Resumen de Datos de teledetección que atestiguan la ocupación para la regularización de la tenencia de la tierra en Brasil
B. de A. M. Trovão, André Brunckhorst, J.D. Teixeira, C.E.P. Sturm
The implementation of Brazil’s Land tenure Regularization policy always has been a challenge due to the size of the country, social and economical differences. It is required by law, for the Land Tenure Regularization, to prove occupation prior to 2008 and to present the Environmental Rural Register – CAR. This one is mandatory for all rural properties and has the objective to receive the related environmental information for the Rural Environmental Regularization policy.
The year 2008 is important for both regularization process. For the CAR it is available a 2008 map of consolidated rural areas for the whole country as a reference, on the other hand, the Land Tenure Regularization does not have anything like that. The objective of the present paper is to assess the accuracy of the consolidated areas map to identify the occupation in 2008 for the Tenure Regularization Process.
The vector base for consolidated areas of 2008, from Brazilian Forest Service (SFB), here called (BR-CAR), is on a spatial scale of 1:50,000. This base was elaborated from the processing of RapidEye 2011 and Landsat 2008-2011 images with spatial resolutions of 5 and 30 m, classified by a methodology that combines machine learning algorithms (machine learning) and visual interpretation.
Since the (BR-CAR) was elaborated from the processing of medium and low- resolution images (RapidEye and Landsat), this paper aims to verify the agreement of this base with the visual interpretation of images with high spatial resolution (SPOT 2008).
The BVR-SFB showed better applicability for parcels with an area ≧ 10 hectares, since it reached a percentage agreement of 93.40%. For the size classes (0,1 - 2 hectares), (0,1 - 6.25 hectares) and (0,1 - 10 hectares) the percentage indices of "non-agreement" obtained were 22.13%, 13.44% and 12, 39% respectively, values that indicate a higher risk of misunderstanding. As for the Kappa indices, they showed similar behavior to the percentage indices. The best result (0.46) was obtained for the size class ≧ 10 hectares, considered moderate. For the other classes (0,1 - 2 hectares), (0,1 - 6.25 hectares) and (0,1 - 10 hectares) the values were lower, classified as reasonable. In view of these results, consideration should be given to the application of the Reference Base for Consolidated Areas (BR-CAR) in the analysis of occupation and practice of effective cultivation of properties with a size of less than 10 hectares.
