Classification of brake pads in four different categories, based on their surface structure. Traditional camera vision systems are not applicable, because copper particles are distributed irregularly within the pads.

A camera connected to the S7-1500 TM NPU takes a picture of each brake pad. A Neural Network analyses the pictures and generates a quality indicator for the classification.

E.P.F. was requested to build a vision-based classification system for brake pads. These must be classified to 4 different categories. The classification happens accordingly to the irregular surface structure of pad, which cannot be . handled by a traditional camera vision system. Moreover, this system is intended to monitor and improve the prescending working steps.