|
遥感学报 2010
Multi-source remote sensing image matching based on contourlet transform and Tsallis entropy
|
Abstract:
There are a lot of differences in multi-source remote sensing images from various sensors about the same scene. Maximization of mutual information can be used for the multi-source image matching, but the accuracy and efficiency of image matching need to be further improved. Therefore, an algorithm for multi-source remote sensing image matching was proposed in this paper, based on contourlet transform, Tsallis entropy based mutual information and improved particle swarm optimization. Firstly, the target image and reference image were decomposed to the low resolution image using contourlet transform, respectively. Then, a new image similarity measure criterion, the Tsallis entropy based mutual information, was used to achieve the global optimization. Meanwhile, a modified extremum disturbed and simple particle swarm optimization algorithm was applied to match the lowest resolution remote sensing images. Based on the preliminary result, the matching between the higher resolution images could be implemented stepwise up to the full resolution images. The experimental results show that, compared with those of other existing remote sensing image matching methods, the proposed algorithm has the high accuracy, strong robustness and requires much fewer operations.