Journals Proceedings

Abstract

The Amazon is one of the most complex and diverse biomes in the world and their level of environmental preservation has global scope of impact. However, apart from the environmental aspects, the complexity of the region involves other different views such as: social, economic and cultural. Indeed, such views are inextricably linked, so that, for example, cultural aspects can affect the relationship between the land use and land cover and, in the case of Amazon, depending on the scale of these changes, there may be consequences planetary. Depending on the criticality involving the region, various government actions, organizations and social movements and the international community, has been trying to rationalize the land use and land cover, in order to create a sustainable relationship of exploitation of natural resources. An important government program is based on the creation of settlements, with their support, with regard to certain infrastructure financing, machinery, seeds and seedlings, technical assistance, among others. However, despite the work involved, the solution is extremely complex, given the correlation of factors to be assessed and combined in search of success and improvement of such programs. So on behalf of all complexity involved, is of paramount importance the use of technological tools to measure the degree of evolution of theland use and land cover in the regions that are part of the Amazon forest. One of the fundamental assessments is the observation of the evolution of land use change, because it is a temporally changing variable, with severe implications on the natural landscape and the environment as a whole.This study aims to evaluate the performance of computational techniques SVM, Neural Networks, Maximum Likelihood and Decision Tree to better classify the images of Landsat 5 satellite on the scene of the Brazilian Amazon. Data mining techniques are being used increasingly used for satellite image classification purposes, among various techniques there is the decision tree. This technique improved the classification accuracy and very efficient computational way, as the results were generated automatically and optimized time. The study area, concentrated in the settlement 26 de Março / Maraba, are facing serious problems in its reforestation. This settlement is only a reflection of occurring in the Amazon biome and its complex and diverse environmental system