Automatic Identification Method Based On Mechanical Image

Mechanical image is an important representation of fault diagnosis information. Due to its high dimension of mechanical images, it is difficult for the computer to identify directly. At present, the identification of mechanical image is mostly dependent on the human eyes. In order to recognize the mechanical images automatically, data reduction technology for the mechanical images were studied, and the support vector machine is used to recognize the mechanical images automatically in this paper.The advantages of non-negative matrix factorization were analyzed with respect to principal component analysis, linear discriminant analysis and kernel principal component analysis. And the four methods are applied to the face images in the face database. The results show that the non-negative matrix factorization is superior to other three methods. Then the non-negative matrix factorization and sparse non-negative matrix factorization are mainly discussed for the data reduction of ORL face database. The experimental results of image reconstruction show that the sparse non-negative matrix factorization has better characteristics.Finally, the sparse non-negative matrix factorization is applied to the mechanical images to reduce the data dimension and the support vector machine is used to identify these images. The results show that a higher recognition rate can be gotten by this method.