%0 Conference Proceedings
%4 sid.inpe.br/marte2/2017/10.23.19.37.26
%2 sid.inpe.br/marte2/2017/10.23.19.37.27
%@isbn 978-85-17-00088-1
%F 59238
%T Classificação supervisionada entre soja Bt e soja não-Bt, em imagem RGB gerada por drone, a partir da ferramenta Pixel Explorer
%D 2017
%A Marinho, Carlos Alberto Branco,
%A Luz, Priscila Maria Colombo da,
%A Baptista, Gustavo Macedo de Mello,
%A Specht, Alexandre,
%@electronicmailaddress carlosabmarinho@bol.com.br
%E Gherardi, Douglas Francisco Marcolino,
%E Aragão, Luiz Eduardo Oliveira e Cruz de,
%B Simpósio Brasileiro de Sensoriamento Remoto, 18 (SBSR)
%C Santos
%8 28-31 maio 2017
%I Instituto Nacional de Pesquisas Espaciais (INPE)
%J São José dos Campos
%P 1455-1461
%S Anais
%1 Instituto Nacional de Pesquisas Espaciais (INPE)
%X The image classification is an important tool used by remote sensing professionals, but the classification of features that have very similar features is a work of extreme difficulty, since the bands of electromagnetic radiation in the portion of the visible many confuse the analyst and commercial software of classification and, thus, similar features are commonly classified as equals. This work aims to demonstrate that it is possible to perform a supervised classification from images obtained by remote sensors, even those coming from sensors with low spectral resolution, as in the case of recreational UAVs, and with it distinguish not only two different kinds of coverage vegetable, but differentiate two variations of the same plant species, as is the case of Bt-soybean evaluated in relation to non-Bt-soybean. Using the computational tool named Pixel Explorer (PE), developed in Matlab by the first author of this work, as dissertation composition and later thesis, a classification was made in an experimental area of EMBRAPA, resulting in the separation of the parcels containing two kinds of genetically different soybean, being classified material composed of images collected by a drone model: Phanton 3 Professional, with spectral resolution restricted to bands RGB, with oblique view and without gyro stabilization, leading to the hypothesis that the result can be even more reliable if the same methodology is used in images generated by sensors with high spatial and spectral resolutions and target nadir for both vegetation and geology.
%9 Processamento de imagens
%@language pt
%3 59238.pdf