Enhancement of Satellite Images Using Contrast-Limited Adaptive Histogram Equalization With Hybrid Grey Wolf and Particle Swarm Optimization

Improving the quality of Satellite images using an enhancement technique is an important and complex phenomenon. These images are mainly used in applications like urban development, navigation, and climate monitoring, and conducting research with these applications requires a high-quality enhancement of the image to provide more information. In fact, the traditional techniques for image enhancement work less well in areas with high noise or low light levels because they are not designed for striking a balance between contrast improvement and noise amplification. The proposed method is an advancement of the existing method using contrast-limited adaptive histogram equalization with a hybrid metaheuristic approach. The proposed hybrid approach is developed by combining the exploration ability of the original particle swarm optimization with the exploitation refinement of the grey wolf optimizer. The proposed method is validated on satellite images taken from different platforms. The enhanced images obtained by the proposed method are compared with the three popular traditional image enhancement techniques, namely Histogram Equalization, Contrast-Limited Adaptive Histogram Equalization, and Adaptive Gamma Correction with Weighting Distribution, as well as with four different metaheuristic-based methods, such as the original Particle Swarm Optimization, Grey Wolf Optimizer, Arithmetic Optimization Algorithm, and Elk Herd Optimizer. Further, the proposed method is also compared with four different kinds of other recent state-of-the-art methods, which include recent traditional and optimization algorithms, as well as recent deep learning and transformer-based models. The comparison of the proposed method with the other methods is carried out using several image quality metrics such as peak signal-to-noise ratio, structural similarity index, feature similarity index, entropy, mean squared error, etc. The proposed method shows quality improvements in images with a peak signal-to-noise ratio of 34.51 dB, a structural similarity index measure of 0.914, and entropy of 7.76, thus proving its ability in satellite image enhancement. This work impacts tasks relevant to image processing, such as improving the quality of satellite images to provide good visual clarity, in situations like a lack of visual appearance and details of the image, and, therefore, makes it much better to make decisions in higher-level applications involving satellite images. The proposed method, with an approach of automated parameter optimization, is better in performance compared to traditional techniques, as well as with deep learning and transformation-based approaches. It establishes a premise for further and more efficient, and accurate satellite image enhancement techniques.