Scholarlink

<p>Precision agriculture, or agriculture 4.0, leverages modern technologies such as the IoT, cloud services, and AI to gather precise data on crop growth, soil moisture, and temperature. This data assists farmers in making efficient management decisions crucial for sustainable water resource use. Predicting plant irrigation needs is essential, particularly given the ongoing decline in freshwater reserves, largely attributable to agricultural consumption. This paper proposes a novel CNN-MLP hybrid approach for precise daily irrigation water amount prediction. The CNN model aims to standardize time intervals by reducing the dimensionality of the per-minute parameter data records and extracting relevant information without losing pertinent details. The MLP model uses this data to predict the daily plant irrigation needs. Results show the MLP model, utilizing CNN-reduced data, achieved a small error margin (Mean Absolute Error (MAE) = 0.5&#xa0;l/m<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(^2\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mrow /> <mn>2</mn> </mmultiscripts> </math></EquationSource> </InlineEquation> and <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(R^2\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>R</mi> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> score of 0.82). Comparing our solution with various models, techniques, and data harmonization methods, our proposed approach outperformed the baseline models and methods across evaluation metrics for prediction, data harmonization, and feature reduction, highlighting its efficacy in optimizing water resource management for sustainable and precision agriculture.</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

A Novel Deep Learning Framework Based on Heterogeneous Temporal Data Harmonization for Irrigation Water Amount Prediction

  • 期刊论文

  • 发表日期：

    2026年4月5日

查看全文
  • Hamed Laouz,
  • Soheyb Ayad,
  • Labib Sadek Terrissa,
  • Samir Merdaci,
  • Noureddine Zerhouni

摘要

Precision agriculture, or agriculture 4.0, leverages modern technologies such as the IoT, cloud services, and AI to gather precise data on crop growth, soil moisture, and temperature. This data assists farmers in making efficient management decisions crucial for sustainable water resource use. Predicting plant irrigation needs is essential, particularly given the ongoing decline in freshwater reserves, largely attributable to agricultural consumption. This paper proposes a novel CNN-MLP hybrid approach for precise daily irrigation water amount prediction. The CNN model aims to standardize time intervals by reducing the dimensionality of the per-minute parameter data records and extracting relevant information without losing pertinent details. The MLP model uses this data to predict the daily plant irrigation needs. Results show the MLP model, utilizing CNN-reduced data, achieved a small error margin (Mean Absolute Error (MAE) = 0.5 l/m \(^2\) 2 and \(R^2\) R 2 score of 0.82). Comparing our solution with various models, techniques, and data harmonization methods, our proposed approach outperformed the baseline models and methods across evaluation metrics for prediction, data harmonization, and feature reduction, highlighting its efficacy in optimizing water resource management for sustainable and precision agriculture.