2023AutonomousandAutonomousandControllableAIIndustryenablingplatformChina’sThe?rstfullyopensource,fullyfunctionalindustrialleveldeep-learningplatformOpen,complete,andsecureautonomousdrivingsoftwareandhardwareintegratedsolutions,openplatform,andecosystemTian-SuanDataLakeanalysisplatformTian-GongEdgeconvergenceIOTPlatformTian-HeCloudnativedevelopmentplatformTian-LianCloudareaplatformTian-XiangIntelligentmultimediaplatformKai-WuIndustrialInternetPlatformChina’s?rstfull-functionAIchipwithHighestPower/HighCost-Effective/EasytoUseHong-HuFar-?eldvoiceinteractionchip,automotive-gradestandard,ultra-largememory,lowpowerconsumption#encoding:utf-8importcvimportcv2face_cascade=cv2.CascadeClassi?er('haarcascade_?les/haarcascade_frontalface_default.xml')eye_cascade=cv2.CascadeClassi?er('haarcascade_?les/haarcascade_eye.xml')img=cv2.imread('west.jpeg')gray=cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)#轉(zhuǎn)為灰度圖importpaddlehubashubmodule=hub.Module(name="ulmb_640")res=module.face_detection(paths=["./test.jpg"],visualization=True,output_dir='face_detection_output')#檢測臉部faces=face_cascade.detectMultiScale(gray,CoordinateBoxSelectionCoordinateBoxSelectionminNeighbors=5,minSize=(30,30),MatrixOperationprint('Detected',len(faces),"face")MatrixOperationHaarFeatureExtractionHaarFeatureExtractionFaceRecognitionRecursiveOperationImageReatingRecursiveOperationImageReatingAdaboostTrainingAdaboostTraining1000000+DataTrainingcv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,1000000+DataTrainingImageIntegrationlabel='Result:Detected'+str(len(faces))+"facesImageIntegration0.8,(0,0,0),1)10000+Data0.8,(0,0,0),1)10000+DataAcquisitioncv2.imshow('img',img)cv2.destroyAllWindows()OpenCVEnvironmentalInstallationIndustryDevelopmentCenterIntelligentIndustrialiIndustryDevelopmentCenterIntelligentIndustrializationIndustrialIntelligence【TechnologicalEmpowerment】enterprise【ProjectCooperation】【Scienti?cResearchCooperation】GovernmentConnectingPlatformAIIndustryCrowdsourcingPlatformObtainProjectOpportunities｜Supply-DemandInformationConnecting|ProjectManagementAIProjectVisualisationAIProjectVisualisationDevelopmentPlatformProjectProgressManagement｜ProjectProgressTracingProcessTechnicalSupportTianGong/KaiWu/SmartProductionLine/IndustrialVision/ElectricalAutomation/IndustrialControlUnitDemandSideEco-EnterprisesDemandSideEco-EnterprisesAIIndustryStandardisedTradingPlatformApplicationScenarioReplication｜StandardisedProductTrading|IntellectualPropertyTransferAITalentCommunityandPrecisionServiceAITalentCommunityandPrecisionServicePlatformIndustry-educationintegrationtraining|professionaltalentconnecting|technicalexchangecommunityEasyData/EasyEdge/UNITAIAIIndustryDevelopmentCenterProfessionalTechniciansManufacturingEmpowerment92%GovernmentProcurement49%Case8% 1TextileIndustryAINewExperienceinFabricInspection 2AgricultureYingdeRedTeaWitheringProcessPractice 3AnimalHusbandry 4ConstructionIndustryAIProtectingTowerCraneSafety 5UrbanManagementAll-weatherRoadDiseaseInspectionTextileIndustryAINewExperienceinFabricInspection?LowE?ciency：Manualfabricinspectiontimeisabout15meters/minute,andrepetitivetaskssuchasmarkingandrecordingdatatakemoretime.Traditionalreportsrequiremanualcalculationand?PoorQuality：Poordetectionofdefects,manualinspectionispronetofatigueandsubjectiveerrors,withanaveragedetectionrateofabout70%.