Infrared passenger flow collection system based on RBF neural net

Abstract

As the current level of the people-counting, on the consideration of the cost of the acquisition equipment and the acquisition accuracy, the customer-counting system is constructed based on RBF neural net using technology of infrared photoelectric sensor. For better differential count, extant data segmentation and the method of feature extraction is improved based on the feature of passenger-counting data continuous space-time sequence. Compared with traditional customer-counting using sensor, the accuracy of real-time customer-counting is improved in this method and the situation of customers who entry at the same time can be identified with lower error rate. It is helpful for both principle research and actual application.