@InProceedings{KuxSanParAhePie:1997:CaRADa,
author = "Kux, Hermann Johann heirinch and Santos, Joao Roberto dos and
Pardi Lacruz, Maria Silvia and Ahern, Frank J and Pietsch, Ron W",
affiliation = "Instituto Nacional de Pesquisas Espaciais, Divis{\~a}o de
Sensoriamento Remoto (INPE, DSR)",
title = "Capabilities of RADARSAT data for land use land cover analysis in
SWmazonia (test site Acre, Brazil)",
year = "1997",
pages = "35--43",
organization = "Geomatics in the Era od RADARSAT International Conference.",
keywords = "VEGETACAO, Amaz{\^o}nia (Regiao), Acre (Estado), Radarsat,
florestas, desmatamento, desflorestamento, mapeador tem{\'a}tico
(Landsat, miss{\~a}o Sarex, imagens de radar, sar, radar de
abertura sint{\'e}tica, monitoramento, fotografia a{\'e}rea.",
abstract = "The area under study, located in SW Amazonia, Acre State, Brazil,
has been studied in the last few years, using both optical
(TM-Landsat)and microwave (airborne C-band SAR, ERS-1,
SIR-C)remote sensing data. Recently data from RADARSAT in both
Standard and Fine mode beams were acquired. The objective of this
study is to show the capabilities of RADARSAT and merged
RADARSAT/Landsat TM images for land use/land cover discrimination
in SW Amazonia. The concept of the study is to analyze the
feasibility for systematic mapping of similar land use / land
cover features in the tropical rainforest areas. Two distinct
periods of the tropical hydrological cycle, i.e. at the end of
both the rainy (April/May)and at the dry season
(September/October)were investigated. The main features
discriminated were: phenological conditions of pasture, recent
forest burning, and changes in regrowth patterns. The separation
of small (50-100 ha)and large (200 to several 1000 ha)clearing is
also feasible. Furthermore, a stratification of primary forest and
regrowth is also in progress. The RADARSAT products used were
pre-processed at CCRS. It consisted of scaling the data from 16 to
8 bit real integers, antenna pattern correction, attenuation of
speckle and linearly contrast stretch. The RADARSAT data was then
registered to Landsat TM Bands3, 4 and 5 with IHS/RGB transforms,
to integrate the data. This study is supported by both color
aerial photographs (scale 1:20.000)and physiognomic-structural
profiles of the vegetation types.",
conference-location = "Ottawa, CA",
conference-year = "May 25-30 1997",
copyholder = "SID/SCD",
label = "8328",
organisation = "Canadian International Development Angency",
targetfile = "INPE 7015.pdf",
urlaccessdate = "16 jun. 2024"
}