Research On Traffic Sign Recognition Method Based On CNN

Due to the uncontrollable conditions and poses of traffic signs,traffic signs detection and recognition have been challenging the robustness and efficiency of the existing algorithm.The former has a lot of false detected windows,and the latter is influenced by environmental factors seriously.In this work,this paper presents a method of traffic signs detection based on HOG and Boolean Convolutional Neural Networks(HOG-BCNN).A cascade classifier is trained based on HOG to detect the candidate regions of traffic signs.These regions as proposal windows input a special CNN,which is like a Boolean logic,which can eliminate the error detection so that improve the detection rate in traffic signs detection.A method of fast traffic sign recognition based on convolutional neural networks(FTSR-CNN)is proposed to recognize the signs in this paper.A new architecture of convolution neural network is designed,which extracts feature using convolution kernels by sliding filter,and reduces dimension by pooling.In forward propagation,the loss of the whole network is computed.Then stochastic gradient descent method is exploited to minimize loss in step of back propagation.The parameters and activation function of the network are adjusted to optimize the performance.Experiments demonstrate that the proposed methods are sufficient to attain high detection rate and classification accuracy when ensuring the efficiency.In the end,this paper builds the traffic sign recognition system by combining the methods of traffic signs detection and recognition.The method has been evaluated on multiple video,captured by vehicle traveling data recorder,and obtain great effect.