%0 Report
%4 sid.inpe.br/mtc-m21b/2017/06.30.16.59
%2 sid.inpe.br/mtc-m21b/2017/06.30.16.59.25
%@ INPE-1206-PE-117
%T A bayesian approach to edge detection in images
%D 1978
%9 RPQ
%P 30
%A Mascarenhas, N. D. A.,
%A Prado, L. O. C.,
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%I Instituto Nacional de Pesquisas Espaciais
%C São José dos Campos
%K image processing, edge detection, statistical decision theory.
%X New statistical techniques for the edge detection problem problem in images are developed. The image is modeled by signal and noice which are independent, additive, Gaussian and autorregressive in two dimensions. The optimal solution, in terms of statistical decision theory, leads to a test which decides among multiple, composite, overlapping hypotheses. A computationaVly attractive suboptimal test, involving non-overlapping hypotheses, is proposed. Results are presented with simulated data and real satellite images. A comparison with standard gradient techniques is made.
%@language en
%3 publicacao.pdf