%0 Conference Proceedings
%4 dpi.inpe.br/marte/2011/07.21.13.11
%2 dpi.inpe.br/marte/2011/07.21.13.11.47
%@isbn 978-85-17-00056-0 (Internet)
%@isbn 978-85-17-00057-7 (DVD)
%T Registro Semi-automático de Imagens CBERS utilizando Coeficiente de Correlação de Pearson
%D 2011
%A Alves, Andrei Olak,
%A Meloni, Raphael Belo da Silva,
%A Brito, Jorge Luís Nunes e Silva,
%@affiliation K2 Sistemas e Projetos Ltda
%@affiliation K2 Sistemas e Projetos Ltda
%@affiliation Universidade do Estado do Rio de Janeiro - UERJ
%@electronicmailaddress andreiolak@k2sistemas.com.br
%@electronicmailaddress rmeloni@k2sistemas.com.br
%@electronicmailaddress jsilvabr@gmail.com
%E Epiphanio, José Carlos Neves,
%E Galvão, Lênio Soares,
%B Simpósio Brasileiro de Sensoriamento Remoto, 15 (SBSR).
%C Curitiba
%8 30 abr. - 5 maio 2011
%I Instituto Nacional de Pesquisas Espaciais (INPE)
%J São José dos Campos
%P 2546-2552
%S Anais
%1 Instituto Nacional de Pesquisas Espaciais (INPE)
%K Remote Sensing, CBERS imagery, semi-automatic image registration, Pearson Correlation Coefficient, Sensoriamento Remoto, imagens CBERS, Registro semi-automático de imagens, Coeficiente de Correlação de Pearson.
%X In recent years, there was a considerable increasing in researche that uses information derived from free remote sensing imagery . For example, speaking only from the series of CBERS satellites, more than 300,000 images were freely downloaded from the Internet. However, one major problem is that the CBERS imagery is available only with a partial geometric correction. Such a partial correctionis characterized by a displacement that can range from hundreds of meters to some kilometers, which creates the needing that those images must be properly registered before being used. Due to the fact thatmanually image registration is a timeconsuming, tedious and error-prone task, it is necessary to search for new possibilities of automation of image registration process. Thus, the purposal of this study is to use the Pearson correlation coefficient to identify homologous features in two images, in order to automatically gathering the coordinates of control points required for image registration. For accomplishment of this goal we developed an algorithm in C + +. This algorithm allows to correlate homologous features of two images and delivers the values of coordinates of the homologous points, that are then saved in a file that can be used inthe software SPRING. The results show the quality of the correlation coefficients for homologous features, the small errors in the control points, and the visual quality of CBERS´s image registration.
%9 Processamento de Imagens
%@language pt
%3 p0563.pdf