Tracasa multiplies by four the resolution of the images of the Sentinel 2A satellites with deep learning techniques

The public company of the Goverment of Navarra improves all the metrics previously obtained in this field with SENX4, a new enabling technology that is at the cutting edge of superresolution

Tracasa, public company of the Goverment of Navarra, has multiplied by four the resolution of the images of the Sentinel 2A satellites with deep learning techniques. Tracasa has reached this international milestone with SENX4, an enabling technology that improves all the metrics previously obtained in this field and that is at the cutting edge of superresolution. Here you can find more about SENX4, in Tracasa´s website, in the section Innovative stories.

“Sentinel 2A satellites of the European Space Agency offer multispectral images with a spatial resolution of 10 meters per pixel, which sometimes can limit their uses. In Tracasa, with our SENX4 model we have improved the RGB and NIR bands resolution, going from 10 meters per pixel to 2,5 meters per pixel, with a 4X factor” , explains Carlos Aranda, Tracasa Research and Innovation manager.

To achieve this result, Tracasa has taken a new step in the field of artificial intelligence and has used deep learning techniques in photogrammetry. These techniques use deep learning algorithms to copy and improve the human capacity to process and recognize patterns in images.

Aranda says that in order to increase the resolution of the images “we have not done postprocessing actions, such as color adjustments or sharpening (a technique that is used to remark the details of the images but spoils its spectral information). Our model only uses artificial intelligence to learn the patterns of the data and get an image as real as possible”.

Taking as a reference the combination of PSNR and SSIM indexes, a measurement system widely used in the analysis of images, the images obtained with SENX4 model offer an important improvement in the metrics compared to the images obtained with traditional interpolation models, achieving, as a final result, more defined images, clearer and more adjust to reality.

This milestone in the field of images superresolution opens the door of its aplication in different sectors. “SENX4 model is a technological tool that was unthinkable just a few years ago. From this moment, its use can be really relevant, for example, in the generation of maps related to topics such as the environment, emergency care or security”, explains Aranda.

Innovative stories: a meeting point in Tracasa´s website

In the section Innovative stories, Tracasa is going to present its most cutting-edge investigations, those tools that have a disruptive technology and that can have an important development in the next years. “Innovation and motivation to find new paths in the technological development are part of Tracasa´s DNA. We want that our progresses can converge in a single meeting point: Innovative stories”, Aranda says.

In this way, Tracasa plans to provide new innovative approaches in subjects such as artificial intelligence and data science, automated analytical models of remote sensing and geoscience, specialized digital models of the territory and 3D cities, smart cities platforms and digital twins (Internet of Things) and smart mapping in the field of mobility.

 In the last years, Tracasa has made a firm commitment to develop the use of artificial intelligence. In this sense, the public company of the Goverment of Navarra has developed, in collaboration with Tracasa Instrumental, 16 projects related to artificial intelligence since 2017, 11 of them in the last year. Besides, Tracasa is part of the IA Technology Capabilities Map in Spain, and has recently incorporated its services to the Digital Innovation Hub promoted by the Goverment of Navarra.

SENX4: At the cutting edge of superresolution technology

WHAT DOES SENX4 OFFER?

When we talk about extracting information of images from multispectral sensors deployed on satellites, their spatial resolution is one of the main factors that will define their future uses. The ESA constellation of Sentinel 2A satellites offers to the scientific community multispectral images every five days in an open way. But one of the drawbacks of these images is their spatial resolution (10 meters per pixel), limiting their future applications.

There are a lot of methods to increase the spatial resolution of the images. Some solutions try to improve the 13 bands of each satellite, taking all of them to 10 meters using the bands with the máximum resolution. Other solutions develop models with 2X and 4X scaling factors from the 13 bands, thinking that is possible to use this factor to take the bands from 10 meters to 2,5 meters. But nobody has yet created a model to transform the 10 meters bands into 2,5 meters with a contrast of real bands.

In Tracasa, with our SENX4 model we have improved the RGB and NIR bands resolution, going from 10 meters to 2,5 meters with the 4X factor.

HOW HAVE WE GOT IT?

Until now, to increase the resolution were used the traditional technics of interpolation and resampling: linear interpolation, bilinear, cubic and bicubic. In our case, to increase the resolution in a 4X factor, we have used deep learning technics and with cutting-edge convolutional neural networks we increase the resolution of the images without including unwanted patterns, improving the traditional methods.

We have not used sharpening or color adjustments. We avoid generating very synthetic images, far from the reality of the sensor. Our model only uses artificial intelligence to learn the patterns of the data and get an image as real as possible.

 

VALUING THE IMPROVEMENT

To interpret the results we have used a measuring system that is frequently used in the analysis of images: the PSNR and SSIM combination. These measures are indicators of the quality of the results based on an objective image to achieve. The higher the value in the metric, the better the image will be adjusted to the real result. Taking bicubic like the baseline to improve in the traditional models of interpolation, we have achieved an important improvement in the metrics, that offers, as a result, a more defined image, clearer and more adjusted to reality.

 

 

EXAMPLES

 

 

USE CASES

Semantic segmentation (Streets and buildings)

With the new image SENX4 (2,5 meters) we can have a muticlass semantic segmentation more correct and accurate. We improve the segmentation metrics that we got with the data of Sentinel 2A 10 meters.

Clustering algorithms

With SENX4 resolution we optimize the results that we get with clustering algorithms. We add more precision in the spatial limits of the clusters and in their representative average.