SuperSAR is the built-in algorithm of our platform, that makes it possible to allow a constant, ultra-high resolution monitoring of the Earth, in any atmospheric condition and at any time.

A satellite technology that, thanks to artificial intelligence, will transform the way space monitoring happens. An application that can translate images from the SAR domain to the optical one, bringing up to 10 times improvements in resolution. . Let’s see how our algorithm works in detail.

SuperSAR, an algorithm that overcomes all obstacles

SuperSAR was created for many reasons. The first one is to significantly increase the availability of satellite imagery in areas with heavy cloud cover. Cloud cover, in fact, has always been an obstacle to satellite monitoring.

To observe, monitor and analyze phenomena occurring over the territory has always been difficult for satellites such as Copernicus Sentinel-2, which do not have the chance to see through clouds. That’s why areas characterized by the presence of massive cloud formations, for long periods of the year, have always been more challenging to reach.

SAR (Synthetic Aperture Radar) images, acquired by a sensor installed aboard satellites like Copernicus Sentinel-1, have tried to provide a possible solution to the problem. This sensor, in fact, can emit a signal directed toward the Earth’s surface, recording its response overcoming the interferences caused by the presence of the clouds.

The most challenging part, however, is studying these images: they’re not easy to analyze and decode as they have something like a layer of black and white dots.

To cope with this situation, our R&D division – active and vigilant in the field of research and experiments in applying artificial intelligence algorithms to satellite images – identified a technique that would allow a radar image to be “transformed” into an optical image. In this way the content is made intelligible even to the human eye.

Thanks to this technique it was possible to “transform” the source image resolution from 20 m to 10 m. Despite this, its limitations were still visible, as there was no improving effect on spatial resolution.

That’s the exact moment we had the idea to apply our super-resolution algorithm, based on artificial intelligence, to a SAR image, to make its transformation into an optical image more effective, thus bringing its spatial resolution from 10 m to 1 m.

It’s currently being validated and, specifically, it will make it possible to:

  • increase the availability of images in areas characterized by heavy clouds;
  • make monitoring activities for the construction of new buildings or infrastructures, checking their compliance with development plans and authorization processes.

Get in Touch

1 + 12 =