Expert AI for green

Next June, the European parliament will discuss the new EU Nature restoration Law, as part of the EU Biodiversity Strategy for 2030. This proposal reflects many challenges and values close to Sobolt, such as the preservation and restoration of endangered ecosystems. Current nature monitoring and management practices still require laborious field work. Despite the romanticism of working outside and being one with nature, these traditional approaches are often slow.

Luckily, artificial intelligence has been found to be a key player for high impact nature monitoring solutions. In the last decade, significant applications were developed that rely on satellite borne and airborne data. This allows unprecedented insights of large natural areas through a simple laptop. However, these techniques are not without limitations, hindering their ability for addressing all facets of adequate nature management. A new study from the European Forest Institute highlights the need for new technologies in order to turn into reality the 2030 Agenda for Sustainable Development.

What do these technologies look like? Well, the direction of research appears to be moving towards combining into unique solutions different sources of data. In this year 2022 IEEE GRSS data fusion contest a specific trak named BNI: Brave New Ideas was included to push researchers to be creative and find new ways to combine together data that might seem to not share much. An example is combining satellite images and Light Detection and Ranging (LiDAR) to leverage the best from both worlds. Satellite data is optimal due its increasing accessibility and high resolution, allowing frequent estimation of vegetation indices on large areas. However, due to the nature of 2D images, they fail to account for volumetric data. On the other hand, LiDAR is unmatched for generating volumetric maps, but is known to be noisy, and lacks the rich information provided by cameras and hyperspectral sensors. By combining them together, we can create Digital Elevation Models (DEM) (see picture below), not only cool 3D visualisations but really powerful tools to measure how much our forests are growing and how they are changing. Moreover, DEM can also be further combined with yet different data sources, like heatmaps, or map conveying statistics like species distribution, to further expand the range of predictions and trends we can generate! Add to the recipe the decreasing cost of technical hardware like unmanned aerial vehicles (Drones for friends) with their ability to fly and gather data even underneath tree crowns, and it becomes clear the above mentioned limitations of artificial intelligence will soon be a memory.

The practical challenge of course is to have the resources & skills to gather and process all these different data. Sobolt has experience in tackling & embracing solutions that use satellite, drone imagery and LiDAR. Leveraging these techniques for optimal impact has been in our DNA since 2017. In other words, ready to move to the next step. Stay alert for the hottest green solutions or book a demo for an early glimpse.

Comments are closed.