| Multi-spectral image registration is used to match two or more images that get from the same sense, derived from different time, different sensors, different perspectives and different conditions. Multi-spectral image registration is the foundation of image analysis, image processing, and image fusion.Aiming at the problem of the standard mutual information which only uses the pixels intensity but ignores their spatial relationship, leading to poor registration,this paper takes second-order mutual information as similarity measure, and improve the similarity measure. In the process of calculating the entropy of image information, not only to consider each point of the image pixel, but also give full consideration to every point of the pixel neighborhood pixel, thus increasing the stability of the registration, inhibiting the local maxima which is generated by the interpolation in the process of the registration, smoothing curve of the information entropy, so the accuracy of the image registration is improved.Aiming the similarity measure has many local maxima and the initial searching parameters are not given, this paper takes the Ant Colony Algorithm as the optimization algorithm for the strong global search ability and without relying on the initial search points. The transfer criterion of the ants between each layer and global pheromone updating process are improved for the disadvantage that the algorithm may fall into the local maxima, and achieve good results.Considering the strong global search ability of Ant Colony Algorithm, this paper combines the ACA with Powell which has strong ability in local optimization. Taking the advantages of the two kinds of algorithms and abandoning the shortcomings to reduce the searching time and the speed of registration is improved.The experiments prove that the algorithms and methods mentioned in the paper can effectively improve the accuracy and the speed of Multi-spectral image registration. |
PDF Full Text Request