Wheel surface defect detection method using laser sensor and machine vision

The safety of the railways is maintained by regular maintenance of tracks, signals, and rolling stocks to keep them in good conditions. Notably, such maintenance is ensured using a manual inspection method that involves skilled maintainers, however automated inspection systems are only partially used. To maintain a good condition of railways, the development of the automated inspection systems is crucial. Herein, we focused on automatic surface defect detection of wheels because wheels are especially important components of the rolling stocks. We have already previously evaluated wheel tread profiles using laser instruments. However, in this research, to confirm the accuracy of the measurement equipment, we prepared special test pieces that include intentionally processed surface defects. Experiments were conducted using different materials, different passing speeds by employing laser sensors, and a machine vision technique to confirm the effectiveness of the proposed inspection system.