An Automatic and High-precision Registration Method Based on Point Features for Multi-source Remote Sensing Images
基于点特征的多源遥感影像高精度自动配准方法

This paper proposes an automatic and high-precision registration method based on point features for multi-source remote sensing images.The proposed method is a two-step process including pre-registration and fine-tuning registration.Firstly,the method detects the matching points by the Scale Invariant Feature Transform(SIFT) algorithm and the input image is pre-registered by using linear polynomial model.As a result,the input image is transformed to the same spatial pixel size and the reference coordinate system as the reference image.After a large number of feature are detected based on the Harris corner detector in the image pre-registered,tie point pairs are found in a small search scope which is identified in the reference image by correlation coefficient.Tie point pairs with large errors are pruned by Baarda data snooping method.Finally,the images are divided into a number of triangular regions by constructing the Triangulated Irregular Network(TIN),and the high precise rectification and registration can be fulfilled on each triangular region one by one.Experiments demonstrate that the method achieve precise registration of multi-source remote sensing images.