；?HighCost：Highrecruitment,management,andtrainingcosts.Informationrecognizedbyhumanscannotbeeffectivelytransmitted,andtheindustryfacespainpointssuchasforminguni?edstandards.ThissolutionThissolutioniswidelyapplicabletofabricqualityinspectionatvariousstagesoftextileindustrymanufacturing,printinganddyeing,garmentmaking,etc.,andissuitableforinspectionofsurfacedefectsandcolordi?erencesofknittedandwovenplainfabrics.Moreaccurate,fasterinspectione?ciency,andlowerinspectioncostsmakeintelligentfabricinspectionthebestchoiceforthefuturetextileindustry.?ImageAcquisitionModulecollectsimagedataoffabrics,includingcolor,texture,shape,etc?ImageProcessingModulepreprocessesthecollectedimages,includingdetailenhancement,noiseremoval,contrastadjustment,etc.comprehensivelyjudgesthequalityandquali?cationoffabricsbasedonfactorssuchasfabricmaterial,color,size,andhistoricalinspectiondata,andgeneratesinspectionreportsthatmeetrequirements.2Agriculture2ProcessPracticeYingdeProcessPracticeGuangdongHongyanTeaIndustryCo.,Ltd.isanimportantenterpriseintheYingderedteaindustryandatechnologytransformationplatformoftheTeaResearchInstituteofGuangdongAcademyofAgriculturalSciences.Byrelyingontechnologicaladvantages,itiscommittedtoR&Dandproduction,representingthehighestlevelofspecialtyfamousteaproductsinGuangdong.However,HongyanTeaalsofacesindustrypainpointssuchastightlaborinteagardens,inconsistentpickingstandardsforfreshleaves,unevenwitheringstandardsforredtea,andcontradictionsbetweenproductstandardsandteagardenproduction.Witheringisanimportantprocessforformingthequalityofblacktea.Currently,intraditionalwitheringprocesses,freshleavesarespreadonwitheringtroughs,andproductionpersonnelcontroltheairvolumeandtimeofthewitheringtroughblowertowitherthefreshleaves.Sincedifferentteamakershavedifferentjudgmentsontheactivityofwithering,itdirectlyaffectsthequalityofeachbatchoftealeaves.BycombiningBycombininginfraredthermalimagingtechnology,AIintelligentrecognitionmodels,andhigh-de?nitioncameramonitoringtechnologytoformintelligentwitheringequipment,itisappliedtoteawithering.Theblower'sstartandstopstateandairvolumesizeareautomaticallycontrolledbasedonthechangesofvariousfactorsinthewitheringprocessoffreshleaves,accuratelymasteringthewitheringstandardsoftealeaves."SmartTeaProcessing"isoneofthekeyresearchandimplementationobjectsinthefutureteaindustry.Therefore,theestablishmentandapplicationofintelligentwitheringequipmenthavegreatprospects.Moreover,byestablishingmodelsforthestandardstateofappropriatewitheringoftealeaves,itcanmoreaccuratelydeterminethereal-timestateoffreshleafwithering.Additionally,thewitheringequipmentcanautomaticallycontroltheblower'sstartandstopstatebasedonthewatercontentandphysicalstateofthefreshleaves,trulyrealizing"fullyintelligentwitheringoffreshleaves."3AnimalHusbandry3Intelligent"IDCard"forDairyCowsTheclientfocusesonmovablepropertypledgesupervisionbusiness,specializinginmovablepropertypledgesupervisionbusiness,dedicatedtoprovidingenterpriseswithconvenient,high-quality,safe,andef?cientsupervisionservices.Currently,ranchsupervisionmainlyreliesontraditionalhumanmonitoring,supplementedbyconventionalsupervisionsystems,stillrequiringmanualcountingandcompletionofstatisticalreports.?AutomatedImageQualityScreening?CreateaVisualMonitoringPlatform?StaticCountingFunction?DynamicTrackingFunction4ConstructionIndustry4AIProtectingTowerCraneSafetyBasedontheBasedontheinfrastructure,deeplearningframework,andCValgorithmmodelconstructionofBaiduArti?cialIntelligenceIndustryEmpowermentCenter,includingtheapplicationofimagetargetdetection,imageenhancementandsegmentation,keypointrecognition,etc.,arti?cialintelligencescenarioapplicationisempoweredtotheexistingtowercraneconstructionmachinerysafetymanagementsystem.5UrbanManagement5All-weatherRoadDefectInspectioninrecentyears,withanunprecedentedgrowthrate.However,duetotherapidincreaseinthenumberofmotorvehicles,theda