• Home
• Organization
• Business solutions
• For governments
• Our Research
  • Energy technology
  • Environmental modelling
  • Environmental risk and health
  • Materials Technology
  • Remote Sensing
  • Separation and conversion technology
  • Transition energy and environment
• Events
• Jobs
• News
• Contact
• Links

Description

Project Objectives

In Europe, the main regulations for biodiversity protection are found in the Habitats and Birds Directive, which provide the legal basis of the Natura 2000 network. To meet the reporting needs of these Directives, detailed, reliable and up-to-date habitat distribution maps are needed that stretch further than merely attributing a vegetation patch to a habitat type. Remote sensing methods can be employed to support these needs, but existing data and classification methods fall short of the purposes of habitat reportage in several aspects: (a) the issue of habitat structure which are most important for assessing habitat quality, has not been addressed; (b) most methods are directed at mapping vegetation species/types, and not at (Natura 2000) habitat patches; (c) most existing remote sensing methodologies have not been tested vigorously for operational purposes. The overall objective of the project is to provide  a state-of-the-art classification framework that can be used as a tool that allows better status reporting on habitats using remote sensing data. 

Methodology
For two study areas, airborne hyperspectral and hyperspectral-oriented multi-angle satellite data have been acquired: the Kalmthoutse heide (BE) and the Ginkelse en Ederheide (NL). These areas mainly consist of heathland habitat types with a typical large spatial heterogeneity in both habitat patch type and quality. Field work was performed in 2007 and 2009 to serve both the training and validation of the developed classification framework. Both spectral and (spatial) contextual information is combined in a two-step mapping process. In a first step, the hyperspectral image is classified into vegetation and land cover classes which can serve as habitat quality indicators, but as well provide the spatial building blocks for the reconstruction of habitat patches. The second step consist of a spatial interpretation algorithm combined with expert ecological knowledge to develop a habitat patch map. Combination of both the land cover and the habitat patch map enables the assessment of each specific habitat patch. To enhance the process, various remote sensing image interpretation topics are under investigation: ensemble classifiers, spectral unmixing and spatial classification to increase the classification robustness, spatial resolution enhancement through super-resolution techniques, ...

Results

• An integrated ecology-remote sensing classification framework to support the habistat status reporting needs under the EC Habitats Directive (92/43/EEC)
• A multitude of publications in peer reviewed journal papers and presentations/proceedings papers at various conferences
• Two international Workshops (first took place on 24 October 2008, second on 13 October 2010)

VITO Contribution
VITO is the coordinator of the Habistat project. The project research activities consist of enhancement of the classification of vegetation, habitat-quality indicators and habitat patch maps. VITO is also in charge of the final integration of the developed methodologies into a pre-operational framework.

Partners

• Universiteit Antwerpen – VisionLab (Belgium)
• Research Institute for Nature and Forest (INBO – Belgium)
• Vrije Universiteit Brussel (Belgium)
• Alterra/Universiteit Wageningen (the Netherlands)

Illustration

From left to right: true colour excerpt image of the hyperspectral acquisition over Kalmthoutse heide in June 2007; Land cover classification including habitat quality-indicating classes; and a habitat patch classification map.

Contact:

Birgen Haest
Tel. + 32 14 33 67 11
Fax + 32 14 32 27 95
Send a message to Birgen Haest