Prediction of Soil pH Hyperspectral Spectrum in Guanzhong Area of Shaanxi Province Based on PLS

1st International Conference on Frontiers of Materials Synthesis and Processing

材料科学;化学

Liu, Jinbao^1,2,3,4 ; Zhang, Yang^1,2,3,4 ; Wang, Huanyuan^1,2,3,4 ; Cheng, Jie^1,2,3,4 ; Tong, Wei^1,2,3,4 ; Wei, Jing^1,2,3,4

Institute of Land Engineering and Technology, Shaanxi Provincial Land Engineering Construction Group Co. Ltd., Xian

710075, China^1

Key Laboratory of Degraded and Unused Land Consolidation Engineering, Ministry of Land and Resources, Xian

710075, China^2

Shaanxi Provincial Land Engineering Construction Group Co. Ltd, Xian

710075, China^3

Shaanxi Provincial Land Consolidation Engineering Technology Research Center, Xian

710075, China^4

关键词: First-order differentials;    Hyperspectral spectrum;    Partial least-squares method;    Root mean square errors;    Root-means-square errors;    Spectral characteristics;    Spectral prediction model;    Spectral reflectances;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/274/1/012020/pdf
DOI  :  10.1088/1757-899X/274/1/012020

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

The soil pH of Fufeng County, Yangling County and Wugong County in Shaanxi Province was studied. The spectral reflectance was measured by ASD Field Spec HR portable terrain spectrum, and its spectral characteristics were analyzed. The first deviation of the original spectral reflectance of the soil, the second deviation, the logarithm of the reciprocal logarithm, the first order differential of the reciprocal logarithm and the second order differential of the reciprocal logarithm were used to establish the soil pH Spectral prediction model. The results showed that the correlation between the reflectance spectra after SNV pre-treatment and the soil pH was significantly improved. The optimal prediction model of soil pH established by partial least squares method was a prediction model based on the first order differential of the reciprocal logarithm of spectral reflectance. The principal component factor was 10, the decision coefficient Rc2 = 0.9959, the model root means square error RMSEC = 0.0076, the correction deviation SEC = 0.0077; the verification decision coefficient Rv2 = 0.9893, the predicted root mean square error RMSEP = 0.0157, The deviation of SEP = 0.0160, the model was stable, the fitting ability and the prediction ability were high, and the soil pH can be measured quickly.