智慧蔗田“空－天－地"协同智能监控系统设计与应用

打开文本图片集

中图分类号：S126 文献标志码：A 文章编号：2096-9902（2025）24-0006-05

Abstract：Inviewoftheproblemsofinsuffcientmonitoringaccuracy，incompletedatacoverageandlowintelligencelevelin theconstructionofadigitalmanagementsystemforGuangxi'ssugarcaneindustrythisstudyproposedandimplementedaetof "air-sky-ground"integratedcollborativeintellgentmonitoringsystembasedondroneremotesensingtechnologyanddeep learningmethods.Thesystemreliesontheexisting infrastructureofthetowerbase，combinesdroneautonomousairportaerial photographytechnologywithartificialinteligenceimagesegmentationalgorithm，mediumandhighpointvideosurveillnce technology，Iteetofingssensingmitoringtecogtc.touilda"pereptiotransisio-platfoaplicati"four layerarchitecturesystem，andintegratescollaborativedronetaskmanagement，orthophotoanalysis，imagesegmentationand quantitativeevaluationmethodsbasedontheimprovedU-Netmodelandmulti-terminalcolaborativedecision-makingtobuild anintegratedcoremonitoringfunctionmodule.TheoverallsystemadoptsB/Sarchitecture，isdesignedseparatelyfromthefront andbackends，andusesadistributeddatabasetoachieveeficientstorageandaccessoflarge-scalecoleteddata.Theactual measurement results show thatthe coverage radius of single drone tower base monitoring and acquisition can reach 2.5km （about 1963.5hectares），coveringatotalareaofabout3OOoOacres，andtheerorofautomaticareacalculationofsugarcanefieldsis controlled within 5% . Theresearchresultsshow that thesystem has outstanding performance in field identification accuracy，data visualizationandagriculturalmanagementeficiencyThiscanprovidedatasupportforprecisemanagementoftheentireproce， allfactorsandallweatherconditionsofsugarcanefieldproduction，andpromotethedigitaliationofGuangxi'ssugarcane industry，with important application value and promotion prospect for intelligent construction.

Keywords：remotesensing technology；deeplearning；crop identification；areameasurement；yieldestimation；image segmentation

广西作为我国糖料蔗产业主要地区，当前糖料蔗“双高"（高产、高糖）基地面积已超1100万亩（1亩约等于 667m2 ，下同）。（剩余4906字）