A Hybrid Method to Enhance Low-Resolution Images for Smart Applications

Low-resolution images often lack fine detail and are prone to blurring, noise, and distortions, making accurate assessment challenging. This issue becomes even more challenging for applications, such as land mapping, environmental monitoring, disaster response, and military surveillance. In such situations, enhanced image resolution is essential for precise analysis and decision-making. This paper addresses these challenges and proposes a method to improve the quality of low-resolution images. We introduce an image resolution enhancement technique that employs the Discrete Wavelet Transform (DWT) to decompose the image into four frequency bands. These bands are then processed by a deep learning network, which uses generative adversarial networks (GANs) to generate high-resolution bands. Finally, the Inverse Discrete Wavelet Transform (IDWT) is applied to reconstruct the high-resolution image. This integrated method effectively reconstructs high-resolution images from their low-resolution counterparts and preserves critical information. We validated the efficacy of our approach through extensive simulations on two publicly available datasets. The results demonstrate that our method consistently outperforms several state-of-the-art techniques in terms of PSNR, SSIM, FSIM, and UIQI. These performance metrics highlight the potential of our developed method applications including satellite images and building smart cities.