Year: 2005
Type: article
Abstract: In this study we propose a methodology to classify and map logging roads in the Amazon region. Visual interpretation followed by digitization of roads was compared with methods for automatic extraction of linear features (i.e., image slicing, thinning and gap filling). Digitization of roads produced more accurate results than the automatic methods. The extension of logging roads, mapped with visual interpretation, suppressed the extension of the main roads and of the access logging roads by three fold. In addition, the road pattern formed by the logging roads showed that it is possible to separate managed logged forests from unplanned logging. Even though the digitization is a time-consuming process, the proposed method to interpret and classify logging roads has the potential to be automated.
Authors: Brandão Júnior, Amintas de Oliveira, Souza Júnior, Carlos Moreira de
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6999012903