Case Sharing: Deployment of Warehouse Monitoring System in A Certain South China Automated Warehouse

Project background introduction

Selection of storage position status recognition sensor

• Single-point LiDAR: Single-point LiDAR can only emit one beam of laser to form a point on the surface of the object at a time. This ranging method makes it easy to ignore the gap between the carton and the carton or the tray and the tray, and then identify the current storage location as missing goods or empty goods, causing stacking accidents.

• RGB camera: When detecting the target area through Deep Learning, once an object outside the training dataset enters the storage location, it may cause false detection and give incorrect storage location judgment information; in addition, for the storage location judgment with cargo stacking requirements, the information given by the RGB camera lacks cargo height information, which makes it difficult to arrange stacking tasks.

• MRDVS RGB-D camera: can provide three-dimensional data and color information of the library, built-in computing power, and no external industrial computer.

On-site installation photos

Detailed introduction of MRDVS plan

• Comprehensive monitoring: Through the collection of three-dimensional data and color information, combined with AI technology, accurately identify the status of the warehouse.

• Easy deployment: The database status recognition algorithm is placed on the camera side to reduce deployment and maintenance costs.

• Flexible communication: Supports multiple communication methods such as TCP/IP, UDP, HTTP, etc., and reports data to the control system in real-time in JSON format.

• Improve efficiency: Real-time transmission of location information to assist the scheduling system in quickly and accurately assigning tasks.