智能駕駛環(huán)境感知系統(tǒng)認(rèn)知KnowledgeofIntelligentDrivingEnvironmentalPerceptionSystem導(dǎo)入Import如何實(shí)現(xiàn)周圍環(huán)境的感知？Howtoperceivethesurroundingenvironment?PART01環(huán)境感知EnvironmentalPerception環(huán)境感知技術(shù)的感與知SensingandPerceptionofEnvironmentalSensingTechnology作為自動(dòng)駕駛的第一個(gè)環(huán)節(jié)，環(huán)境感知是智能車輛與外界環(huán)境信息交互的關(guān)鍵，其核心在于使智能駕駛車輛更好地模擬、最終超越人類駕駛員的感知能力，準(zhǔn)確的感知并理解車輛自身和周邊環(huán)境的駕駛態(tài)勢(shì)。Asthefirststepofautonomousdriving,environmentalperceptionisthekeytotheinformationinteractionbetweenanintelligentvehicleandtheexternalenvironment.Thecoreliesinenablingintelligentvehiclestobettersimulateandultimatelysurpasstheperceptionabilityofhumandrivers,andaccuratelyperceiveandunderstandthedrivingsituationofthevehiclesandthesurroundingenvironment.環(huán)境感知技術(shù)的感與知SensingandPerceptionofEnvironmentalSensingTechnology智能汽車環(huán)境感知技術(shù)

Intelligentvehicleenvironmentalperceptiontechnology感：傳感器技術(shù)

Sensing:sensortechnology知：識(shí)別技術(shù)

Perception:recognitiontechnology超聲波技術(shù)Ultrasonictechnology毫米波雷達(dá)技術(shù)Millimeterwaveradartechnology激光雷達(dá)技術(shù)

LiDARtechnology夜視技術(shù)

NightvisiontechnologyCCD技術(shù)

CCDtechnology數(shù)據(jù)處理

Dataprocessing數(shù)據(jù)分析

Dataanalysis數(shù)據(jù)理解

Dataunderstanding深度學(xué)習(xí)技術(shù)

Deeplearningtechnology經(jīng)典模式識(shí)別技術(shù)

Classicpatternrecognitiontechnology被動(dòng)紅外技術(shù)

Passiveinfraredtechnology主動(dòng)紅外技術(shù)

ActiveinfraredtechnologyPART02傳感器技術(shù)的不同類型DifferentTypesofSensorTechnologies超聲波雷達(dá)UltrasonicRadar超聲波雷達(dá)是利用超聲波的特性研制而成的傳感器，是在超聲波頻率范圍內(nèi)將交變的電信號(hào)轉(zhuǎn)換成聲信號(hào)或?qū)⑼饨缏晥?chǎng)中的聲信號(hào)轉(zhuǎn)換為電信號(hào)的能量轉(zhuǎn)換器件。超聲波雷達(dá)信號(hào)如圖所示。Ultrasonicradarisasensordevelopedusingthefeaturesofultrasound.Itisanenergyconversiondevicethatconvertsalternatingelectricalsignalsintoacousticsignalswithintheultrasonicfrequencyrangeorconvertsacousticsignalsintheexternalsoundfieldintoelectricalsignals.Theultrasonicradarsignalisshowninthefigure.超聲波雷達(dá)UltrasonicRadar毫米波雷達(dá)Millimeterwaveradar毫米波雷達(dá)是工作在毫米波頻段的雷達(dá)，它通過(guò)發(fā)射與接收高頻電磁波來(lái)探測(cè)目標(biāo)，后端信號(hào)處理模塊利用回波信號(hào)計(jì)算出目標(biāo)的距離、速度和角度等信息。Millimeterwaveradarisatypeofradarthatoperatesinthemillimeterwavefrequencyband.Itemitsandreceiveshigh-frequencyelectromagneticwavestodetecttargets.Thesignalprocessingmoduleattherearendusestheechosignalstocalculateinformationsuchasthedistance,speed,andangleofthetarget.毫米波雷達(dá)的定義及其應(yīng)用DefinitionandApplicationsofMillimeterWaveRadar毫米波雷達(dá)及其應(yīng)用

MillimeterWaveRadarandItsApplication毫米波雷達(dá)Millimeterwaveradar毫米波雷達(dá)

