Advanced Image Registration Method Based on Harris and SIFT Algorithm

Considering the redundant feature points, high computational complexity in image mosaic method based on scale invariant feature transform (SIFT) algorithm, an advanced image registration method is proposed. At first it uses Harris to extract the feature points from two images. The feature points of low quality are canceled. Then SIFT descriptor is utilized to characterize and match those feature points. The voting strategy based on corners' constraint relations is used to delete mismatching points. The advanced PROSAC algorithm is used to compute the transform matrix between the two images. The experiments verify the algorithm's efficiency.