学术会议PPT

ATraffic-congestionDetectionMethodforBadWeatherBasedonTrafficVideoPurposeCurrentproblem:TheresultoftrafficcongestiondetectioninbadweatherisinaccurateOurpurpose:ATraffic-congestionDetectionMethodforBadWeatherIntroductionTrafficcongestiondetectionisofgreatsignificanceincityvehicle,roaddesigning,trafficlightssetting,preventingtrafficcongestionandotherapplicationfields.Therefore,thetrafficcongestiondetectionplaysanimportantroleintransportationfield.Unfortunately,theresultsofanautomaticvehicleidentificationsystemfordetectingtrafficispronetoappearhighfalse-negativerateinbadweathersuchasfog,mist,andrain.Itmayleadcauseoferrorsinstatisticsoftrafficcongestionindexandincreasestheriskoreconomiclossinrelevantdepartmentsandindustry.OurMethodFirstly,thismethodusesdiscrete-framedifferencealgorithmtoextracttheimagesthathavevehicleinformation.Secondly,thismethodusesthehistogramequalizationalgorithmtoeliminatethenoiseoftheimages.Finally,thismethodrecognizesthevehiclefromthevideoandcomputesthetrafficcongestionindexbythecalculationmethodbasedondiscrete-framedifference.MainAlgorithmsImageextractionmethodDiscrete-framealgorithmImagede-noisingmethodHistogramequalizationalgorithmVehicleidentificationmethodBackgroundsubtractionalgorithmCalculationofTrafficcongestionindexAmethodconsiderednumberandspeedMainAlgorithmsCalculationofTrafficcongestionindexAmethodconsiderednumberandspeedWrepresentsthetotalnumberofidentifiedvehicles.k1representsthetotalnumberofframesofvideosectioncontainingvehicleinformation.Cirepresentsthenumberofidentifiedvehiclesiniframesinthevideothathasvehicleinformation.Wmaxrepresentsthemaximumnumberofidentifiedvehicles.Srepresentsthelengthofthevideodetectionregion.Zrepresentsthenumberofintervalframesindiscrete-framealgorithm.Tmaxrepresentsthetotaltimeofthevideo.trepresentsthedifferenceoftwoframes.mrepresentstheminimumlengthofavehicle.ρyrepresentsthetrafficcongestionindexwithoutunitconversion.ExperimentsExperimentofImagesProcessingTimeandeffectsComparisonExperimentofDiscrete-frameMethodtoExtractTrafficImagesTimetestingExperimentofVehicleIdentificationAccuracytestingResultsResultsResultsThankyou