MAGIC EYE

This is a pilot device for monitoring and predicting undesirable conditions of mechanical parts or of the presence of foreign objects, in our case it relates to an overhead hanger conveyor for bodywork, which is the logistic backbone between the individual production sections. This conveyor must handle almost 1,250 cars per day during a three-shift operation without a back-up solution. Its parts are exposed to relatively high forces in use, causing them to wear and deteriorate in a relatively short time. Given that the entire conveyor is at a height of several metres, it is difficult to access. To eliminate all causes of faults that would mean huge financial losses for production, we have introduced the MAGIC EYE into the fray.

Its principle of operation is based on a comprehensive system consisting of several powerful cameras with an analytical computer so that quality image recognition is achieved for the images that have already been taken. By using AI (deep learning / neural networks) we can train the model for precise classification of defects and their extent and location in a short time. Simply put, we will be able to substitute the “eyes” of a maintenance worker for whom it would be very difficult to access these locations.                                                  

The MAGIC EYE has been designed for direct installation onto the travelling trolley of the overhead conveyor. For this reason, the diagnostics and rapid assessment of the condition can be carried out immediately during the actual passage of the entire conveyor. The results of the image comparison are transmitted via WiFi and displayed on the visualization in the central control room where a timely maintenance alert is sent out. This ensures a rapid reaction to (elimination of) potential downtime. The system allows us to perform diagnostics in hard-to-reach places in real time, but also, thanks to powerful computing and machine learning, we can predict individual faults within weeks.