2D Discrete Cosine Transform for fusion of Multitemporal Satellite Images

This paper analyses the quantitative performance metrics on applying image fusion method on satellite images and compares it with the fusion methods when applied to filtered images. The satellite images are obtained from over long distances and are immensely huge in size. During transmission, the satellite images get distorted by the presence of different noise components in the atmosphere. In this paper, satellite image fusion is done using Discrete Cosine Transform and compared with the image fusion after filtering operation on satellite images. The results are compared by calculating the Peak Signal-to-Noise Ratio, Mean Square Error and Entropy. From the results, it can be inferred that the image fusion on filtered images gives improved results in comparison to fusion on unfiltered images.