A registration and matching method for remote sensing images

The analysis of a remote sensing image usually requires the comparison of the analyzed image with another image taken from the same spot. This paper describes a general-purpose, representation independent method for registering and matching remote sensing images. The method is based on the iterative closest point (ICP) algorithm. The algorithm requires a procedure to find the closest point on a geometric entity to a given point to measure the required deformation. A dynamic programming method is used for finding the best approximation of the deformation above. The corresponding points are then mapped into each other. A hypothesis propagation and verification method finds the features that match in the two images. An example from the Ottawa area illustrates the procedure