Exploring the Correlation Between Gaseous Emissions and Phenological Phases in Tomato Crops Through Machine Learning

Nowadays, agriculture is facing significant challenges, including climate change. Precision agriculture might address these issues by optimizing resource use and promoting sustainability. In this work, a case study of tomato crop monitoring is presented, employing a large amount of gas sensor data collected over three years (2020–2022) to develop models for phenological phase classification. A k-NN classifier achieved accuracies above 99% across multiple train/test splits, with AUC, sensitivity, specificity, precision, and F1-score above 98%. Results demonstrate the feasibility of low-computational-cost systems capable of real-time detection of the transition point between plants’ developmental stages.