Avaliação de dados extraídos de imagens de satélite para fins de delimitação de imóveis rurais

Abstract

This work was developed to use techniques of Remote Sensing and georeferencing of rural properties to delimit the area of a rural property already certified by INCRA. With the general objective of comparing the geodetic survey with GNSS positioning methods, and extraction of features by Remote Sensing, to define the perimeter of the property, in a rural area already georeferenced according to the standards of Federal Law 10.267, of 2001. And with the specific objective of comparing images of three different satellites and with different spatial resolution to know which satellite is the best resolution for this type of work. The area selected for the study is located in the municipality of Itaqui, RS, near the mouth of the Itú river, when it meets the Ibicuí river. The georeferenced polygon of the property was accessed through the INCRA land assets, and its survey was carried out by an accredited professional considering the indications of the technical standard for the georeferencing of rural properties. The property used as a laboratory for the present study has part of its limit defined with fence and another limit with river. We used the images of the RapidEye satellite with 5 meters of spatial resolution, the satellite CBERS-4 PAN camera with 10 meters of spatial resolution, the satellite Landsat 8 OLI instrument with 30 meters of spatial resolution. For the necessary processing, the following programs were used: QGIS 2.14.4, open source GIS system, TerrSet 18.21 and AutoCAD. The vectorizations performed in satellites increased the difficulty according to the spatial resolution, the higher the resolution the better the vectorization. The difference in area calculated from the hydrographical line drawn from each satellite image compared to the area already georeferenced and certified by INCRA was -4.72 hectares of RapidEye, 3.66 hectares of CBERS-4 and 7.23 hectares of Landsat 8. The image of the RapidEye satellite was the best to do the vectorization, it was an image that did not suffer as much with the sinuosity of the river, while with the images of the satellite CBERS-4 and Landsat 8 obtained alteration in the vectorizations. This satellite was the best resolution to do the vectorization, although the result was not the closest to the one certified by INCRA, it is the one that presents the best geometric characteristics of the vectorized features.