基于车辆声频信号的车型识别算法研究

《基于车辆声频信号的车型识别算法研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、y978‘l二长安大学硕士学位论文基壬奎箍直叛值墨数奎型迟别篡洼硒究瞳鱼申请学位级别塑主学科名称叁煎焦皇苎塑垒垫剑论文提交FI期2QQ6。区论文答辩日期2QQ互生互旦iQ目AbstractVehicleclassificatjonisanimportantpartofintelli2enttransportsystem．Itisusedintoll、transponationstatistjcsandsoonfafandwide．Ttaditionally'electriccablesandjnductioncoilsareburiedund

2、erroads；vehicles’photostakenbycamerarecordsystem．However，thismethodwillcausedamagetotheroad，柚ddemandequipmentmaintenancefbrlongte哪use．ThisthesisisfbcusedonthestudyofthevehicleclassificatjOnbasedonthea∞usticsignaI．First，theDfesentrcsearchsituationofsounddetectionanditsfeasib

3、iIitvwasreviewed．Thenthegenerationandpmpagationofacousticsi2皿alwa击studied．ThisthesiscollectsmanVkindsofsignal，de-noiseandsynthesizespecimenusingthetheoryofwaveletaIldarraVsignal．AndthenextractfeatufesfromsignalIFeatureexIractionisthekevtotheclassificationprocess．Intheapproa

4、ch，featuresafeextractedfromspectr02rambaseontraditionalmethods．Spectmgr锄denotesthespectmmbywayofpictureintwodimensions：timeand仃equency．Experimentsshowthatthercarcmanydi丘brenthndsofcolorandtexturcforpowerinspectm盯amcontmsttothepowerspectnlm．Thisthesisextradthecontmst、entropy

5、aIldene唱yasourfeaturcs．Thelastpanofthisstudyinvestigatedpattemrcco印itionproblemovertexturefcaturcs．AimatthecharacteristicSofthesignalthethesisadoptfIlzzVc．meansal舯ritllIll．Forcomplexfeaturesthethesisistakenanewa190rithmcaIledKemel-basedfIlzzVandpossibilisticc—meansclusterin

6、2waStaken．Someresultsare西ventoilIustrateIheadvantagcsofthepmposedalgorithmsoVertheFCM(fIIzzyc·me卸s)aIldPCM(possibilisticc—means)a120rithms．Thea190rithmofthisthesisisef：fectivetoclassthevehiclesbaseonthevehideacousticsigllal．Usingtheacousticnoisesi印algeneratedbydriVillgmotor

7、VehiclestovehicleclassificatiOnisanewandpromisingmethod．Keywords：vehicleacousticsi弘a1．vehicleclassification．spectm莎lIIltextufctheory．Kemel．basedfuzzy锄dpossibilisticc—meansclustering论文独创性声明本人声明：本人所呈交的学位论文是在导师的指导下，独立进行研究工作所取得的成果。除论文中已经注明引用的内容外，对论文的研究做出重要贡献的个人和集体，均已在文中以明确方式标明。

8、本论文中不包含任何未加明确注明的其他个人或集体已经公开发表的成果。本声明的法律责任由本人承担。论文作一1撕游钿石日论文知识产权权属声明本人在导师指导下所完成的论文及