A Novel Approach to Satellite Image Fusion Based on Half Up and Down Motion Blur

Abstract

Fusion of digital images is an important phase in digital image processing. It combines information from multiple images into a single image, improving quality and clarity. This paper simulates motion blur with different block sizes of 3, 5, 7, 9, and 11 pixels. The point spread function was programmed to produce linear motion blur. Motion blur simulation was applied to half of the image, in two parts (the top and bottom halves). The recommended techniques in this study use weights defined by standard deviations to integrate corrupted grayscale images, unlike conventional approaches that use addition and multiplication. The suggested approaches employ a color conversion procedure from the RGB (red, green, and blue) color system to the HSV (hue, saturation, and value) color space to combine the top and bottom halves of colored blur images. For the upper, lower, and merged images, added, distorted, and typical edges in the original image were detected. Different thresholds were used to visually compare edges. The standard deviation-based fusion technique was found to be the most effective method for preserving and enhancing edge information among the evaluated methods, getting a 50% improvement in performance over the multiplicative and additive methods.