Multi-scale image fusion using Contourlet Transform with spectral factorization

Multi-modularity fusion is the most popular technique in the field of image fusion. Image fusion is the process which combines the relevant information from a pair of multi-scale images. This work demonstrates fusion for Multi-Modularity images like infrared and visible light sensor images. Firstly preprocessing had done by Gaussian filter for noise removal. Subsequently Contourlet Transform decomposition using a double filter bank structure along with spectral factorization has performed with a specific decomposition level. Here edge detection using Laplacian filter and target extraction processed in terms of watershed transform. The extracted sub-band coefficient is fused based on averaging operation for approximation image and choose max fusion rule for detail image and a inverse transform is performed. This fusion framework investigated for sensor images yields avoidance of reconstruction errors and ghost effect. Experimental results in comparison with existing method show better performance and robustness.