Automobile Speedometer Reading Recognition for Automated Testing

Automated testing of manufactured products reduces the lead time to considerable extent in the process of production to delivery. Products like automobiles demand automated testing, for which robots and vision systems are widely employed. The basic functionality of a vision system in automation is to detect an object and then recognize it. In current automotive industry such systems are being used for robotic guidance, component tracking, dimensional gauging etc. There is a need to test the proper functionality of a speedometer fitted on a motorbike in the production line itself.

Focused work on detection and recognition of Analog type and Digital type speedometer console reading of a motorbike is described in this paper. A vision based system is proposed which recognizes the speedometer reading instantaneously at the desired time. Image binarization, connected component analysis combined with character recognition algorithms are used to achieve the desired recognition, which resulted in reduced lead time hence contributing to lean manufacturing