Rotation and Translation Invariant Template Matching Based on Edge Points for Remotely Sensed Imagery from Two Sources

【摘要】：正An edge based rotation invariant algorithm(EBRI) is suggested to estimate the arbitrary rotation angle between images to be matched.Based on the basic geometric invariants and a way of a voting scheme,rotation angle is estimated by edge points located in the correspondence rings selected. Simulating annealing algorithm is applied as the search strategy. By computer simulations with IRS-C(pan) satellite images as the input images,and SPOT(pan) images as the reference images,the performance of the method proposed is analyzed.It is found that the proposed method can significantly improve the success rate of matching for the images with arbitrary rotation, translation and large gray aberrance.The method may be used for position estimation of aircraft,and can also be used for object recognition,image retrieval,and motion tracking.