Community Research and Development Information Service - CORDIS

Abstract

A classification method, using global statistics defined over individual fields, is proposed for discrimination and inventory of agricultural land occupation classes. It is compare with pixel by pixel classification in the case of SPOT simulated visible IR data. It is also tested with SAR data and combinations of visible IR data where the speckle effect is a major difficulty. The results show an improvement around 18% in the classification accuracy in the first case and a discrimination performance better than 90% for 5 agricultural categories in the second case.

Additional information

Authors: MEGIER J, JRC ISPRA ESTAB. (ITALY);TECHNICAL UNIVERSITY OF BERLIN (GERMANY);MEHL W, JRC ISPRA ESTAB. (ITALY);TECHNICAL UNIVERSITY OF BERLIN (GERMANY);RUPPELT R JRC ISPRA ESTAB. (ITALY), JRC ISPRA ESTAB. (ITALY);TECHNICAL UNIVERSITY OF BERLIN (GERMANY);TECHNICAL UNIVERSITY OF BERLIN (GERMANY), JRC ISPRA ESTAB. (ITALY);TECHNICAL UNIVERSITY OF BERLIN (GERMANY)
Bibliographic Reference: 18TH INTERNATIONAL SYMPOSIUM ON REMOTE SENSING OF THE ENVIRONMENT, PARIS (FRANCE), OCT. 1-5, 1984 WRITE TO CEC LUXEMBOURG, DG XIII/A2, POB 1907 MENTIONING PAPER E 32171 ORA
Availability: Can be ordered online
Record Number: 1989123091400 / Last updated on: 1987-01-01
Category: PUBLICATION
Available languages: en