Study on Tire Sidewall Marking Recognition Based on Moments

Tire sidewall marking are the information for customer usage, safety, regulatory of national and manufacture internal trace ability. The consequence is high severity when marking are not correct. Hence, the marking verification is very important, because tire sidewall marking is arc-distribution, it is difficult to extract the feature vector. To solve this problem, a moment-based method is presented in this paper, which avoids stretching and correction during the recognition. The extracted vector is scale and rotation invariance. Experimental shows that the recognition rate of presented method is above 94.7%, which indicates that the method can be put into partial application to substitute for manual verification.