Abstract:

The most common automobile meter is the pointer meter. The pointer detection of the automobile meter in industry is mainly concentrated on the detection of the pointer angle and the measurement of the light guide property of pointer. Accurate detection is the key of automobile quality. Currently, manual meter detection is widely used in Chinese automobile industry, which is tedious and with a high variability of quality control. According to the image characteristics of automobile pointer meter, a minimum gradient comparing subtraction method is used to extract the meter pointer from two images with the same background and pointers in different positions. Then, the pointer angle can be recognized by the binary method, the thinning algorithm and the Hough transform. Meanwhile, the extracted pointer is utilized to realize the detection of pointer light guide property. Experimental results show this proposed method is robust, accurate and can be performed in realtime. It has been successfully applied to an automatic detection system of automobile meter based on computer vision.   

Key words: automobile meter;pointer detection;light guide;minimum gradient comparing subtraction;Hough transform