Skip to Main Content
The BP networks have the ability of nonlinear mapping, so they are widely used in pattern recognition and classification. However, BP networks need to be trained again when the training set is changed. Meanwhile, the larger the network is, the slower convergence rate is, and the poorer result of classification and recognition is. So, two-hierarchy neural network classifier for recognition of traffic signs is presented: the first hierarchy classification which consists of a single BP network is used to coarsely classify indicative signs, warning signs and prohibitive signs; the second hierarchy classification including of three BP networks is designed to concretely identify each traffic signs. The simulation results show that the correctness of recognition and classification is up to 100% for testing set with white Gaussian noise. To reduce the scale of the first classification training set and improve the adaptability of training set, two incomplete training set are used: (a) taking part of samples as training set, and (b) obtaining smaller training set by artificial selection. The two-hierarchy neural network classifier and incomplete training set could improve the convergence rate and identification ability; meanwhile it is proved that the first hierarchy classification is robust to the training set.