The characterization of uses and coverage of the land is currently an important element for the management and monitoring of agricultural territories and natural resources with the aim of facilitating the decision-making of governmental and non-governmental entities. Given this, the objective of the present investigation was to generate information based and updated on the main agricultural production systems through teledetection techniques within the study area and with it, to elaborate a methodological process that implements technological tools that automate the generation of geographic information through spatial observation of the earth. Para el processing de los datos se adaptaron codes de lenguajes computacionales dentro de la plataforma de Google Earth Engine (GEE) y comos insumos base se utilizaron imágenes Sentinel-2 para el año 2021 y los datos obtenidos en la gira campo realized in ese same year. The mosaic of images was classified with three classification algorithms and validation was applied to each of them, giving the result that the algorithm that presented the highest global accuracy was “Ramdom Forest” with 82.95%, followed by “Classification and Regression Tree (CART)” with 81.50% and lastly “Minimum Distance” with a total of 57.51%. Based on the results obtained by "Ramdom Forest", it was determined that the main land uses with major occupation were forest cover with a total of 136,162, 78 ha, followed by the classes of pastures with 25,497,24 ha, sugarcane cultivation with a total of 11,081,02 ha and coffee cultivation with 5,512,17 ha. Por último, se elaboraron dos mapas de aptitud agrículo para los cultivos de caña y coffee para determinar las zonas potenciales y no potenciales para el desarrollo de estos cultivos. It is hoped that this investigation will encourage the application of these tools for the management of territories at the national level and benefit the agricultural sector.
Palabras clave: usos y coberturas del suelo, algoritomos clasificadores, clasificación supervisada, sentinel-2, google earth engine, aptitude agrícola.