基于高光谱数据结合BorutaShap_CatBoost算法的柑橘叶面积指数反演

打开文本图片集

Abstract：InordertomakefulluseofhyperspectraldataandimprovetheaccracyofcitrusLAIinversion，thispapertakedcitrusas theresearchobject，usedhyperspectraldata toconstructthefeatureset，usedBorutaShaptoompareBorutaforfeaturesreening，and combinedCatosdG，ndespecielyndfourgeioodelstsuctightbatiofels， andthenadjuststheparametersthrough Bayesianadjustment.AftertheBayesianadjustment，theoptimalmodelforinvesionofcitrus LAIwas selectedthroughtheaccuracyassessmentoftheregressionresults.Theresultsshowedthat BorutaShaphadfewerbandsand better performance than Boruta.Among the eight combined models，BorutaShap ⊢ CatBoost was the best model for inverting citrus LAI， andthecoefntsofeteraoftitedodelndealiatedmodelere8812nd.6，esptivelychowed thatBorutaShap_CatBostcowldaccuratelypredictcitrusLAIand generate LAIinversionmaps toprovidedatasupportforrapid monitoringandmanagementofcitrusfruitreegrowth.Inconclusion，TheinverseLAImapanprovidedatasupportfortherapidmoitor ing and management of citrus fruit trees.

Keywords：UAVhyperspectral; BorutaShap;CatBoost;leaf Area Index

中国是全球最大的柑橘出口和消费国之一，推动柑橘产业的精细化管理以提高柑橘的质量、产量，对实现农民的脱贫致富、助推产业兴旺、实现乡村振兴具有重要意义。（剩余11314字）