Comparação de índices de vegetação no mapeamento da cobertura da terra no semiárido: estudo de caso no Município de Martins/RN

Abstract

Studies related to sustainable development and proper planning of the use of natural resources is one of the challenges of today's society in the search for instrumental support for the characterization of environmental standards as, for example, the survey of land cover. With the advent of geotechnology, studies about the land cover in the semiarid region from vegetation indices are paramount in the study of natural resources, making it possible to distinguish different types of coverage and land use. Thus it is intended with this work compare the NDVI and SAVI, classification and land cover spatial distribution in Martins / RN. Therefore, satellite images were selected Landsat 8 and a map of land use and land cover developed by INPE. These images have gone through stages of preprocessing, where applied radiometric and geometric corrections, and then applied the NDVI and SAVI index in ERDAS software 9.2, with quantification of the land cover classes and preparation of layout in ArcGIS 10.2 software. In the field, they were observed and collected the coordinates of 125 control points, and then confusion matrix was designed to evaluate the accuracy of the maps from the Global Accuracy index and the Kappa index for both maps (NDVI, SAVI, INPE ). Initially, before the satellite images treatment were observadadas field eight land cover classes in the city of Mantins / RN, as follows: Seasonal Forest, Savannah-Estépica forested, wooded Savannah-Estépica, permanent and Agriculture Temporary Soil exposed , urban area and water body. Given the confusion matrix, prepared from 125 control points obtained in the field, with a Kappa index of 66.96%, the SAVI index showed better results compared with NDVI to map Martins municipal land cover / RN. In assessing the accuracy of the land cover mapping from an array of confusion, the indexes to evaluate the accuracy of accuracy of mappings (global and Kappa accuracy Index) showed to be great options with regard to the analysis of vericidade such data, thus obtaining better results for SAVI index. Thus, it is concluded that the use of satellite from remote sensing images in the application of vegetation indices, has proved to be relevant tools in the study of land cover, along with the Global accuracy rates and the Kappa index, which in turn, proved to be relevant alternatives regarding the accuracy of land cover maps.