Determination of manufacturer, size and tread depth on tires

Many products are only encountered in classic machine vision as new products. When it comes to tires, however, the requirements are often more diverse. During tire production, tire fitting, logistics and used tire disposal, these products are encountered in different states and must be automatically detected and recorded. The latter, the disposal of used tires, is the subject of the machine vision task described here. In this area, the brand names, tire size and tread depth are of interest.

Laser triangulation sensors are used to record the aforementioned features. The tires are roughly mechanically centered and rotated in front of the sensor system via a roller drive. An algorithm from the field of artificial intelligence is used to classify the manufacturer and tire size. Due to the large number of tire manufacturers, it is essential to build up a database. The corresponding manufacturer lettering is trained. This continuously increases the system’s recognition performance over time (deep learning).

The tread depth measurement of the tire is determined at different points on the tread. The system was developed as a stand-alone testing system in cooperation with our mechanical engineering partner for used tire recycling plants.

The machine vision system already demonstrated its modularity and flexibility during the development phase. This allowed the development to take place in stages and the investment risk to be kept low.