Industry 4.0 - Smart production with computer vision, AI and Robotics

This is demonstration of pick and place task implementation where two kinds of objects (tubes and cans), that are randomly distributed in container in multiple layers where picked from the container and sorted in two different boxes by industrial robot. Object detection and recognition is implemented using Artificial Intelligence based computer vision.

The core component of the demonstrator of the proposed system is UR5 collaborative robot from Universal Robots with controller CB3. It consists of 6 rotating joints, payload is 5 kg, reach is 850 mm and repeatability is +/– 0.1 mm. 3D computer vision is based on Intel RealSense D415 camera. It is active IR stereo camera with depth resolution up to 1280 x 720, 90 Hz frame rate, and depth range 0.16-10m. For CV and AI processing algorithms computer with following specs is used: Intel Core i7-8700K 3.70GHz CPU, 16GB DDR4 2666MHz RAM, NVIDIA GeForce GTX 1050 Ti 4 GB GPU.

More information can be found in following publications -

Janis Arents, Ricards Cacurs, and Modris Greitans, “Integration of Computervision and Artificial Intelligence Subsystems with Robot Operating System Based Motion Planning for Industrial Robots”, Automatic Control and Computer Sciences, 2018, Vol. 52, No. 5, pp. 392–401

Elvijs Buls, Roberts Kadiķis, Ričards Cacurs, Jānis Ārents. “Generation of Synthetic Training Data for Object Detection in Piles”, 2018 The 11th International Conference on Machine Vision, November 1-3, 2018 | Munich, Germany

Also, please, check: https://www.trinityrobotics.eu/

Видео Industry 4.0 - Smart production with computer vision, AI and Robotics канала Institute of Electronics and Computer Science, Latvia / Elektronikas un Datorzinātņu Institūts (EDI)