JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Estimating soil profile salinity under vegetation cover based on UAV multi-source remote sensing

  • 0College of Hydraulic Science and Engineering, Yangzhou University, Yangzhou, 225009, China.
Scientific Reports +

|

January 21, 2025

|

January 21, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed machine learning models using multi-source remote sensing to estimate soil salinity at various depths under barley. Gaussian Process Regression and Random Forest models showed high accuracy, offering a new method for soil salinization monitoring.

Area Of Science

  • Environmental Science
  • Remote Sensing
  • Agricultural Science

Background

  • Soil salinization is a major land degradation issue in arid, semi-arid, and coastal China.
  • Estimating soil salinity at different depths under vegetation cover remains challenging.
  • This research addresses the need for effective soil salinity monitoring techniques.

Purpose Of The Study

  • To develop and evaluate machine learning models for estimating soil salinity at various depths.
  • To assess the performance of different algorithms and variable combinations.
  • To investigate the influence of crop coverage on salinity estimation.

Main Methods

  • Field-controlled experiments with multi-source remote sensing data collection.
  • Derivation and filtering of feature variables using the Boosting Decision Tree (BDT) method.
  • Application of four machine learning algorithms (including Gaussian Process Regression and Random Forest) with seven variable combinations.

Main Results

  • Gaussian Process Regression (GPR) models achieved high accuracy (R² up to 0.774) for 0-10 cm and 30-40 cm depths.
  • Random Forest (RF) models showed superior performance (R² up to 0.714) for 10-20 cm and 20-30 cm depths.
  • The study confirmed the effectiveness of machine learning and remote sensing for quantitative soil salinity estimation.

Conclusions

  • Machine learning models coupled with multi-source remote sensing data provide a robust approach for estimating soil salinity at various depths.
  • This method offers a valuable tool for monitoring soil salinization, especially under vegetation cover.
  • The findings support improved land management and agricultural practices in affected regions.

Keywords: