针对车型特征,提出了一种基于BP神经网络的识别方法.从图像中提取车型特征向量,用BP神经网络设计分类器,并进行有效的训练与测试.应用改进的BP算法(尺度化共轭梯度法)对网络进行训练,使网络很快得到收敛,解决了一般算法收敛慢的缺陷. Aiming at vehicle model characteristics, a recognition method based on BP neural network is proposed, vehicle feature vectors are extracted from the image, the classifier is designed by BP neural network and effective training and testing are carried out.Based on the improved BP algorithm Gradient method) to train the network so that the network converges very quickly and solves the problem that the general algorithm converges slowly.