A novel method to evaluate the gear quality based on machine vision technology is introduced.The method included: fast acquisition of the contact area of the gear

calculation of the deviation from the contact area to the centroid of the gear surface and the fractal dimensions of the contact area

a multiple linear regression model is established in which the gear quality is dependent variable

and the deviation and the fractal dimensions are independent variable.To validate this method

a field experiment is carried out in the workshops of First Automobile Works.The results show that the identification efficiency of the proposed method has a good consistency with the identification by an experienced worker（the consistency rate reached up to 92.26%）

which means this machine vision method can be used to assess the quality of a gear effectively.