Camera-based identification systems for empties

Reusable deposit systems, in which the bottles are cleaned after use and reused directly, have a lower requirement for raw materials and energy than disposable systems. A disadvantage of this system is the high logistical effort - the empties have to be delivered from the end customer back to the beverage manufacturer. The processes required for this are summarized under the term empties management and subdivided into external and internal empties management. External empties management refers to the direct acceptance of empties from the end customer in the beverage trade through to presorting and handing over the empties to the beverage manufacturer. Internal empties management, on the other hand, includes the delivery of empties to the beverage manufacturer itself, the recording and storage of the goods, cleaning and inspection, as well as providing empties for production. Since efficient production is directly dependent on recording the correct stocks, this process must meet particularly high quality criteria. The implementation poses great challenges for manufacturers.

The pre-sorting of the external empties management system achieves a relatively high grade purity within the beverage crates. For transport to the beverage manufacturers and storage the crates are also placed on pallets. When empties are received at the factory, the pallets are unloaded from the truck by forklift trucks, which can pick up several plates per unloading operation.
In order to transmit information (amount, type) about the delivered crates to the manufacturer's inventory management system, the forklift driver must manually inspect the loaded pallets. As the different types of crates are very similar in design or colour, there is often confusion and consequently incorrect postings in the inventory management system. This incorrect information in the inventory quantity can lead to delays or failures in the subsequent goods production. Therefore, regular manual visual inventories are carried out to check the stock information.

To make the identification of the empies at goods receipt more reliable and economical, a fully automatic camera-based system is to be developed which relieves the forklift driver during the inspection process. Important requirements are mobility, real-time quantity and classification of the empties crates on the pallets during the unloading process. For this purpose, a camera system captures a complete all-round view of the load and evaluates it using image processing technologies such as neural networks.

Four aspects are considered for the implementation:

  • Frontal (camera) sensor technology on the forklift
  • Lateral (camera) sensor technology k
  • Communication unit for synchronization between the camera modules and data transmission to computer unit
  • Computer unit for image data evaluation

Due to the close cooperation with the industry, the project is carried out and evaluated on a real application case. This results in a direct benefit of the research project for the project partners from industry.