Fusion of optical and radar satellite data and AI support for monitoring of clear-cuttings

Purpose and goal:

Among its tasks, the Swedish Forest Agency (SFA) dates clear cuttings and monitors the regrowth. Today, optical satellite data is used. During winter and cloudy periods, data can be hard to get.

The project will use radar satellite data to support the monitoring. It will also investigate how radar data reacts to felled deciduous trees. In optical imagery, these areas can appear to be felled when defoliated. The project will provide the SFA with a prototype for this. Finally, the project will adapt the prototype to foreign vegetation types, aiming for a world market of forestry entities.

Expected results and effects:

The expected result is increasing SFA’s precision in dating cuttings and plantings using radar satellite data, as well as understanding of how radar data sees felled deciduous forest, which, in optical imagery from defoliated periods, looks different than when leafed.

A successful project will lead to the SFA gainfully using a prototype from Spacemetric and Qamcom. It will also lead to Spacemetric and Qamcom being able to start preparing for marketing and selling this prototype in a version adapted to other countries, aiming for the world market of national forestry entities.

Approach and implementation:

The SFA will formulate requirements and test the prototype solution, giving user experience feedback, allowing the project consortium to develop it into a product that SFA can use in its day-to-day work. Spacemetric will lead the work and contribute its image data processing and management expertise.

Qamcom Research and Technology will develop the machine learning (ML) algorithms for analysing the combined satellite and radar data, acting as a computational back end in the end-to-end system. Spacemetric and Qamcom will include the results in its marketing and sales efforts.