About |
Physics Guided Machine Learning for Satellite Imagery Analysis at Scale
Our technology builds on decades of research on land cover change detection across multiple sectors. The algorithms are not just data driven but also incorporate physics laws which makes them more powerful than traditional black box machine learning techniques.
Proven. Algorithms have been validated through extensively using non-trivial reference datasets.
Robust. The use of physical principles make the algorithms much more robust to atmospheric disturbances such as clouds, shadows, and aerosols that are major issues in satellite imagery analysis.
Patented. Our technology portfolio consists of patented algorithms that span multiple disciplines such as water, agriculture, forestry and urbanization.
Seamless. Our processing pipelines do all the heavy lifting and produce relevant physical quantities that are easy to integrate in existing workflows.