Automatic Haze Removal in Remote Sensing Images for Enhanced Environmental Monitoring and Analysis

Haze is a serious issue for satellite imaging because it obscures crucial features and lowers image quality, which ultimately compromises the accuracy of remote sensing. This work uses paired datasets to compare Inception Pix2Pix and CycleGAN for dehazing satellite images. The outcomes show that Inception Pix2Pix works better than CycleGAN, restoring image quality with an SSIM of 0.942 and a PSNR of 22.73 dB for thin haze. Better noise reduction and structural preservation are facilitated by its better feature extraction and adversarial training approach. According to these results, Inception Pix2Pix is a more successful method for dehazing satellite images, which helps remote sensing applications including environmental monitoring, urban planning, and disaster relief.