Application of support vector machines method on the classification recognition of wear mark

This paper applies the wavelet packet method to extract the characteristics of striation wear mark made by using two different types of guns to shoot the same types of bullet. By introducing the fractal dimension parameter of the profile curve, i.e. characteristic wavelet packet dimension as factor component, this paper first categorized the sample data after normalization into training sample collection, experiment sample collection and generalization sample collection. According to the basic principle of the support vector machines, SVM classification model is established by the training sample and tested by the experiment sample collection. Adjusting the value of parameter U and C frequently, this study found out an ideal parameter to establish this ideal SVM classification model. After that, the generalization sample is used to test the generalization ability of the model. The study indicates that when the characteristic wavelet packet dimension is used as sample collection, there can be a better classification of the same model of bullet shot by different types of gun. The modern scientific thought such as support the vector machine is introduced into mark examination, which will bring about new breakthrough for mark examination theories.