GEOSCAN Search Results: Fastlink

GEOSCAN Menu


TitleLand cover classification from RADARSAT data of the Tapajós National Forest, Brazil
 
AuthorShimabukuro, Y E; Amaral, S; Ahern, F; Pietsch, R W
SourceCanadian Journal of Remote Sensing 24, 4, 1998., https://doi.org/10.1080/07038992.1998.10874703
Year1998
Alt SeriesEarth Sciences Sector, Contribution Series 20042880
PublisherInforma UK Limited
Documentserial
Lang.English
Mediapaper; on-line; digital
File formatpdf
Released2014 07 31
Abstract(Summary)
The objective of this research was to analyse RADARSAT images for forest types and land cover classification. Image processing techniques used to enhance RADARSAT images and forest types discrimination are presented. An area including the Tapajós National Forest and its immediate surroundings, located in Pará State, Brazil, was chosen for this investigation. The area to the east of the national forest is comprised of numerous small agricultural plots and abandoned areas with early secondary successional forest while the national forest itself is mainly undisturbed primary forest. RADARSAT standard and fine mode images were used in this study. Generally, the RADARSAT standard mode images showed good association with an existing vegetation map as characterized primarily by topographic features. Also, these images show recent clear cut areas, especially the S7D image which has a high incidence angle. Several other image enhancement techniques were applied to the RADARSAT data for land cover assessment. For a more detailed analysis, the same methodological approach was applied to the RADARSAT fine mode images acquired over a small portion of the Tapajós National Forest. RADARSAT data present much more information through visual inspection than one can attain by using digital classification algorithms that are based on per-pixel classifications. This study shows the potential of RADARSAT (standard and fine mode) images for refining the vegetation cover type classification and for updating the land cover class boundaries of an existing vegetation map.
GEOSCAN ID219682

 
Date modified: