As identification of machine components on-line becomes more and more in important in automatic production. Accomplishing the process automatically is practically significant. In this paper, we present an automatic and effective framework for machine components identification based on the figure features of digital image. Using industrial camera lens instead of man eyes as sensor, then a series of operation, such as photoelectric transform, image sampling, image process, image analyzing and identifying, machine components identification can be realized. For the first time, the feature of shape edge variety is presented in this paper, and the shape description is moving insensitive, rotating insensitive and zooming insensitive. It is validated by experiments. By the experiments, it shows that the method is effective and efficiency.
|