The scene is captured by 2D cameras and those images are run through our proprietary AI technology. The images are combined into a 3D understanding of the scene, where single objects are identified as pickable.
The robot is given pose estimation and path planning information through the Apera Vue software embedded in its controller. A safe path that avoids collisions is planned. Because the object is trained into an AI neural network using CAD drawings or 3D scans, these steps can happen very quickly.
The object is picked and placed with precise and accurate results.