Brake Machine Target Recognition Method Based On Svm Research

With the development of urban railway traffic, Automatic Fare Collection system is becoming more and more important. Gate is the core part of AFC system, which can recognize and judge the behaviors of the passengers and use to automatically examine tickets by means of its intelligent recognition system. Both the relation to the passengers and the function of ticket’s examination of gate AFC system, require it must have very high security, reliability, stability and accuracy and so on. Therefore, the intelligent recognition system has become one of the hot spots of urban railway traffic. This paper aims at the characteristics of the targets that go through the gate, thus the targets can be divided into two types which are pedestrians and objects. Then we put forward a SM recognition method based on Support Vector Machine, and research the pedestrian pattern recognition and object pattern recognition. Meanwhile, this paper makes a pretreatment to the data collected by the sensor based on the concept of the standard recognition elements, studies the digital characteristics and distribution rules of pedestrian and object included in SVM training set, and establishes a SMDF recognition predictive model based on LibSVM. According to the testing set, we research the model’s classification results of the two kinds of targets in the condition of different kernel functions, different parameters and different punishment parameters and research the correlation between the classification effect of the model and the training set size. The experiment of gate target recognition shows that both the SM recognition method and the SMDF recognition model are feasible, and they meet the automatic ticket system’s requirements of real-time and accuracy.