Page 47 - 水利学报2025年第56卷第4期
P. 47
thelong - termeffectsduringthegrowingstages.Theresultsillustratethatreducingthecontroldepthofdrainagede
creasesgroundwaterdepth,drainageamount,anddurationwhileincreasingsoilmoisturecontent.Themostsignif
icantvariationoccursafterwaterloggingandduringthebranchingstage.Afterwaterlogging ,theplantheight,leaf
areaindex(LAI),leafchlorophyllcontent(SPAD),netphotosyntheticrate(Pn),stomatalconductance(Gs)
andtranspirationratedecreasefirstandthengraduallyrecovercomparedwiththecontrol.Themaximum relative
valueofSPAD,PnandGsafterrecoveryandtheincrementofrecoveryperiodarethelargestatthecontrolled
depthof50cm,andtherecoverydurationofSPADistheshortest.Atflowering - poddingstage,thecompensation
ofplantheight,LAI,PnandGsarethelargest,andbothPnandGscompensationarethemostobviousatthe
controlleddepthof50cm.Theabove - grounddrymatteraccumulationisreducedbywaterlogging ,andthereduc
tionfollowtheorderofleaf<fruit<pod,andtheyieldreductionismainlyduetothedecreaseofpodnumberper
plant.Thereductionoftheabove - grounddrymatter ,yieldandyieldcompositionarethesmallestatthecontrolled
depthof50cm.Controlleddrainagedepthof50cm,willprovidesuitablesoilwaterenvironmentforsoybean
growthinthesandyloamareaofHuaibeiPlain.
Keywords:drainagecontrol;afterwaterlogging;soilmoisture;growthstage;yield
(责任编辑:王 婧)
(上接第 454页)
Visualmonitoringmethodfortherollingandpavingstatusofthe
rollercompactedconcretedam
1
2
1
1
1
1
SONGBenyang,CUIBo,ZHANGTianwei,WUBinping,RENBingyu,TONGDawei
(1.StateKeyLaboratoryofHydraulicEngineeringIntelligentConstructionandOperation,TianjinUniversity,Tianjin 300072,China;
2.HuanengLancangRiverHydropowerInc.,Kunming 650200,China)
Abstract:Thecontrolofconstructiontimeforrollercompactedconcrete(RCC)dam significantlyimpactsthe
qualityofpouring.However ,theenvironmentofRCCdamsiscomplex,andthetraditionalmethodofassigning
dedicatedpersonneltomanuallyrecordatfixedtimesandlocationsfaceschallengessuchashighpersonnelcosts ,
lowefficiency,anddifficultiesinsafetymanagement.Toaddresstheaboveissues,thisarticleproposesavisual -
basedintelligentmonitoringandanalysismethodfortherollingandpavingstatusofRCCdams.Usingavideosur
veillancesystemarrangedontheconstructionsitetoobtainimagedataofconstruction,theQueryInstdeeplearn
ingalgorithmautomaticallyidentifiestheconstructionstatusofvariousareasonthedamsurface.Consideringthe
relationshipoftheimagedata ,statisticalanalysisisconductedontheconstructionstageandconstructiontimeof
eachstageonsite.ThismethodhasbeenappliedtothemonitoringprocessoftheconstructionstatusofaRCCdam
inSouthwestChina.Theresultsshowthattheproposedmethodcanaccuratelyandeffectivelyidentifythecon
structionstatusofvariousareasonthewarehousesurfaceandcalculatethedurationofeachconstructionstage ,
providingrichdatasupportandreferencebasisforbettermanagementlevelatconstructionsites.
Keywords:rollercompactedconcretedams; constructionofhydraulicengineering; videomonitoring; deep
learning ;computervision;rollingconstruction
(责任编辑:韩 昆)
— 4 6 5 —