To implement an automatic system for quality control on meat to be packed.
To evaluate the quality of meat by highlighting defects (such as excessive fat or unusual color), evaluating the correct positioning and appropriate sealing of the pack.
Currently, all controls concerning the quality of the meat and the correct positioning are carried out by specialized operators, taking a lot of time and with a high risk of error due to the subjectiveness of the judgment.
The vision system already installed on the production line was re-used to acquire images. These images are loaded directly into AI-go Studio and used for the model specialisation to identify the different types of meat and evaluate their quality on the basis of customer-defined classes. In addition, the operator is completely autonomous in the creation of new models, related to new meat varieties or defect classes.
A second model evaluates the correct positioning of the meat in the pack, avoiding the possibility that meat placed too close to the edge could compromise vacuuming.
The trained and tested models were then deployed in production through AI-go Runtime, ensuring full compliance with the cycle times of the packaging line.
Quality evaluation objectification. The quality evaluation is more reliable over time.
Better anomaly detectability – even in sub-optimal visual conditions.
Data-driven approach. For the client, the system is an important building block towards high quality standards across the production line.