Estimation of Soil Salinity by Combining Spectral and Texture Information from UAV Multispectral Images in the Tarim River Basin, China DOI Creative Commons

Jiaxiang Zhai,

Nan Wang, Bifeng Hu

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(19), P. 3671 - 3671

Published: Oct. 1, 2024

Texture features have been consistently overlooked in digital soil mapping, especially salinization mapping. This study aims to clarify how leverage texture information for monitoring through remote sensing techniques. We propose a novel method estimating salinity content (SSC) that combines spectral and from unmanned aerial vehicle (UAV) images. Reflectance, index, one-dimensional (OD) were extracted UAV Building on the features, we constructed two-dimensional (TD) three-dimensional (THD) indices. The technique of Recursive Feature Elimination (RFE) was used feature selection. Models estimation built using three distinct methodologies: Random Forest (RF), Partial Least Squares Regression (PLSR), Convolutional Neural Network (CNN). Spatial distribution maps then generated each model. effectiveness proposed is confirmed utilization 240 surface samples gathered an arid region northwest China, specifically Xinjiang, characterized by sparse vegetation. Among all indices, TDTeI1 has highest correlation with SSC (|r| = 0.86). After adding multidimensional information, R2 RF model increased 0.76 0.90, improvement 18%. models, outperforms PLSR CNN. model, which (SOTT), achieves RMSE 5.13 g kg−1, RPD 3.12. contributes 44.8% prediction, contributions TD THD indices 19.3% 20.2%, respectively. confirms great potential introducing semi-arid regions.

Language: Английский

Fine-resolution baseline maps of soil nutrients in farmland of Jiangxi Province using digital soil mapping and interpretable machine learning DOI
Bifeng Hu, Yibo Geng,

Kejian Shi

et al.

CATENA, Journal Year: 2024, Volume and Issue: 249, P. 108635 - 108635

Published: Dec. 9, 2024

Language: Английский

Citations

1
Estimation of Soil Salinity by Combining Spectral and Texture Information from UAV Multispectral Images in the Tarim River Basin, China DOI Creative Commons

Jiaxiang Zhai,

Nan Wang, Bifeng Hu

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(19), P. 3671 - 3671

Published: Oct. 1, 2024

Texture features have been consistently overlooked in digital soil mapping, especially salinization mapping. This study aims to clarify how leverage texture information for monitoring through remote sensing techniques. We propose a novel method estimating salinity content (SSC) that combines spectral and from unmanned aerial vehicle (UAV) images. Reflectance, index, one-dimensional (OD) were extracted UAV Building on the features, we constructed two-dimensional (TD) three-dimensional (THD) indices. The technique of Recursive Feature Elimination (RFE) was used feature selection. Models estimation built using three distinct methodologies: Random Forest (RF), Partial Least Squares Regression (PLSR), Convolutional Neural Network (CNN). Spatial distribution maps then generated each model. effectiveness proposed is confirmed utilization 240 surface samples gathered an arid region northwest China, specifically Xinjiang, characterized by sparse vegetation. Among all indices, TDTeI1 has highest correlation with SSC (|r| = 0.86). After adding multidimensional information, R2 RF model increased 0.76 0.90, improvement 18%. models, outperforms PLSR CNN. model, which (SOTT), achieves RMSE 5.13 g kg−1, RPD 3.12. contributes 44.8% prediction, contributions TD THD indices 19.3% 20.2%, respectively. confirms great potential introducing semi-arid regions.

Language: Английский

Citations

0