摘要
针对现有车型识别算法的耗时长、特征提取复杂、识别率低等问题,引入了基于深度学习的卷积神经网络(convolutional neural network,CNN)方法。此方法具有鲁棒性好、泛化能力强、识别度高等优点,因而被广泛使用于图像识别领域。在对公路中的四种主要车型(大巴车、面包车、轿车、卡车)的分类实验中,改进后的卷积神经网络LeNet-5使车型训练、测试结果均达到了98%以上,优于传统的SIFT+SVM算法;另外,还研究了改进网络中的Dropout层对车型识别效果的影响。与传统算法相比,经过改进后的卷积神经网络LeNet-5,在减少检测时间和提高识别率等方面都有了显著提高,在车型识别上具有明显的优势。
Aiming at the problem of time-consuming,complex feature extraction and low recognition rate,this paper introduced the convolution neural network method based on deep learning. This method was widely used in the field of image recognition because of its good robustness,strong generalization ability and high recognition rate. In classification experiments of the 4 main vehicle types( bus,microbus,car,truck) on road,the improved convolution neural network LeNet-5 achieved vehicle type training and test accuracy both over 98%,which outperformed traditional SIFT + SVM algorithm. In addition,this paper also studied the effect of the Dropout layer in improved network on the vehicle type recognition. Compared with the traditional algorithm,the improved convolution neural network LeNet-5 has a significant improvement in reducing the detection time and improving the recognition rate,and has obvious advantages in vehicle type recognition.
引文
[1]蔡伯根,赵济民,王剑,等.三轴AMR车辆检测器的车型分类设计实现[J].交通运输系统工程与信息,2014,14(4):46-52.
[2]崔莹莹.智能交通中的车型识别研究[D].成都:电子科技大学,2013.
[3]华莉琴,许维,王拓,等.采用改进的尺度不变特征转换及多视角模型对车型识别[J].西安交通大学学报,2013,47(4):92-99.
[4]Kazemi F M,Samadi S,Poorreza H R,et al.Vehicle recognition using curvelet transform and SVM[C]//Proc of the 4th International Conference on Information Technology.Piscataway,NJ:IEEE Press,2007:516-521.
[5]Rahati S,Moravejian R,Mohamad E,et al.Vehicle recognition using contourlet transform and SVM[C]//Proc of the 5th International Conference on Information Technology:New Generations.Piscataway,NJ:IEEE Press,2008:894-898.
[6]Dalal N,Triggs B.Histograms of oriented gradients for human detection[C]//Proc of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Piscataway,NJ:IEEE Press,2005:886-893.
[7]景阳,刘琳,黄鸿志,等.基于头肩HOG特征的快速行人检测[J].电视技术,2014,38(15):227-230.
[8]吴月凤,何小海,张峰,等.SURF算法和RANSAC算法相结合的台标检测与识别[J].电视技术,2014,38(13):191-195.
[9]Dollar P,Tu Zhuowen,Perona P,et al.Integral channel features[C]//Proc of British Machine Vision Conference.London:BMVC Press,2009:247-252.
[10]Hinton G E,Osindero S,Teh Y W.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554.
[11]Le Cun Y,Bottou L,Bengio Y,et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE,1998,86(11):2278-2324.
[12]Lawrence S,Giles C L,Tsoi A C,et al.Face recognition:a convolutional neural network approach[J].IEEE Trans on Neural Networks,1997,8(1):98-113.
[13]Lauer F,Suen C Y,Bloch G.A trainable feature extractor for handwritten digit recognition[J].Pattern Recognition,2007,40(6):1816-1824.
[14]Tivive F H C,Bouzerdoum A.An eye feature detector based on convolutional neural network[C]//Proc of the 8th International Symposium on Signal Processing and Its Applications.Piscataway,NJ:IEEE Press,2005:90-93.
[15]Mate S,Akira Y,Munetaka Y,et al.Pedestrian detection with convolutional neural networks[C]//Proc of IEEE Intelligent Vehicles Symposium.Piscataway,NJ:IEEE Press,2005:224-229.
[16]张松松,詹智财.基于SIFT特征与支持向量机的车辆分类系统研究[J].电脑知识与技术,2012,8(17):4277-4280.