Computer Vision : Case Studies

Conveyor Line

⚡ Problem:

Used a labor intensive approach to count each loaf on each tray and to ensure that the correct quantity of loaves was packaged. This method was costly, and left the process open to errors.


For our Client, a bread manufacturing company, the automation of counting and rejecting bakery products on the conveyor line became imperative.

On the conveyor line loaves are packed into trays, with each tray containing 8 loaves for further transportation along the conveyor line. These loaves come in various types distinguished by their packaging.

The computer vision system we developed was able to meticulously track the product count for each SKU and swiftly identify and reject trays with irregular loaf quantities or with quality defects.

Conveyor Line
⚙️ Results and Solutions
A connection to the installed cameras via RTSP was established and the video stream 'scanning' by the neural network was configured to track the product count and to Identify and reject trays with irregular loaf quantities and/or with defects.
Through the utilization of machine learning techniques, dataset labeling was completed for detecting trays, counting the number of loaves, and classifying loaves by their SKU.
The project implementation surpassed the customer's target metrics, achieving an accuracy rate exceeding 95%.
95%

Accuracy rate

Target reached. The model’s precision outperformed human accuracy.
80%

Reduction

Production inconsistencies
50%

Reduction

In customer’s claims
Contact us today to schedule a consultation and embark on your journey
to data-driven success.
TRY TRY Contact Contact