Community Research and Development Information Service - CORDIS

Abstract

An experiment has been conducted to investigate the possible use of combined ERS-1 SAR data and Landsat TM data for mapping land cover in complex Mediterranean landscapes. TM and SAR data were obtained from a test area in Portugal and geometrically co-registered. The SAR data were speckle filtered using a multiple regression technique based on the TM imagery. Classification of the multi-sensor data set was then performed using a multi-layer perceptron neural network. The addition of the SAR data yielded a small improvement in overall per-pixel classification accuracy, with the most significant gains in the discrimination of cereals, forests and urban areas.

Additional information

Authors: WILKINSON G G, JRC Ispra (IT);KANELLOPOULOS I, JRC Ispra (IT);MEHL W, JRC Ispra (IT);HILL J, JRC Ispra (IT)
Bibliographic Reference: Paper presented: PECORA-12 Remote Sensing Conference on Land Information from Space-based Systems, Sioux Falls, South Dakota (US), August 24-26, 1993
Availability: Available from (1) as Paper EN 37822 ORA
Record Number: 199311324 / Last updated on: 1994-11-28
Category: PUBLICATION
Original language: en
Available languages: en
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top