SAR and Optical Image Registration Method Based on Quantum Particle Swarm Optimization

Image registration is the most important step of SAR and optical image fusion. In order to improve the registration accuracy and efficiency of the strategic targets, a new image registration method based on quantum particle swarm optimization is presented. The proposed method consists of three steps in sequence. Firstly, it decomposes the optical image into the high frequency detail components and the low frequency approximation component with the wavelet transform, and extracts feature corner points with Harris corner detection algorithm for the low frequency component. Secondly, it constructs the similarity measure criterion by migrating the pixel feature points to SAR image, and searches the most optimal registration parameter for the SAR image using quantum particle swarm optimization algorithm with individual optimal selection constraint. Finally, it completes affine transform according to the optimal registration parameters, and finishes image registration through interpolation resampling. The experimental results prove that the proposed method is effective and feasible, it can achieve high accuracy of sub-pixel level for the large strategic targets.