%0 Journal Article
%4 sid.inpe.br/plutao/2019/06.10.16.19.58
%2 sid.inpe.br/plutao/2019/06.10.16.19.59
%@doi 10.18605/2175-7275/cereus.v11n1p184-194
%@issn 2175-7275
%F lattes: 5142426481528206 3 RenatoGarciaBragaHideFrag:2019:OdViNa
%T Odometria Visual para a Navegação Autônoma de VANT
%D 2019
%9 journal article
%A Braga, José Renato Garcia,
%A Shiguemori, Elcio Hideiti,
%A Campos Velho, Haroldo Fraga de,
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@electronicmailaddress jgarciabraga@gmail.com
%@electronicmailaddress elcio@ieav.cta.br
%@electronicmailaddress haroldo.camposvelho@inpe.br
%B Revista Cereus
%V 11
%N 1
%P 184-194
%K Unmanned Aircraft Vehicle (UAV), Visual odometry, UAV autonomous navigation.
%X The use of Unmanned Aircraft Vehicle (UAV) has being grown with many applications such as: ecological monitoring, precision agriculture, search and rescue operations, and engineering projects. An important objective of cientific community is to perform the UAV autonomous navigation. There are several strategies to develop an autonomous flight system, including the use of an inertial sensor combined with GPS, computer vision and visual odometry. The latter scheme is the focus of this article. Visual Odometry is applied and tested on the UAV RMAX helicopter. In order to implement the OV positioning system, the SURF and RANSAC algorithm were used as descriptors of the points of interest and post-processing to remove the false points of correspondence, respectively. The visual odometry method presents a cumulative error, but in the test performed, the maximum positioning UAV error was below 20 meters, which is acceptable when compared with the GPS error.
%@language pt
%3 2697-8857-1-PB.pdf