Survey Reviews AI Models for Soil Moisture Estimation and Classification.

Modeling soil moisture is a challenging task due to its complex spatiotemporal nature, involving non-linear environmental interactions, diverse data sources, and often limited ground observations. Traditional physics-based models, while accurate, are computationally intensive and struggle with scalability for large-scale deployments. Artificial intelligence methods offer a flexible alternative, capable of identifying empirical relationships between soil moisture and various environmental factors with fewer underlying assumptions. This survey provides a structured overview of these AI-based approaches for both estimating and classifying soil moisture levels. The paper organizes current methods into five main categories: statistical time-series models, geostatistical techniques, classical machine learning algorithms, deep learning models, and probabilistic or Bayesian methods. These models typically utilize historical soil moisture data, meteorological information, vegetation indices, topographical features, soil characteristics, and geolocation data to perform their respective regression or classification tasks.