IMPROVEMENTS OF THE DIVIDE AND SEGMENT METHOD FOR PARALLEL IMAGE SEGMENTATION

Conteúdo do artigo principal

Anderson Reis Soares
Thales Sehn Körting
Leila Maria Garcia Fonseca

Resumo

Remote Sensing is an important source of data about the dynamics of Earth's land and oceans, but retrieve information from this technique, is a challenge. Segmentation is a traditional method in remote sensing, which have a high computational cost. An alternative to suppress this problem is use parallel approaches, which split the image into tiles, and segment each one individually. However, the divisions among tiles are not natural, which create inconsistent objects. In this work, we extended our previous work, which used non-crisp borders computed based on graph-theory. By applying this non-crisp line cut, we avoid the post-processing of neighboring regions, and therefore speed up the segmentation.

Downloads

Não há dados estatísticos.

Métricas

Carregando Métricas ...

Detalhes do artigo

Como Citar
SOARES, A. R.; KÖRTING, T. S.; FONSECA, L. M. G. IMPROVEMENTS OF THE DIVIDE AND SEGMENT METHOD FOR PARALLEL IMAGE SEGMENTATION. Revista Brasileira de Cartografia, [S. l.], v. 68, n. 6, 2018. DOI: 10.14393/rbcv68n6-44486. Disponível em: https://seer.ufu.br/index.php/revistabrasileiracartografia/article/view/44486. Acesso em: 16 abr. 2024.
Seção
Artigos

Artigos mais lidos pelo mesmo(s) autor(es)

1 2 > >>