3D Vision + AI system for automatic depalletizing

NTN-SNR, a subsidiary of the Japanese NTN Group, is a world leader in the manufacture of bearings and power transmission solutions. Based in Annecy, France, the company provides innovative solutions for the automotive, aerospace, rail and energy sectors. It stands out for its quality, technical expertise and commitment to research and sustainable development. As part of this innovative approach, NTN-SNR asked INNODURA to integrate INNOROBOT, our adaptive robot guidance and 3D vision software, into their depalletizing process.

INNODURA has developed an innovative automatic packaging unit for NTN parts kits. This unit is based on an advanced 3D vision system capable of locating unit parts boxes (UBs) in space. The main challenge lies in the variability of BU size and shape, with around 30 different types and the potential addition of new models in the future. The depalletizing system performs several key functions:

- Fully depalletize each BU handling unit, layer by layer, without picking boxes from the bottom layer before completing the current layer.

- Accurately locate BUs by transmitting their position (X-Y-Z), orientation and dimensions to the robot, while avoiding collisions.

Thanks to a patented algorithm combining AI and 3D vision developed by INNODURA, INNOROBOT is able to manage this diversity by precisely segmenting images and generating a point cloud for each BU.
The INNOROBOT system is capable of locating the boxes, calculating their grip vector and checking for collisions with the pallet. The system is fully integrated into the production line via a PLC. INNODURA also provides on-site installation and validation. Our solution guarantees accurate detection, even with screen prints or sizes not known in advance. Should exceptions occur in the future, we can re-train the model with the new data to maintain a high level of performance.