How do you know when biscuits are baked to perfection? At Belgian biscuit company La Confiance, quality control is currently subject to the trained eye of individual employees, who evaluate samples at regular intervals and adjust the baking process accordingly.
With the BBISQUIT project, La Confiance wants to switch to a more objective and continuous quality control system. Cameras and intelligent software will evaluate the colour of the baked biscuits and make the necessary adjustments to oven parameters. The benefit is optimised production and increased food safety – and always uniform quality.
In addition to La Confiance, the international consortium behind this project consists of AI experts from Yazzoom in Belgium and oven automation specialists from Goldstream in Holland.
Final report summary
The goal of this project was to introduce vision and AI techniques to optimise typical food processes like baking. The objective was to reduce variability in quality and waste by automatic control, replacing the manual and visual inspection in use at the start of the project. The main challenges the team found were dealing with the non-uniformity of the products and with the natural variations in the ingredients.
The outcome of the project is an automatic system for optimising the baking process. A supervisory controller, using smart vision/AI techniques, identifies biscuit characteristics representative for the baking process. The system allows for more throughput with a more constant quality and less waste, despite ingredient variations.
For more information, see the S3FOOD article here.