To automate traditional extraction and classification of features such as waterbodies, roads, forestry, gravel and building outlines using an artificial intelligence deep learning algorithm.
Until now, feature extraction and classification has been achieved using labour and time-intensive processes that must be started from scratch with each new data set. Orbica's goal was to create a solution that would automate the process and improve upon it with each new data set.
Orbica has developed an algorithm that extracts and classifies - with a high degree of accuracy and speed - building outlines, roads, forestry and surface water types from 3-band imagery from any source - it doesn't require multi-spectral. It can adapt to any other natural or man-made feature and has fantastic applications for environmental management, Civil Defence disaster management, change reporting and compliance. Using satellite or drone imagery - rather than traditional aerial imagery - significantly reduces carbon emissions.
This solution saves time and money, and the algorithm improves accuracy with each new data set that it processes.
Orbica won the Thyssenkrupp Drone Analytics Challenge and the People's Choice Award at the Beyond Conventions event in Essen, Germany, in February 2018 with our AI feature extraction and classification solution.
We proposed to automate building progress reporting by flying drones to collect imagery, building 3D point cloud models then using our feature extraction and classification algorithm to identify built features and materials. That then enables a comparison to be made with previous versions to accurately identify progress made.
Orbica is working with Thyssenkrupp to advance our concept for commercial use.