人民珠江2025，Vol.46Issue(7)：101-114,14.DOI:10.3969/j.issn.1001-9235.2025.07.010
基于高光谱特征的土壤水力特性参量反演方法研究
Inversion Methods of Soil Hydraulic Parameters Based on Hyperspectral Characteristics
摘要
Abstract
To enrich remote inversion methods of soil hydraulic parameters and achieve a fast and nondestructive prediction,59 high-spectral in-situ soil samples in Xiangzhou District,Zhuhai City,were collected using Field Spec4 spectrometer.The remote hydraulic parameter models were established for two land types:grassland and bare land(saturated water content θs,residual water content θr,inverse of air entry value α,shape coefficient n,saturated hydraulic conductivity Ks,and soil moisture content θ).The results are as follows.① The spectral curves exhibit clear linearity in three bands:700～750 nm,830～1 100 nm,and 1 520～1 620 nm,with the mean values of coefficient of determination for the linear fitting R2 all over 0.94.② Among gradient boosting regression(GBR),partial least squares regression(PLSR),and random forest(RF),GBR performs the best and shows high sensitivity to the linear fitting parameter lga2(the logarithm of the slope)in the 830-1100 nm band.It accurately predicts θs,α,n,and θ for grassland types(Relative Percent Deviation,RPD>1.4).For the bare land types,the coefficient of determination RP² for all hydraulic characteristic parameters,except soil moisture content θ,is greater than 0.9,and RPD exceeds 2.0.③ In most hydraulic characteristic models predicting the grassland types,the feature importance of the leaf area index(LAI)is the highest.Therefore,an effective inversion model is built based on field spectral data and LAI,with simple access to input parameters and good prediction effect,providing a potential for wide application in the future.关键词
非饱和土/水力特性/高光谱/机器学习/遥感反演Key words
unsaturated soils/hydraulic parameters/spectrum/machine learning/remote sensing inversion分类
建筑与水利引用本文复制引用
吴炎镅,朱涵,李佶,李紫昕,何健良,林凯荣,庄鲁文,董春雨..基于高光谱特征的土壤水力特性参量反演方法研究[J].人民珠江,2025,46(7):101-114,14.基金项目
国家自然科学基金项目(42377057、42007165) （42377057、42007165）
广东省基础与应用基础研究基金(2023A1515030122、2023B1515040028) （2023A1515030122、2023B1515040028）
广东省水文局项目(440001-2023-10716) （440001-2023-10716）
