The recognition of aluminum alloy wheel hub is an important part of the intelligent production management platform. Based on the powerful machine vision open source library of the OpenCV,the Microsoft Visual Studio platform and MFC was used to build the system interface which provided the real-time photographs of wheel hubs. The shapes of the rough castings with a larger burrs were polygonally simplified and their features were extracted. Then the order number of the images was generated to match and classify the wheel hub images to finally conduct the real-time classification. The system can improve effectively the production efficiency,and realize the intelligent hub classification and data transmission.