Artificial intelligence for sustainable soil systems: Advances in prediction, mapping and sustainable management of soil health and properties: A critical review
持続可能な土壌システムのための人工知能：土壌の健康と特性の予測、マッピング、持続可能な管理の進歩：批判的レビュー (AI 翻訳)
Venkatesan VG, Sankar R, Muhilan G, Gokulakrishnan B, Naveen N, Kavin P, Sivaganeshan K, Rajan Babu B
🤖 gxceed AI 要約
日本語
本レビューは、AI（機械学習・深層学習）を用いた土壌特性の予測・デジタルマッピング・精密管理の進展を体系的にまとめ、ランダムフォレストなどが高い精度を示す一方、解釈性やデータ品質、途上国での導入障壁を指摘。説明可能AIやフェデレーテッドラーニングなどの新技術が今後の鍵となる。
English
This critical review synthesizes a decade of AI (ML/DL) applications in soil science for predicting properties, digital mapping, and precision management. Random Forest achieves 90-95% accuracy for soil organic carbon, but challenges remain in interpretability, data quality, and scalability in smallholder systems. Emerging solutions like explainable AI and federated learning promise more accessible and trustworthy tools.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本では、農地の炭素貯留能評価や精密農業へのAI活用が進む中、本レビューは土壌健康管理の手法比較と導入課題を整理しており、国内のスマート農業やカーボンファーミング政策に関連する。
In the global GX context
Globally, soil health is critical for climate regulation and food security. This review evaluates AI methods for soil property prediction and management, informing sustainable agriculture practices and carbon sequestration strategies relevant to global climate goals.
👥 読者別の含意
🔬研究者:A comprehensive review of AI applications in soil science, highlighting strengths and limitations of various algorithms.
🏢実務担当者:Provides insights into AI tools for precision soil management and potential cost savings in agricultural operations.
📄 Abstract（原文）
Soil is fundamental to food security, biodiversity and climate regulation, yet faces accelerating degradation from intensive agriculture, climate change and land scarcity. Artificial Intelligence (AI), encompassing Machine Learning (ML) and Deep Learning (DL), offers transformative capabilities for handling heterogeneous, high-dimensional soil data, enabling accurate prediction of properties (e.g., texture, organic carbon, moisture, nutrients), digital mapping, real-time monitoring via sensors/IoT, and precision management. This critical review synthesizes developments over the past decade, based on a systematic search across major databases (Google Scholar, Scopus, Web of Science, IEEE Xplore, ScienceDirect, MDPI, Nature). We evaluate dominant algorithms, Random Forest (RF), Support Vector Machines (SVM), Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), highlighting their ability to capture non-linear patterns while critiquing limitations in interpretability, data quality, regional bias and scalability for smallholder systems. AI consistently outperforms traditional methods in predictive accuracy (e.g., RF achieving 90-95% predictive accuracy for soil organic carbon across multiple studies), yet barriers persist: high costs, limited farmer training, infrastructure gaps in developing regions, and model opacity. Emerging solutions such as explainable AI, federated learning, digital twins, hybrid physics-informed models and edge computing promise greater accessibility and trustworthiness. We outline priorities for interdisciplinary collaboration to advance equitable, sustainable soil management aligned with global goals for food security and climate resilience.
🔗 Provenance — このレコードを発見したソース
- openalex https://doi.org/10.33545/2618060x.2026.v9.i6se.5867first seen 2026-07-20 04:47:06
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。