Millimeterwaveradar24G赫茲：用于盲點(diǎn)監(jiān)測(cè)和變道輔助24GHzForblindspotmonitoringandlanechangeassistance77G赫茲：用于實(shí)現(xiàn)緊急制動(dòng)和自適應(yīng)巡航77GHzForachievingemergencybrakingandadaptivecruisecontrol毫米波雷達(dá)Millimeterwaveradar激光雷達(dá)通過(guò)向目標(biāo)發(fā)射激光束，然后將接收到的從反射目標(biāo)返回來(lái)的信號(hào)與發(fā)射信號(hào)進(jìn)行對(duì)比，做適當(dāng)處理后，就可以得到目標(biāo)的距離，方位，高度，速度，姿態(tài)，形狀等參數(shù)。LiDARemitsalaserbeamtothetarget,andthencomparesthereceivedsignalreturnedfromthereflectiontargetwiththetransmittedsignal.Afterproperprocessing,thetarget'sdistance,azimuth,altitude,speed,attitude,shapeandotherparameterscanbeobtained.激光雷達(dá)Laserradar激光雷達(dá)

Laserradar通過(guò)掃描可以得到周圍環(huán)境的三維模型，運(yùn)用相關(guān)算法對(duì)比上一幀和下一幀環(huán)境的變化，可以輕松探測(cè)出車輛和行人，并進(jìn)行障礙物的追蹤A3Dmodelofthesurroundingenvironmentcanbeobtainedbyscanning.Byusingrelevantalgorithmstocomparethechangesintheenvironmentbetweenthepreviousframeandthenextframe,vehiclesandpedestrianscanbeeasilydetectedandobstaclescanbetracked.三維環(huán)境感知3Denvironmentalperception指激光雷達(dá)通過(guò)掃描到的點(diǎn)云數(shù)據(jù)實(shí)現(xiàn)同步創(chuàng)建地圖。TheLiDARusesthescannedpointclouddatatorealizethesynchronouscreationofmaps.slam加強(qiáng)定位slamenhancedpositioning夜視技術(shù)Nightvisiontechnology一種介于光電成像器件實(shí)現(xiàn)夜間傳感器的光電技術(shù)，其分為主動(dòng)紅外夜視技術(shù)和被動(dòng)紅外夜視技術(shù)。Itisanoptoelectronictechnologythatrealizesnightsensorsbetweenoptoelectronicimagingdevices,includingactiveinfrarednightvisiontechnologyandpassiveinfrarednightvisiontechnology.CCDCCD成像原理是當(dāng)光線與圖像從鏡頭透過(guò)投射到CCD表面時(shí)，CCD就會(huì)產(chǎn)生電流，將感應(yīng)到的內(nèi)容轉(zhuǎn)換成數(shù)碼資料儲(chǔ)存起來(lái)。CCD像素?cái)?shù)目越多，單一像素尺寸越大，收集到的圖像就會(huì)越清晰。TheimagingprincipleisthatwhenlightandimagesareprojectedfromthelensontothesurfaceoftheCCD,theCCDwillgenerateanelectriccurrenttoconvertthesensedcontentintodigitaldataforstorage.MoreCCDpixelsandlargersizeofasinglepixelresultinclearercollectedimage.數(shù)據(jù)識(shí)別技術(shù)DataIdentificationTechnology?數(shù)據(jù)處理：從大量的，可能是雜亂無(wú)章的，難以理解的數(shù)據(jù)中抽取并推導(dǎo)出對(duì)于某些特定的人們來(lái)說(shuō)是有價(jià)值的，有意義的數(shù)據(jù)。Dataprocessing:Itextractsandderivesdatathatisvaluableandmeaningfultocertainspecificpeoplefromalargeamountofdatathatmaybedisorganizedanddifficulttounderstand.?數(shù)據(jù)分析：適當(dāng)?shù)慕y(tǒng)計(jì)方法對(duì)收集來(lái)的大量數(shù)據(jù)進(jìn)行分析，將它們加以匯總和理解，以求最大化的開(kāi)發(fā)數(shù)據(jù)的功能，發(fā)揮數(shù)據(jù)的作用。Dataanalysis:Itusesappropriatestatisticalmethodstoanalyzethelargeamountofdatacollected,summarizeandunderstandtheminordertomaximizethedevelopmentofdatafu