A Selection Algorithm of Training Set Based on Similar Classification

摘要

License Plate Recognition (LPR) combines computer vision technology and pattern recognition technology and plays an important role in Freeway Toll System, Urban Road Monitoring System and the Intelligent Parking Lot Management System. Therefore, it has attracted an ever increasing number of scholars from home and abroad. Despite many years of unremitting effort which has resulted in breakthrough achievements, it remains unsatisfactory in meeting real world application requirements. LPR primarily employs pattern recognition and digital image process technology. This paper is focused on the study of pattern recognition. The segmented characters are trained utilizing the BP neural network. Selecting the ideal training set from the usually large sample set we have is the first step to train a good network which has a high recognition rate. At present, training sets are randomly selected, which affects the accuracy of recognition as well as its speed. Thus, selecting the best training sets is of uttermost importance. In this paper, Similarity Comparison Sampling method is proposed to improve the training results.