欢迎来到天天文库
浏览记录
ID:51417307
大小:5.95 MB
页数:7页
时间:2020-03-23
《基于红外标记视觉的安全带佩戴规范性检测.pdf》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、汽车工程2017年(第39卷)第7期AutomotiveEngineering2017(Vol.39)No.7doi:10.19562/j.chinasae.qcgc.2017.07.006*基于红外标记视觉的安全带佩戴规范性检测1,21311钟铭恩,黄伟,温程璐,黄波,黄杰鸿(1.厦门理工学院机械与汽车工程学院,厦门361024;2.福建省客车先进设计与制造重点实验室,厦门361024;3.厦门大学信息科学与技术学院,厦门361005)[摘要]针对安全带佩戴规范性仍未引起充分重视的现实问题,借助具有红外敏感表面涂层标记的三点
2、式安全带,基于标记视觉和图像处理技术提出一种安全带佩戴规范性的集成检测方法:利用增强现实技术快速识别定位安全带标记,在此基础上分别设计数量阈值算法、模糊聚类阈值算法、曲线拟合松弛度算法和肩部外轮廓交点定位算法,用以进行安全带是否佩戴和佩戴形式、松紧程度、高低位置是否规范等4种判断,从而实现安全带是否规范佩戴的检测目标。实验结果表明,在各种成像条件和图像背景干扰下,对应上述4种判断的识别准确率分别为95.4%,93.1%,79.5%和85.3%,而安全带佩戴规范性的综合识别准确率达89.5%,验证了所提出算法的准确性、有效性和适
3、应能力。关键词:安全带检测;佩带规范性;标记视觉DetectiononProperWearingofSeatBeltBasedonInfraredMarkVision1,21311ZhongMingen,HuangWei,WenChenglu,HuangBo&HuangJiehong1.SchoolofMechanicalandAutomotiveEngineering,XiamenUniversityofTechnology,Xiamen361024;2.FujianProvincialKeyLaboratoryofBusA
4、dvancedDesignandManufacture,Xiamen361024;3.SchoolofInformationScienceandEngineering,XiamenUniversity,Xiamen361005[Abstract]Inviewoftheproblemthatwhetherseatbeltisproperlyusedornothaven’tyetattractedsuffi-cientattention,anintegrateddetectionmethodoftheproperuseofseat
5、beltisproposedbymeansofthethree-pointsafetybeltwithinfrared-ray-sensitivecoatmarksandbasedonmarkvisionandimageprocessingtechnologies:Aug-mentedrealitytechnologyisusedtorapidlyidentifyandlocatethemarksonseatbelt,basedonwhichthealgorithmsofnumberthreshold,fuzzycluster
6、ingthreshold,curvefittingrelaxationandshoulderprofileintersectionlocatingaredevelopedrespectivelytoconductfourjudgmentsofweatherseatbeltisusedornot,thewearingformsofseatbelt,belttightnessandbeltheight,andhenceachievethedetectiongoaloftheproperuseofseatbelt.Experimen
7、tresultsshowthatundervariousimageformingconditionsandimagebackgroundinterferences,theidentificationaccuraciesofthealgorithmsusedcorrespondingtoabovementionedfourjudgmentsare95.4%,93.1%,79.5%and85.3%re-spectivelywithanoverallidentificationaccuracyof89.5%,verifyingthe
8、correctness,effectivenessandadaptabilityofthealgorithmsproposed.Keywords:seatbeltdetection;properuseofseatbelt;markvision辆发生交通事故时可大幅降低车载乘员
此文档下载收益归作者所有