SpaceNed verbindt
de Nederlandse ruimtevaartsector

Developing new applications for satellite data

In the last decades various earth observation instruments have found their way in orbit. This brings a phenomenal amount of data that can be exploited in applications that help us to monitor our living environment. S[&]T has over 20 years of experience in processing and analysing Earth observation data. Based on this, we have developed a wide range of remote monitoring and inspection applications. At S[&]T we specialize in combining data from all types of satellites, (multispectral) optical, radar and atmospheric, using advanced data science techniques.

This knowledge is currently being successfully used by some of the S[&]T spinoffs, such as Orbital Eye and Sensar. However, at S[&]T we are always looking for new applications of our knowledge. One of the ways we do that is by participating in the Small Business Innovation Research (SBIR) programme of the Dutch government. In this programme, the government challenges companies to come up with innovative solutions to societal problems.

In the last year we have successfully worked on a number of these SBIRs, together with our partner 52impact. In all cases we used a combination of radar, multispectral and high resolution satellite data and machine learning algorithms. One example of such projects is MEANDER (which stands for MutatiE ANalyse, Detectie En Registratie), which focusses on detecting changes in and around waterways.

Detecting changes near water ways

Water management is crucial for quality of life in the Netherlands and access to clean water. Rijkswaterstaat (Directorate-General for Public Works and Water Management) and the Dutch water boards play an important role in the proper management of water systems. Good information provision is necessary so that Rijkswaterstaat and water boards can intervene in time when there are relevant changes in a water system, and effective policy can be developed. Currently, most information for surveillance and enforcement is obtained through visual inspections on site. This is a time-consuming and therefore expensive activity, which means that it is not possible to register all necessary mutations in the water system in time. In addition, the supervisory area is increasing, while it is expected that capacity will decrease due to staff turnover.

In order to support Rijkswaterstaat and water boards in carrying out these inspections more efficiently, we have developed the concept for MEANDER. MEANDER is a service that automatically detects and classifies changes in and around water systems.

In the first phase of the SBIR, we successfully showed that changes in and around water ways can be detected and classified using a combination of radar and multispectral data. In addition we demonstrated that using multiple data sources yields a more reliable service. Since radar data is always available, because it is not dependent on weather conditions, MEANDER can provide reliable and timely information.

Our solution was selected as one of the two to continue on to the next phase of the SBIR, where we will develop our concept into a service.

CAPTION: Overview of open satellite data (Sentinel-2, Sentinel-1 change detection) and Superview high resolution images (obtained from the Satellietdataportaal) covering the area around an embankment who’s presence changes with the seasons. The radar based change detection (in red) can be pin pointed in time using Sentinel-2 time series, which clearly show 2 transitions (early April and November). High resolution imagery (50 cm resolution) confirm that these transitions took place.