USE OF UAV PLATFORM AS AN AUTONOMOUS TOOL FOR ESTIMATING EXPANSION ON INVADED AGRICULTURAL LAND

Abstract

For a long time, in many countries, questions involving disputes about land ownership has generated demand for geoinformation and documentation. In most cases, access for researchers is restricted or humanely impossible by eminence of conflicts, even armed. In these cases, researchers use Remote Sensing and Photogrammetry to enable their studies. However, the dynamics of the phenomenon being studied often requires approaches that traditional techniques become unviable or unable to fulfil. This work shows the results of an approach that used a photogrammetric UAV platform to take pictures of an invaded rural area in Brazil and estimate its expansion over two years. From the taken images, mosaics were generated and then classified using Decision Tree to identify tents. Then it was developed a Matlab algorithm, to detect and quantify the tents on the classified Images. It was possible to infer that there was an expansion of 7.3% between the two analyzed dates and probably more than three thousand people occupied the invasion site.