SAR Despeckling Using a Denoising Diffusion Probabilistic Model

Speckle is a type of multiplicative noise that affects all coherent imaging modalities including Synthetic Aperture Radar (SAR) images. The presence speckle degrades the image quality and can adversely affect performance SAR applications such as automatic target recognition change detection. Thus, despeckling an important problem in remote sensing. In this paper, we introduce SAR-DDPM, denoising diffusion probabilistic model for despeckling. proposed method employs Markov chain transforms clean images to white Gaussian by successively adding random noise. despeckled obtained through reverse process predicts added iteratively, using predictor conditioned on speckled image. Additionally, propose new inference strategy based cycle spinning improve performance. Our experiments both synthetic real demonstrate leads significant improvements quantitative qualitative results over state-of-the-art methods. code available at: https://github.com/malshaV/SAR_DDPM.