IDENTIFICAÇÃO DE EMBARCAÇÕES EM IMAGENS AEROTRANSPORTADAS DE RADAR DE ABERTURA SINTÉTICA (R-99 SAR) NA ÁREA MARÍTIMA DO BRASIL
DOI:
https://doi.org/10.5380/bcg.v17i3.24601Keywords:
Sensoriamento Remoto, Caracterização de Embarcações, Processamento de Imagens SAR, Realce EspectralAbstract
O objetivo deste trabalho foi identificar embarcações em imagens de radar obtidas pela aeronave R-99 da Força Aérea Brasileira. Dados de amplitude, obtidas na banda L e nas polarizações HH, HV, VH e VV da região de Porto de Tubarão, ES, foram processados por meio de diferentes tipos de realces, filtros, classificadores e transformadores espectrais. As imagens com maior potencial para identificar embarcações foram ainda analisadas para diferenciar embarcações militares de mercantes, considerando-se os cinco elementos de interpretação (forma, tamanho, sombra, tonalidade e fatores associados, isto é, o contexto em que as embarcações
se encontram nas imagens) e as cinco fases de interpretação de imagens (detecção, reconhecimento, análise, dedução e classificação). A combinação de processamentos mais favoráveis foi o realce com contraste 50-200, seguido de filtro
abertura ou erosão e classificador SVM (Support Vector Machine) ou transformação SCI (Synthetic Color Image). Foi possível discriminar embarcações nas fases de detecção e reconhecimento, enquanto a diferenciação entre embarcações mercantes e militares foi obtida nas fases de análise e dedução. No nível de classificação, não foi possível definir o tipo de embarcação militar (e.g., fragata ou contratorpedeiro) ou o tipo de embarcação mercante (e.g., petroleiro ou graneleiro).
Downloads
Published
How to Cite
Issue
Section
License
Submission of an original manuscript to the Journal will be taken to mean that it represents original work not previously published, that is not being considered elsewhere for publication.
The BCG allows the author(s) to hold the copyright without restrictions and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
The BCG also allows the authors to retain publishing rights without restrictions.
