Normalized Difference Vegetation Index – NDVI processed in the cloud with Google Earth Engine

Abstract

Introduction: Artificial Intelligence (AI) has revolutionized the way environmental studies are processed, allowing the production and analysis of large volumes of spatial data with precision and greater agility. The Normalized Difference Vegetation Index (NDVI) on the Google Earth Engine (GEE) platform is possible due to the integration of several databases such as the United States Geological Survey-USGS in the USA, which stores satellite images, and in Brazil the IBGE, which provides download of vector files, among others, that corroborate the assessment of changes in land cover. Cities in the interior of Rondônia located on the BR 364 axis, such as Ariquemes, with agricultural activity, require environmental monitoring for the sustainable management of natural resources. Objectives: This research proposed to develop a continuous time series of Normalized Difference Vegetation Index with Landsat satellite image processed with cloud-based artificial intelligence to analyze changes in vegetation from the 1980s to the present day and correlate with land use and occupation in the municipality of Ariquemes-RO. Methodology: The theoretical basis used a systematic review of publications in the SciELO, EBESCO and Periódico Capes databases. Geographic data processing was performed using cloud computing in GEE, a free version with JavaScript language. A script was structured and stored the geometry, variable and function codes used to create a mosaic, the NDVI band, graphs and merging of images from Collection 2 of Landsat 8 Sensor OLI/TIRS and Collection 2 of Landsat 5 Sensor TM, both with a spatial resolution of 30 m and corrected atmospheric surface reflectance. Results: A continuous time series was obtained, 1984-2024 for NDVI, which showed a decrease in vegetation after 2013, confirmed by a trend line for two areas near a fish tank in a rural area. Final considerations: Based on the temporal and spatial variations of NDVI using graphs and maps generated in GEE, it is possible to infer intensification of land use from 2014 onwards, considering that from 1984-2012 the NDVI reached a maximum between 0.8-0.9 and from 2012 to 2024 it fell with NIDVI below 0.5.