A Use of Object-Oriented Classification Methods to Map Coastal Regions from High-Resolution...

A Use of Object-Oriented Classification Methods to Map Coastal Regions from High-Resolution Satellite and Airborne Digital Imagery

SHARE

Louis Demargne
EarthData International,
USA
Email: [email protected]

Chad Lopez
EarthData International
Email: [email protected]

Kass Green
Alta Vista

This paper presents the results of various studies conducted by EarthData International and Alta Vista on the use of object-oriented image classification methods for mapping land use/land cover and benthic habitats in coastal regions. Source imagery for these studies included multispectral QuickBird satellite imagery and airborne Leica ADS40 digital imagery. Image classification was completed using a combination of CART analysis, object-oriented classification, and manual interpretation. Different methods and softwares for polygon delineations were explored for accuracy and cost-effectiveness. The softwares used were Visual Learning System’s Feature Analyst and Definiens eCognition.