TrafficSignRecognitionBasedonConvolutionalNeuralNetwork

doi:10.3969/j.issn.1006-2475.2018.07.021

Abstract

Abstract: Withthedevelopmentofscienceandtechnology,therecognitionoftrafficsignsforunpilotedvehicleisbecomingmoreandmorepractical.Andthemostpopularmethodtorecognizethetrafficsignistousetheconvolutionalneuralnetwork.Theexperimentismadebasedonthismethod.Afterthecollectionofactualtrafficsignsimagedata,weusethehistogram-equalizationandothermethodstoachieveimagepreprocessingandconstitutethetrainingsetandtestset.ThenwebuildtheconvolutionneuralnetworkofLeNet5bythetrainingset,andusethetestsettotesttheaccuracyofthemodel.Inthefirstexperiment,therecognitionaccuracyofthemodelishigherthanothertraditionalidentificationmethods,butonly87.5%.Aftertheadjustmentoftheneuralnetwork’sbatchsize,theconvolutionkernel,thenumberoftrainingandotheraspects,therecognitionaccuracyisincreasedto93.2%.

Key words: convolutionalneuralnetwork, trafficsignrecognition, Python

CLC Number:

TP301

CHENBai-li，LINNan. TrafficSignRecognitionBasedonConvolutionalNeuralNetwork[J]. Computer and Modernization, 2018, 0(07): 103-.

References

［1］吴锦晶,张仁杰,唐春晖.交通路标图像的分割与识别方法研究［J］.光学仪器,2011,33(5):34-41.
［2］AkatsukaH,ImaiS.Roadsignpostrecognitionsystem［C］//Proc.ofSAEVehicleHighwayInfrastructure:SafetyCompatibility.1987:189-196.
［3］LiuHX,RanBin.Vision-basedstopsigndetectionandrecognitionsystemforintelligentvehicle［J］.TransportationResearchRecordJournaloftheTransportationResearchBoard,2001,1748(1):161-166.
［4］LoyG,BarnesN.Fastshapebasedroadsigndetectionforadriverassistancesystem［C］//Proceedingsof2004IEEE/RSJInternationalConferenceonIntelligentRobotsandSystems.2004:70-75.
［5］FangChiung-Yao,ChenSei-Wang,FuhChiou-Shann.Road-signdetectonandtracking［J］.IEEETransactionsonVehicularTechnology,2003,52(5):1329-1341.
［6］徐梅.基于内容的图像数据库检索技术分析［J］.通讯世界,2016(6):277-278.
［7］董浩,程鹏,李玲玲.深度学习算法在电力系统短期负荷预测中的应用［J］.电气时代,2017(2):82-84.
［8］丁婷,孙玉环.基于BP神经网络的交通标志识别［J］.物联网技术,2013(5):18-20.
［9］丰晓霞.基于深度学习的图像识别算法研究［D］.太原:太原理工大学,2015.
［10］庞振.基于径向基函数神经网络的图像识别研究［D］.长沙:长沙理工大学,2011.
［11］张安胜.深度学习在蛋白质二级结构预测中的应用研究［D］.合肥:安徽大学,2015.
［12］冯霞,石超,郝振平.基于差分图像的灰度信息匹配算法在轮胎检测中的应用［J］.电脑知识与技术,2015,11(3):183-184.
［13］张艳丽.基于多信息融合的门禁系统的设计［D］.天津:天津工业大学,2006.
［14］赵志宏,杨绍普,马增强.基于卷积神经网络LeNet-5的车牌字符识别研究［J］.系统仿真学报,2010,22(3):638-641.
［15］王珂娜,邹北骥,黄文梅.一种基于神经网络的畸变图像校正方法［J］.中国图象图形学报,2005,10(5):603-607.
［16］孙劲光,蒋金叶,孟祥福,等.一种数值属性的深度置信网络分类方法［J］.计算机工程与应用,2014,50(2):112-115.
［17］吴杰.基于深度学习的手势识别研究［D］.成都:电子科技大学,2015.
［18］陈宇.基于深度置信网络的中文信息抽取方法［D］.哈尔滨:哈尔滨工业大学,2014.
［19］胡锦城.基于稳定特征的交通标志检测与识别技术研究［D］.长沙:湖南大学,2012.
［20］邵桂珠.道路交通标志识别算法的研究［D］.长春:吉林大学,2008.
［21］黎华东.智能交通中的违章识别算法研究［D］.哈尔滨:哈尔滨工业大学,2014.

[1]	ZONG De-cai , WANG Kang-kang. A General Fast Design Method for Floating Point Arithmetic Instruction Based on TEC-XP16 Teaching Machine [J]. Computer and Modernization, 2021, 0(08): 77-84.
[2]	HEJun-jie,XIAOKe,LIUChang,CHENSong-yan. TFT-LCDCircuitDefectsDetectionBasedonFasterR-CNN [J]. Computer and Modernization, 2018, 0(07): 33-.
[3]	HE Bim, WANG Min-shuai. Application of Pywinauto in Software Performance Test [J]. Computer and Modernization, 2014, 0(8): 135-138.
[4]	LUO Pan1,2, XU Jian-nan2, LI Zhong2. Distributed Full-text Retrieval Method Based on Coreseek and Python [J]. Computer and Modernization, 2014, 0(6): 7-11.