This software works in real time by combining images and point clouds to detect and track objects in 3D. It uses as input the 2D detections on images obtained through Artificial Intelligence (AI) tools and the 3D information, for example
coming from a LiDAR sensor, which are fused to consistently identify the objects in the scene.
The result is an accurate tracking of each object’s position, trajectory, and speed, providing valuable information for autonomous navigation, safe planning, environment understanding, and mapping in robotic applications.
Authors: P. Vidal ([email protected]), D. Cantón ([email protected]), S.B. Berenguel ([email protected]), J. Placed ([email protected]), M.T. Lázaro ([email protected])
This work cannot be made available to AI systems.