Avaliação da qualidade posicional de pontos de controle para geração de modelos digitais agrícolas.

Abstract

Remote sensing is a way of obtaining information that allows a study about an object, without having to come into direct contact with it, generating a model of the content to be studied. The proximity of the model obtained from this technique to reality makes the analysis and decisionmaking more precise, corresponding to real needs. The present work seeks to base itself on this information in order to assess the positional quality of models generated from aerial remote sensing, comparing the coordinates obtained by GNSS to those found in the orthomosaics generated from the composition of the flight images. Coordinates were obtained through static survey of 81 control points in and around mega plots and static survey of the base. This experiment used a total of 22 control points to generate orthomosaic and digital terrain models. These models were created by placing the points in different number and spatial distribution at each processing, to assess the importance of this distribution being performed correctly. The best results were found in the models that used the base to create the models, as it stayed longer at the coordinates collection point, bringing a more precise response in relation to other points. The best distribution of control points, from the analysis of results, was the orthomosaic with 22 points.