Emerging Technologies for Soil Evaluation Using Spectrometry Sensing, Internet of Things and Machine Learning

Florin Nenciu, Mihai Gabriel Matache, Iuliana Gageanu, Ioan Catalin Persu, Florin Bogdan Marin, Iulian Florin Voicea

The transition from conventional laboratory-based soil analysis to real-time, data-driven evaluation has become essential for advancing precision agriculture and ensuring sustainable resource management. This review provides a comprehensive and structured synthesis of emerging technologies for soil evaluation, focusing on the integration of spectrometric sensing, Internet of Things (IoT) systems, and machine learning approaches. A systematic analysis of peer-reviewed studies published between 2012 and 2026 was conducted to assess the performance of these technologies in terms of accuracy, robustness, and scalability under variable environmental conditions. Spectrometry techniques, including visible–near-infrared and mid-infrared sensing, enable rapid and non-destructive estimation of soil chemical properties, while IoT-based sensor networks facilitate continuous in situ monitoring of key parameters such as moisture, pH, and nutrient content. Machine learning models further enhance soil assessment by enabling predictive analytics, data fusion, and high-resolution mapping. Despite their significant potential, challenges related to data quality, model transferability, sensor calibration, and implementation costs remain critical barriers to large-scale adoption. The review highlights the need for standardized evaluation frameworks, improved multimodal data integration, and increased focus on model interpretability and real-world applicability. Overall, the synergistic use of these technologies supports more efficient input management, reduces environmental impact, and contributes to the development of resilient and sustainable agricultural systems.