{"id":35199,"date":"2025-09-15T13:48:31","date_gmt":"2025-09-15T11:48:31","guid":{"rendered":"https:\/\/nr.no\/en\/?post_type=bc_project&p=35199"},"modified":"2025-09-16T09:39:26","modified_gmt":"2025-09-16T07:39:26","slug":"sharper-satellite-images-with-deep-learning-superai","status":"publish","type":"bc_project","link":"https:\/\/nr.no\/en\/projects\/sharper-satellite-images-with-deep-learning-superai\/","title":{"rendered":"Sharper satellite images with deep learning (SuperAI)"},"content":{"rendered":"\n
\n
\n

SuperAI develops cutting-edge deep learning methods to enhance the resolution of Earth observation data. By combining super-resolution techniques with image fusion, the goal is to improve the spatial detail of Sentinel-2 images, enabling more accurate analysis of the Earth\u2019s surface. These enhancements open new opportunities across areas like agriculture, climate and environment, and urban planning.<\/strong><\/p>\n\n\n\n

\"Arctic
Satellite image of an Arctic region. While not from Sentinel-2 data, it illustrates an area in Earth observation that the SuperAI project aims to enhance.<\/em> Image: USGS\/Unsplash.<\/figcaption><\/figure>\n\n\n\n

Detecting small-scale changes with deep learning<\/strong><\/h2>\n\n\n\n

While satellites like Sentinel-2 already provide valuable information about the Earth\u2019s surface, much of the data is captured at moderate spatial resolution. This limits its use for detecting small-scale changes in landscapes, ecosystems, or infrastructure.<\/p>\n\n\n\n

In SuperAI we are designing a new hybrid approach to overcome this limitation. Rather than building new satellites or sensors, we are using deep learning to enhance existing Sentinel-2 imagery by increasing its spatial resolution to 2.5 metres across all 12 spectral bands.<\/p>\n\n\n\n

This process, known as super-resolution, will generate sharper, more detailed images, making satellite data even more powerful and adaptable than before.<\/p>\n<\/div>\n\n\n\n

\n

To learn more about SuperAI, please contact:<\/h3>\n\n\n\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\n\t\t\t\t\t
\n\t\t\t\t\"\"\n\t\t\t<\/figure>\n\t\t\t\t
\n\t\t\t

Arnt-B\u00f8rre Salberg<\/p>\n\t\t\t\t\t\t\t

Chief Research Scientist<\/p>\n\t\t\t\t\t\t<\/svg>\n\t\t<\/div>\n\t<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\n\n\n

\n

Project: SuperAI<\/p>\n\n\n\n

Funding: The European Space Agency (ESA)<\/p>\n\n\n\n

Period: 2023-2025<\/p>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n

\n
\n
\n

Training AI to improve satellite images<\/strong><\/h2>\n\n\n\n

We use a diffusion model to sharpen the satellite images. The model is trained on a large and freely available dataset of aerial photos from the National Agriculture Imagery Program (NAIP) in the United States and learns to create high-resolution images from images with lower quality.<\/p>\n\n\n\n

However, aerial photos and satellite images have different properties, differing in resolution, sensor type and how they capture the Earth\u2019s surface.<\/p>\n\n\n\n

To bridge this gap, we have developed a method that simulates the NAIP images into ones that more closely resemble data from Sentinel-2. This process, known as image degradation and harmonisation, allows the model to learn patterns that can later be applied to real Sentinel-2 imagery with reliability and precision.<\/p>\n<\/div>\n\n\n\n

\n
\"Comparison
Comparing different methods for sharpening Sentinel-2 images. The top row shows a 60 m atmospheric band, the middle row a natural-colour view, and the bottom row a false-colour composite. Each column shows (from left to right): the original low-resolution input, a traditional sharpening method (GS pansharpening), and our neural network output. The AI-based method produces sharper images that preserve structure and detail, especially in uniform areas like open land or water, and avoids the colour distortions common in traditional techniques. <\/em>Figure: NR.<\/figcaption><\/figure>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n

Enhancing all Sentinel-2 bands with AI-based image fusion<\/strong><\/h2>\n\n\n\n

An important component in our model, is that it improves all 12 spectral bands of Sentinel-2 imagery, not just the RGB (red, green blue) bands, that most earlier approaches focus on. <\/p>\n\n\n\n

To achieve this, we use a neural network-based fusion pipeline that uses the sharpened RGB image as a guide to improve the resolution of the remaining spectral bands. This results in a much clearer and more detailed version of the full 12-band Sentinel-2 image.<\/p>\n\n\n\n

The fusion model is trained in a self-supervised way, which means it doesn’t require high-resolution reference images. Instead, it learns by simulating lower-resolution inputs from real Sentinel-2 images and trains itself to reconstruct the original. This approach is inspired by the Wald protocol, a well-established method for evaluating super-resolution models.<\/p>\n\n\n\n

Why improved resolution matters<\/strong><\/h2>\n\n\n\n

SuperAI enhances multispectral data to make it sharper and more detailed. This supports a wide range of applications, from precision agriculture and land use mapping to climate monitoring and urban planning, where access to reliable, high-resolution satellite information is increasingly important.<\/p>\n","protected":false},"featured_media":35736,"template":"","meta":{"_acf_changed":false,"_trash_the_other_posts":false,"editor_notices":[],"footnotes":""},"class_list":["post-35199","bc_project","type-bc_project","status-publish","has-post-thumbnail"],"acf":[],"_links":{"self":[{"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/bc_project\/35199","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/bc_project"}],"about":[{"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/types\/bc_project"}],"version-history":[{"count":4,"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/bc_project\/35199\/revisions"}],"predecessor-version":[{"id":35764,"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/bc_project\/35199\/revisions\/35764"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/media\/35736"}],"wp:attachment":[{"href":"https:\/\/nr.no\/en\/wp-json\/wp\/v2\/media?parent=35199"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}