Classification and monitoring of vegetation through NOAA-AVHRR images

Abstract

Knowledge of phenological variations in vegetal cover is today an indispensable aspect for the planning of a coherent and efficient policy of sustainable development, as well as for the understanding and objective evaluation of the co-existence of different ecosystems whether natural, semi natural, agricultural or industrial. The analysis of these variations at regional level and the interconnectionwith the different components of the geographical environment (climate, soil, relief geology, etc.) are part of the necessary information for the understanding of the functioning of ecosystems on a global scale. The monitoring of these variations at regional and global level, today, is only possible due to the development attained in the last 15 years by techniques such as: remote sensing, geoprocessing, satellite global positioning systems, among others. The objective of this project is to undertake the classification and monitoring of the ecosystems of the Center-West region do Brazil (state of Mato Grosso) using a multitemporal series of images from the Advanced Very High Resolution Radiometer (AVHRR) sensor on board the National Oceanographic and Atmospheric Administration (NOAA-11) satellite in High Resolution Picture Transmission (HRPT) format. Notable among the principal aspects of the methodology is the generation and analysis of the fraction images (vegetation, soil and shade) derived from the linear model of spectral mixture applied to the AVHRR images. Among the principal results that should be obtained are the classification of the vegetation of the area study and its seasonal variations, based on the analysis of the Normalized Difference Vegetation Index (NDVI) and of the fraction images (vegetation, soil and shade) derived from the AVHRR images for the period of September 1992 to August 1994. (AU)