Evaluation of Machine Learning Algorithms for Estimating NDVI in Time Series of Multispectral Images for Precision Agriculture

Accurate estimation of the Normalized Difference Vegetation Index (NDVI) is crucial for precision agriculture and environmental monitoring. This study compares five machine learning algorithms LSTM, CNN-LSTM, XGBoost, Random Forest, and Gradient Boosting to predict NDVI using time series of multispectral Sentinel-2 data in a maize crop in Guasave, Sinaloa, Mexico. After data preprocessing, the models were evaluated using cross-validation and metrics such as MSE, MAE, RMSE, MAPE, and R2. The results showed that the LSTM model achieved the best performance in accuracy, while tree-based models struggled to predict extreme values. Recurrent neural networks demonstrated a greater capacity to capture complex temporal dependencies, although they still require improvements to optimize their robustness and precision. Smart citations: https://scite.ai/reports/10.61467/2007.1558.2026.v17i2.597Dimensions.Open Alex.