scale invariant feature transform (SIFT) algorithm/non-rigid registration/random sample consensus (RANSAC)分类
信息技术与安全科学引用本文复制引用
杨朝辉..基于薄板样条的遥感影像非刚性配准方法[J].苏州科技学院学报（自然科学版）,2014,(4):72-76,5.基金项目
江苏省高校自然科学研究资助项目 （）
