Suzuken Co., Ltd. and NEC Corporation (NEC) are aiming for automation and labor saving at Suzuken's wholesale distribution base "Metropolitan area distribution center" scheduled to start operation in March 2024, and NEC's AI and Announced that it conducted demonstration experiments of "automation of arrival slip input" and "automation of packing goods inventory" utilizing advanced technologies such as object recognition from July to November 2021.

In addition to building a distribution system that complies with the GDP guidelines, the center aims to improve both distribution quality and operational efficiency through automation and labor saving. As part of this, Suzuken and NEC will evaluate and verify the balance between accuracy improvement and efficiency in the input of incoming slips and inventory work in the distribution center in this demonstration experiment, and the effectiveness will be confirmed and further operational efficiency will be improved. It is said that it was possible to obtain the necessary knowledge.

● Outline and results of the demonstration experiment (1) Automation of arrival slip input [Experiment outline] Automation of arrival slip input by OCR (Optical Character Reader) utilizing NEC's AI

[Verification Results] In the same experiment, more than 90% of the delivery slips, such as 204 out of 439 formats, were automated. As a result, it was confirmed that the business time required for inputting the arrival slip can be reduced by 75%.

By having OCR using NEC's AI learn more than 6,000 images of delivery slips, it has become possible to read written characters that were difficult to read accurately with conventional OCR with a high recognition rate, and AI. It has become possible for one staff member to remotely input inbound slips that cannot be recognized even by OCR using the above, so that it is possible to handle the work of multiple distribution centers while staying in a remote location.

(2) Automation of inventory of packaged items [Experimental outline] Automation and labor saving of inventory work by individual identification of packaged items using NEC's simultaneous object recognition technology

[Verification results] In the same experiment, 100% automation of the manual inventory work of packed goods using paper forms, handy terminals, etc., using a barcode reader installed on the belt conveyor line. Achieved.

At the same time, by utilizing NEC's simultaneous object recognition technology, the image data of the packaged items such as corrugated cardboard captured by the barcode reader is collated with the master data registered in advance, and the packaged items are individually recognized to change the packaging. Confirmation is possible.

At the same time, it has become possible to read information such as the lot, expiration date, and expiration date written on the side of the package from the image data by OCR and automatically register the system.

The two companies will utilize the knowledge gained from this demonstration experiment to further improve work efficiency, and in the future, further automation and labor saving in the distribution center will be achieved by transport work using robots and autonomous / remote control of forklifts. I am aiming for it.

● System outline diagram of inbound slip input automation (top), system outline diagram of package inventory automation (bottom)