A Hybrid Method for Block-Based Feature Level Image Fusion Technique of PAN and MS Image Using Contourlet Transform with Neural Network

Multisensor image fusion is one of the widely researched areas in the field of image processing. The highest spatial content can be achieved while preserving the spectral resolution of an image with spatial information of a high resolution panchromatic (PAN) image combines with spectral information of a low resolution multispectral image (MS). This paper derived and renovates block-based feature level contourlet transform with neural network (BFCN) mannequin for image fusion. The proposed BFCN mannequin combines Contourlet Transform (CT) with neural network (NN), which performs a vital role in feature extraction and detection in machine learning applications. In the designed BFCN mannequin, the two fusion techniques, Contourlet Transform (CT) and neural network (NN) discussed for fusing the IRS-1D images using LISS III scanner about the places Patancheru, Mahaboobnagar Hyderabad and Vishakhapatnam in Andhra Pradesh, India. Also, Landsat 7 image data and QuickBird image data are used to carry out experiments on The designed BFCN mannequin. The features like contrast visibility, spatial frequency, the energy of gradient, variance, and edge information are studied. Since the learning capability of NN makes it feasible, Feedforward back propagation neural network is trained and tested for classification. The trained NN is then used to fuse the pair of PAN and MS images. The designed BFCN mannequin is analogized with other techniques to evaluate the quality of the fused image. The designed BFCN mannequin is an efficient and feasible algorithm for image fusion, and It is clearly shown that from the experimental results.