Research On Image Fusion Algorithm Based On Combined Features

Image fusion algorithm based on combined features is mostly researched in this paper.Various of image features and their extraction algorithm and multi-scale geometric transformation theory are firstly analyzed, which has verified each features for the high frequency of image obtained effective from the contourlet and wavelet transformation. Since the single feature just only could reflect one attribute of the image, as the parameter index under the frequency domain coefficients fusion rules, it is difficult to get a comprehensive as well as ideal result. In respect to this problem, the image fusion algorithm based on the combined features is presented in this paper. choosing different features in various image fusion. Finally, it has brought the PSO to get the best result by means of combining weights of iterative optimization.The fusion algorithm presented could be choose different features combination for different purpose of image fusion. As for the multi-focus image, to get the sharper result image is the purpose for the multi-focus image,choose the average gradient and fractal feature. While as for the medical image, it need to get the abundant also accurate bone tissue and soft tissue, choose the entropy of information and fractal feature. As for the remote sensing image, the spectral information and the reservation for the detailed information are considered by selecting the texture feature and the average gradient.By comparison with the other two algorithm just using one feature fusion, the experiment result for the algorithm in this paper has shown that the medical, multi-focus and remote sensing image has got an improvement for its decisive evaluation parameter. In solving the image fusion problem based on feature made more significant results, further improve the quality of fusion image, it is a kind of image fusion algorithm with great advantage.