Scholarlink

<p>Soil pH and Electrical Conductivity (EC) are critical indicators of soil health, influencing agricultural productivity and environmental sustainability. This study employs Machine Learning (ML) models to generate spatial prediction maps for soil pH and EC using the laboratory analyzed soil samples, Landsat 8 bands, derived spectral indices in the Google Earth Engine (GEE) platform. Five regression models namely Random Forest (RF), Classification and Regression Tree (CART), Gradient Boosting Regression (GBR), Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) are applied to analyze spatial variations in soil properties. This study integrates field-collected soil parameters with remotely sensed data to enhance predictive accuracy and enable large-scale soil monitoring and generates digital maps for Indukurpet madal, located in Sri Potti Sriramulu (SPSR) Nellore district of Andhra Pradesh. The performances of the models are evaluated using coefficient of determination (R<sup>2</sup>), Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) for both training and testing datasets. The RF model consistently outperformed the other methods achieving the highest prediction accuracy for EC. GBR delivered the the highest prediction accuracy for pH. For EC, the RF model attained an RMSE of 0.05, R² value of 0.86, and MAE of 0.03. while for pH, GBR achieved an RMSE of 0.14, R² value of 0.769 and MAE of 0.09. This demonstrates the robustness of the models in capturing the intricate spatial diversity of soil properties.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine learning-based prediction of soil pH and EC using landsat 8 images

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月15日

查看全文
  • Suneetha C.,
  • Lakshmi Sutha Kumar,
  • Sreenivas K.,
  • Tarik Mitran,
  • Venkatachalam. K

摘要

Soil pH and Electrical Conductivity (EC) are critical indicators of soil health, influencing agricultural productivity and environmental sustainability. This study employs Machine Learning (ML) models to generate spatial prediction maps for soil pH and EC using the laboratory analyzed soil samples, Landsat 8 bands, derived spectral indices in the Google Earth Engine (GEE) platform. Five regression models namely Random Forest (RF), Classification and Regression Tree (CART), Gradient Boosting Regression (GBR), Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) are applied to analyze spatial variations in soil properties. This study integrates field-collected soil parameters with remotely sensed data to enhance predictive accuracy and enable large-scale soil monitoring and generates digital maps for Indukurpet madal, located in Sri Potti Sriramulu (SPSR) Nellore district of Andhra Pradesh. The performances of the models are evaluated using coefficient of determination (R2), Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) for both training and testing datasets. The RF model consistently outperformed the other methods achieving the highest prediction accuracy for EC. GBR delivered the the highest prediction accuracy for pH. For EC, the RF model attained an RMSE of 0.05, R² value of 0.86, and MAE of 0.03. while for pH, GBR achieved an RMSE of 0.14, R² value of 0.769 and MAE of 0.09. This demonstrates the robustness of the models in capturing the intricate spatial diversity of soil properties.