Proposal of a Machine Learning Approach for Early Anomaly Detection in Agricultural Crops Using Hyperspectral Data Conference Paper