Add to Chrome
Add to Firefox
Add to EdgeOptical Image-to-Image Translation Using Denoising Diffusion Models: Heterogeneous Change Detection as a Use Case
Apr 17, 2024
We introduce an innovative deep learning-based method that uses a denoising diffusion-based model to translate low-resolution images to high-resolution ones from different optical sensors while preserving the contents and avoiding undesired artifacts. The proposed method is trained and tested on a large and diverse data set of paired Sentinel-II and Planet Dove images. We show that it can solve serious image generation issues observed when the popular classifier-free guided Denoising Diffusion Implicit Model (DDIM) framework is used in the task of Image-to-Image Translation of multi-sensor optical remote sensing images and that it can generate large images with highly consistent patches, both in colors and in features. Moreover, we demonstrate how our method improves heterogeneous change detection results in two urban areas: Beirut, Lebanon, and Austin, USA. Our contributions are: i) a new training and testing algorithm based on denoising diffusion models for optical image translation; ii) a comprehensive image quality evaluation and ablation study; iii) a comparison with the classifier-free guided DDIM framework; and iv) change detection experiments on heterogeneous data.
