Registration of Remote Sensing Images Based on

Image registration is one of the critical techniques in multi-sensor image fusion. Classical image registration method using similarity measures can only find Registration Control Point (RCP) on discrete points on images, which makes it difficult to get higher registration precision. This paper presents an image registration method based on Gaussian fitting matching, which takes two dimensional Gaussian fit matching on measurement matrix produced by classical search method of image RCPs. This method has the potential of getting more accurate position of RCPs. Experiment shows that using our method can obtain higher registration accuracy than that of classical method. I. I NTRODUCTION

[1]  Venu Madhav Govindu,et al.  Alignment Using Distributions of Local Geometric Properties , 1999, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Taejung Kim,et al.  Automatic satellite image registration by combination of matching and random sample consensus , 2003, IEEE Trans. Geosci. Remote. Sens..

[3]  Sang Wook Lee,et al.  ICP Registration Using Invariant Features , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  Jordi Inglada,et al.  On the possibility of automatic multisensor image registration , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[5]  R. Kwok,et al.  Automated Multisensor Registration: Requirements And Techniques , 1990, 10th Annual International Symposium on Geoscience and Remote Sensing.

[6]  Morgan McGuire,et al.  Techniques for multiresolution image registration in the presence of occlusions , 2000, IEEE Trans. Geosci. Remote. Sens..