Quality Control for Bread Manufacturing
THE CHALLENGE
- A large bakery was experiencing massive bread stock loss variance which they attributed to theft, miscounting and difficulty in implementing effective control processes.
- Despite deploying security guards and a surveillance system, the bakery continued to see a significant reduction in the product reaching its end destination.
TECHNOLOGY USED
- Computer Vision
- Machine Learning
- Azure
THE SOLUTION
The human counting process is inadequate to stop or reduce the theft/loss. Either through collusion, syndicates and intimidation, error or other factors, the counts performed by human beings at each stage of the process did not match and resulted in direct losses.
Using computer vision, visionAI transformed the bakery’s existing camera system into an automated security system. This system does not tire, cannot be distracted and is not open to error or influence.
BUSINESS OBJECTIVE
The bakery’s challenge was evident in the difference between the amount of bread counted in dispatch and the amount that should be delivered.
Fundamental to the problem of understanding the flow of bread within the factory is to understand the movement of palettes from key central areas – accurate counts at each stage of production and delivery became critical to control specific areas.
THE RESULT
The bakery expects a 50% decline in missing loaves from this system that they installed, manage, and optimize themselves.
The visionAI system requires significantly less staff, produces more accurate results, and forms the backbone of an integrated system that affects more than just the bread stock losses.