MULTISPECTRAL SUPER RESOLUTION AND IMAGE QUALITY ASSESSMENT COMPARATIVE ANALYSIS

The satellite image resolution alludes to highest accuracy to capture finer details from scene. This paper addresses five different techniques to improve resolution of multispectral satellite image. Our algorithm generates super resolved multispectral image using advantages of Patch Based processing. The results are then compared with four techniques Bicubic Interpolation, Edge Directed Orientation, Patch Based Processing, Gaussian Process Regression (GPR). Comparative analysis is carried out with reference to Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Image Fidelity, correlation coefficient and similarity measure, processing speed and storage space required. Image quality assessments (IQA) parameters are also performed. Super resolution (SR) is commercial algorithm to improve resolution of satellite image when we compare with image fusion