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

作　　者：杨朝辉[1] 

机构地区：[1]苏州科技学院环境科学与工程学院,江苏苏州215009

出　　处：《苏州科技学院学报（自然科学版）》2014年第4期72-76,共5页Journal of Suzhou University of Science and Technology （Natural Science Edition）

基　　金：江苏省高校自然科学研究资助项目(10KJB42000)

摘　　要：提出一种基于薄板样条的遥感影像非刚性配准方法。首先,根据SIFT算法分别在参考影像与待配准影像中提取特征点;然后,对特征点进行匹配,并利用RANSAC一致性分级检验方法,由粗至精分级排除错误匹配点;最后,利用同名匹配点构建薄板样条配准模型,并完成图像配准结果。实验结果表明,文中方法能有效解决遥感影像在时相变化、几何变形等条件下的配准,具有较高的实用性。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 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]