SAR IMAGE COLORIZATION USING GENERATIVE ADVERSARIAL NETWORKS (PIX2PIX)

Read full text Summarise with VIRAT
Cited by: 0
References: 7

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Synthetic Aperture Radar (SAR) imagery plays a crucial role in remote sensing due to its ability to capture data under all weather and lighting conditions.However, SAR images are grayscale and contain speckle noise, making them difficult to interpret by human analysts and standard computer vision systems.This research focuses on translating SAR images into optical-like RGB images using Generative Adversarial Networks (GANs), specifically the Pix2Pix Conditional GAN architecture.The proposed system utilizes a U-Net generator with skip connections and a PatchGAN discriminator to perform paired image-to-image translation.The model is trained on the SEN1-2 dataset to learn the mapping between SAR and optical domains.Experimental results demonstrate improved visual quality, significant speckle noise reduction, and enhanced interpretability, with the proposed model achieving a PSNR of 28.92 dB and SSIM of 0.91.The study concludes that GAN-based approaches effectively bridge the modality gap between SAR and optical imagery, making them highly valuable for disaster management and environmental monitoring applications.

Fields of Study & Subjects

Generative Adversarial Networks and Image Synthesis Advanced Image Fusion Techniques Advanced SAR Imaging Techniques

Publication details

Journal / source
International Research Journal of Modernization in Engineering Technology and Science
Publication date
31 March 2026
Language
EN