Identification of phenological stages of sugarcane cultivation using Sentinel-2 images

• Published in: 2020 9th International Conference On Software Process Improvement (CIMPS) pp. 110 - 116
• Main Authors: Cruz-Sanabria, Humberto, Sanchez, Maria Guadalupe, Rivera-Caicedo, Juan Pablo, Avila-George, Himer
• Format: Conference Proceeding
• Language: English; Spanish
• Published: IEEE 21.10.2020
• Subjects: Agriculture; Phenological phases; Random forests; Satellites; Sentinel-2; Software; Spectral vegetation index; Sugarcane; Supervised classification; Support vector machines; Time series analysis; Vegetation mapping
• DOI: 10.1109/CIMPS52057.2020.9390095

Sugarcane is a crop of great commercial and economic importance in over 130 countries, including Mexico. One of the main problems faced by sugarcane producers is how to improve the yield of the crop. In order to develop new techniques that can improve the yield of the sugarcane crop, it is key to be able to identify the critical stages of growth and to be able to make timely decisions. In this paper, a method is presented to identify the phenological stages of sugarcane crops using data from the MultiSpectral Instrument sensor onboard the Sentinel-2 satellite. For the development of the proposed method, classification some methods were evaluated: k-Nearest Neighbors, Random Forest, Support Vector Machine, and Naïve Bayes. Also, time series of five vegetation indices were used as input data; the results were validated using the cross-validation technique with k = 10 iterations. The results show that the method Random Forest achieves an accuracy = 92.45 %, which is the best suited to identify the growth stages of sugarcane crops.