The image fusion methods based on DWT suffers from structural distortions, lack of poor directionality, and lack of shift invariance. This paper extends upon the previous approach and derives ‘An improved block based feature level image fusion technique using multi wavelet transform with neural network’ (BFMN) method for fusing PAN and MS images. This proposed BFMN method integrates ‘Multiwavelet Transform’ (MWT) with the block based concepts of feed forward back propagation neural network for fusing Indian Remote Sensing Satellites (IRS-1D), Landsat-7, Quick Bird images. The present study critically compares the fusion results of BFMN method with other existing methods for fusing PAN and MS images. The analysis of multi wavelet transform is the new development in the area of wavelet transform. The fusion process using multiwavelet fusion technique takes place in multi wavelet space with different frequencies. For this, more number of defined features is incorporated in the fused image. MWT offers simultaneous orthogonality, symmetry, compact support and vanishing moments which are not possible with scalar wavelet transforms. The proposed (BFMN) model integrates MWT with neural network, which is one of the feature extraction or detection machine learning applications. In the proposed BFMN model, the two fusion techniques, multi wavelet transform (MWT) and neural network (NN) are discussed for fusing the IRS-1D images using LISS III scanner about the different locations in Andhra Pradesh, India. Also Quick Bird image data and Landsat 7 image data are used to perform experiments on the proposed BFMN model. The features under study are contrast visibility, spatial frequency, energy of gradient, variance and edge information. Feed forward back propagation neural network is trained and tested for classification since the learning capability of neural network makes it feasible to customize the image fusion process. The trained neural network is then used to fuse the source images. The proposed BFMN model using MWT is compared with other techniques to assess the quality of the fused image. From the Experimental results it is clearly prove that the proposed BFMN model is an efficient and feasible algorithm for image fusion.
Volume 11 | 01-Special Issue
Pages: 1593-1602