电子测量技术ELECTRONICMEASUREMENTTECHNOLOGY第46卷第1期2023年1月DOI:10.19651/j.cnki.emt.2210142基于图像匹配与WiFi信号匹配的室内PDR定位方法刘水清1,2(1.华中科技大学武汉430074;2.大众报业集团信息技术部济南250002)摘要:由于智能手机中存在丰富的传感器,可对周围环境进行感知并进而估计位置。本文提出一种融合室内定位方法,融合的信息包含图像信息、WiFi信号指纹信息以及惯性数据观测信息。通过实测实验结果表明,本文提出的方法相比于惯性定位、WiFi定位、图像匹配定位三种单一的定位方法,定位精度分别提高了48%、24%和4%。相比于经典WiFi和图像融合方法,平均定位时间下降了92%。相比于经典融合方法,定位的精度提高了0.7m,并且平均定位时间下降了87%。关键词:室内定位;融合定位;WiFi信号指纹;图像处理;惯性定位中图分类号:TP391文献标识码:A国家标准学科分类代码:510.51040PDRindoorpositioningmethodbasedonimagematchingandWiFifingerprintingLiuShuiqing1,2(1.HuazhongUniversityofScienceandTechnology,Wuhan430074,China;2.ShandongDazhongNewsGroup,Jinan250002,China)Abstract:SmartphonebasedindoorpositioninghasgreatpotentialusageinLbS(LocationbasedService)applications,duetotheabundantsensorsembeddedinsmartphones.Thispaperproposedahybridpositioningmethod,whichadoptedaparticlefiltertofusethreetypesofindependentpositioningresults,includingimagematchingpositioningresult,WiFifingerprintingresultsandPDRpositioningresults,whichachievecontinuousandaccuratepositioningresults.Accordingtorealscenarioexperiments,theproposedmethodhasanaccuracyimprovementofabout48%,24%and4%overpurePDR,WiFifingerprintingandimagematchingbasedmethods.ComparedwithclassicalWiFiandimagematchingfusionmethod,ourmethodcanprovideadecreaseof92%averagepositioningtime.Comparedwithanotherclassicalfusionmethod,theaccuracyhasincreasedby0.7mandthepositioningtimehasdecreasedby87%.Keywords:indoorpositioning;hybridpositioning;WiFifingerprinting;imageprocessing;inertialbasedpositioning收稿日期:2022-05-280引言室内定位服务随着基于位置的服务(locationbasedservice,LBS)的逐渐推广,有着越来越大的需求。然而,由于室内环境中往往无法很好接收全球定位系统(globalpositioningsystem,GPS)的卫星定位信...
