Using machine vision to speed up quality assurance for tempered glass

glaston hero

Glaston is a leading manufacturer of tempered glass. The quality of glass is controlled by a glass fragmentation test which is carried out manually by humans several times a day, taking up to an hour of their workday.

In collaboration with the design agency, Nordkapp, we were challenged with finding an innovative way to use technology to automate this tedious task to reduce labor time. Our deep knowledge and understanding of machine vision technologies meant we were a natural partner for Nordkapp to build a solution that would not only have the needed functionality but live up to the desired brand experience.

Video courtesy of Nordkapp

Project Highlights

  • Utilizing machine vision technology for glass fragment detection.
  • Using a neural network to calculate the number of fragments within a measured area.
  • Additional key technologies used included TensorFlow, AR Core, and AR Kit.



Improved Accuracy

In order to determine the cullet-count of tempered glass, it must be shattered and an average value determined for the number of individual facets of glass created in the process. What was a manual process, could now be taken on utilizing machine vision technology to analyze images of the shattered glass and determine the count instantaneously and automatically.

Story 1


Precision in Your Pocket

The workforce within Glaston is constantly on the move from one area to another, as well as out in the field. By packaging up the technology solution into an app that can run on anyone's smartphone, the benefit of the Siru technology goes everywhere the user does.

Time Saving Automation

Because the machine vision solution can calculate the cullet count in seconds, the time savings per employee for Glaston was tremendous, reducing a process that took up to an hour a day to just a few minutes. Historical records as well as automatic digital reports also helped further improve efficiency.


Technologies used


The Results

Although Glaston Siru is primarily an industry tool, the app is publicly available for iOS and Android. Anyone can download the app and test out the machine vision solution on broken glass anywhere. Just don’t go breaking too much.



circlesml-Jani Jani Seppälä
Head of Computer Vision and AI at Finlabs

This project was a great example of the power of Machine Vision. What used to be a time-consuming manual process is now a more reliable machine vision solution that works in a matter of seconds and is more precise than a human could ever be.

Group 209

Similar cases