基于数学变换和连续小波变换的苹果园土壤总氮含量高光谱反演

打开文本图片集

中图分类号：S153;S661.1 文献标识码：A

文章编号：0439-8114（2025）11-0165-06

DOI：10.14088/j.cnki.issn0439-8114.2025.11.022 开放科学（资源服务）标识码（OSID）：口

Hyperspectral inversion of soil total nitrogen content in an apple orchard based on mathematical transformation and continuous wavelet transform

XIONG Chao-qun，XUNMi，LIJian，ANMiao，LI Guo-tian （ShandongInstituteof Pomology，Tai'an 2710O1，Shandong，China）

Abstract：Torapidlydeterminethesoiltotalnitrogen（T）contentinanappleorchardandacheveprecisefertilization，soilromthe appleorchardattheTanpingLakeBaseoftheShandongInstituteofPomologyastakenas theresearchbject.Itsspectralreflectance （R）wasmeasuredandsubjectedtomathematicaltransformation（reciprocal，logarithm，squareroot，first-orderderivative， etc.），continuous wavelet transform（CWT），andcombinedmathematical tansformationandCWTprocesing.ThePearsoncoelationanalysismethodasusedforfeatureetraction，andahyperspectralinversionmodelforsolTcontentwasonstructedbasedon support vectorregresson（SVR）.Theresultsshowedthatafterprocesingtheoriginalspectraldata withCWT，witinthescalerange of 21 to 210 ，thecorrelationbetweenthewaveletcoeffcientsandsoilTNcontentfirstincreasedandthendecreased.MediumscaleseffectivelysuppressednoiseinterferenceandenhancedthecorelationbetweenthespectrumandTN，withasignificanteffect. BothmathematicaltransformationandCWTefectivelyminedthedetailedfeaturesofthespectrum，withtheefectofCWTgenealy beingsuperiortothatofmathematicaltransformation.ThecombinationofmathematicaltransformationandCWTsignificantlyimproved the model inversion accuracy.Among them，the 1/R-CWT- 28 model demonstrated the best performance （ R2 =0.73，RMSE=0.12 g/kg， RPD=1.85 ）.The fiting curve between the soilTN content predicted by this model and the measuredvalues was closer to the 1：1 line，indicating highodelpredictionaccuracy.Inonlusion，hyperspectralechnologycouldbeusedasanondestructiveesting methodforsoilTNcontentinappleorchards.Thehyperspectralinversionmodelconstructedbycombining mathematical transforation and CWT could more accurately predict soil TN content.

Keywords：mathematicaltransformation;continuous wavelet transform（CWT）；appleorchard;soil；totalnitrogencontent;hyperspectral; inversion

苹果富含矿物质和维生素，是温带地区重要的经济果树。（剩余8600字）