top of page

Radar Ground Truth Generation
The Client is a cutting-edge metamaterials agile radar manufacturer with products in the UAV, security, and autonomous vehicle spaces. The Client was interested in automatically generating radar ground truth from video streams gathered from the test vehicle where both the radar and the cameras are mounted. The business goal was to provide a solution to their customers that enables them to develop deep learning detection, tracking, and classification algorithms.
By using state-of-the-art video point cloud reconstruction techniques in concert with deep learning models trained on video streams, we created a service to automatically detect and classify objects in 3 dimensions that serve as ground truth to train the radar algorithms.
bottom of page