%0 Conference Proceedings
%4 sid.inpe.br/mtc-m17@80/2007/12.17.15.19
%2 sid.inpe.br/mtc-m17@80/2007/12.17.15.19.53
%T Change Detection in the Amazon Rainforest with Radiometric Rotation Technique RCEN Multi-spectral Case Study: Guarayos – Bolivia
%D 2007
%A Ugarte, H. Ferrufino,
%A Niedzwiecki, T. Zawila,
%A Santos, João Roberto dos,
%A Maldonado, F. D.,
%@affiliation University of Applied Sciences Fh – Eberswalde – Forestry Faculty
%@affiliation University of Applied Sciences Fh – Eberswalde – Forestry Faculty
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Instituto Nacional de Pesquisas da Amazônia - INPA
%@electronicmailaddress hferrufino@fh-eberswalde.de
%@electronicmailaddress tzawila@fh-eberswalde.de
%@electronicmailaddress jroberto@ltid.inpe.br
%@electronicmailaddress maldonado@inpa.gov.br
%B IEEE Insternational Geoscience and Remote Sensing Symposium.
%C Barcelona, Espanha
%P 4
%K change detection, degradation, CBERS-2, tropical forest, remote sensing, Guarayos.
%X A working group of three institutions was set up to develop this study: University of Applied Sciences Eberswalde (Germany), National Institute for Space Research (INPE, Brazil) and National Institute for Amazon Research (INPA, Brazil). The main task is to apply in the Guarayos region (Bolivia), the multitemporal change detection algorithm RCEN multi-spectral. The study area is located in Guarayos-Bolivia, characterized by two main high forests landscapes, The Amazon Region and The Brazilian Paranaense Region; the approach of the change detection is taken under multivariate analysis (three spectral bands), with data coming from two kinds of sensor ETM+/Landsat-7 and CCD/CBERS-2. The image detection was transformed from continuous image (floating-point) to thematic, through slicing and labeling process. Hence it is possible to discriminate five thematic classes: two related to degradation, two referring to regeneration and one of no-change. The change detection map shows: in the timeframe studied 11% of all area under study presents deforestation patterns, on the other hand the regeneration class is not significant. In conclusion the methodology has good performance and it is evolving in landscapes with high humidity complications.
%@language en
%3 joão roberto2.pdf