ABSTRACT
Image fusion refers to the process of combining two or more images into single image, integration/composition of the all such image information into a single composite image which is useful for various applications like image interpretation and further studies. The resultant image has higher information content compared to such individual input images being used for fusion. The goal of the fusion process is to evaluate the information at each pixel location in the input images and retain the information from that image which best represents the true scene content or enhances the utility of the fused image for a particular application. Image fusion is a vast discipline in itself, and refers to the fusion of various types of imagery that provide complementary information. In this project work, we propose to study image fusion techniques for satellite remote sensing optical payloads. Due to proliferent information contained in the satellite imagery with temporal resolutions could be composed to create a fused image for better analysis and study applications. Here, we perform such study using machine learning based image fusion techniques on Sentinel imagery using, Principal Component Analysis and Wavelet based image fusion. We focus on the datasets of the Sentinel-2(2 satellites), which is part of Europe's ambitious Copernicus program. We take images of an area that are captured by the Sentinel-2 satellites into consideration and fuse the pixel values of the images in order to retrieve the missing part(s) in the images or make those parts more visible. After the fused image is produced, we perform statistical analysis on the input images and the output images. This analysis is done by calculating mean squared error and structural similarity index between the images. These two statistics are used to compare images. We incorporate these statistics to compare our input and output images. We will further perform visualization using OSG Earth. The main focus of visualization is to give an insight to the area of fused image.