IMAGE SUPERRESOLUTION VIA SRGANS FOR LOWRESOLUTION SATELLITE IMAGERY

This paper presented an SRGANbased image superresolution framework for enhancing lowresolution satellite imagery. The proposed method effectively reconstructs highfrequency details and improves visual realism compared to traditional methods.   Future work includes integrating transformerbased attention modules, incorporating multispectral satellite bands, exploring diffusionbased superresolution models, and optimizing the model for deployment on edge computing devices in remote sensing systems. This proctoring is 90% accuracy.

[1]    Y. Chen et al., “Enhanced SRGAN for Satellite Image SuperResolution,” IEEE Transactions on Geoscience and Remote Sensing, 2023. [2]   M. Kumar et al., “MultiScale Adversarial Network for Remote Sensing Image SuperResolution,” IEEE Access, 2024.   [3]   L. Zhang et al., “AttentionBased GAN for HighResolution    Satellite Image Reconstruction,” IEEE Transactions on Neural Networks and Learning Systems, 2024.   [4]      R. Patel et al., “Lightweight GAN Architectures for RealTime Remote Sensing SuperResolution,” IEEE Internet of Things Journal, 2025.   [5]     H. Lee et al., “Spectral Consistency Optimization in GANBased Remote Sensing Image Enhancement,” IEEE Transactions on Image Processing, 2025.   [6]   A. Sharma et al., “Evaluation Metrics for Remote Sensing Image SuperResolution,” IEEE Access, 2026. [7]  T. Nguyen et al., “Deep Residual Learning for Satellite Image Enhancement,” IEEE Access, 2023.   [8]    P. Singh et al., “Adversarial Learning Framework for HighResolution Geospatial Imaging,” IEEE Transactions on Artificial Intelligence, 2024.   [9]  K. Roy et al., “Perceptual Loss Optimization in Remote Sensing SuperResolution,” IEEE Access, 2025.   [10]     J. Brown et al., “Efficient Training Strategies for GANBased Image Reconstruction,” IEEE Access, 2026.