Remote sensing image fusion based on PCA and wavelets

Abstract

Remote sensing, in the recent past, has witnessed continuous developments in the field of environment, agriculture and security. The satellites obtain the information in two domains—spectral resolution and spatial resolution which are highlighted in the Low-Resolution Multi-Spectral (LRMS) and the Panchromatic (PAN) images, respectively. Remote sensing image fusion aims to integrate this complimentary information of the PAN and the LRMS images. In this paper, this has been achieved through Principal Component Analysis (PCA) and Discrete Wavelet Transformation (DWT). The proposed fusion approach involves extraction of the Ist Principal Component (PC) of the LRMS image while simultaneously performing Morphological Hat Transformation on PAN image. The resultant images undergo Discrete Wavelet Transformation to produce approximation (cA) and detail (cD) coefficients. These coefficients are fused using appropriate fusion rules, and the resultant images are synthesized using Inverse Discrete Wavelet Transformation (IDWT) to produce the final fused image. The results have been evaluated which are presented in the later sections.