AI-enabled vision to identify auto parts in an extreme temperature environment for a global top-20 automotive parts manufacturer

Our client needed to identify heat exchanger parts as they travel along the conveyor from multiple production stations through a brazing furnace with a temperature of over 600° Celcius. Once a part passes through the furnace, any form of tradition ID such as RFID (radio frequency ID) or barcode burns, making it impossible to identify the part and direct it to the appropriate station for further processing.

In addition, special frames are used to protect heat exchangers from damage as they move along the conveyor. It is important to capture the number of times each frame travels through the furnace, to enable regular cleaning and maintenance.

Our task was to design a form of identification that withstands not only extreme heat, but also multiple cycles. To achieve this, we created AI-enabled vision to read metal punch cards (our product Heat ID). Our solution identifies the cards despite the inevitable damage from the heat and the fact that the card (attached to the product) is in continuous motion. In addition, our system tracks each frame and controls its direction of travel.

We also incorporated predictive maintenance, which estimates various ‘dangers’ (such as the cameras’ exposure to extreme conditions, changing light, failed communication with the conveyor) and generates alerts for workers to take action.


  • Identification accuracy of above 99%, despite extreme processing conditions and continuous motion.
  • Accurate real-time location data for each of the hundreds of frames on the conveyor.
  • Significant reduction in production downtime through predictive maintenance.
Want to know more about this project?
Contact urobots
Fill out this field
Please enter a valid email address.
Fill out this field