Mapping the Habitats of Wales
By deploying Definiens’ sofware, the Countryside Council for Wales and the Institute of Geography and Earth Sciences, Aberystwyth advanced the use of remote sensing data for mapping habitats across Wales.
The Phase 1 Habitat Survey has, since the early 1980s, represented the most comprehensive national level map of semi-natural habitats and agricultural land across Wales. A new and revised draft national map (5 m nominal resolution) of Phase 1 habitats for the period 2003-2006 has been generated using satellite sensor (optical) data. The mapping has been funded by the Countryside Council for Wales, with development and delivery by Aberystwyth University and Environment Systems Limited. The classification, undertaken with object-orientated Definiens Developer and Definiens eCognition® Server, avoids traditional supervised classification approaches but instead combines expert knowledge of ecology, landscape and the information content of remote sensing data within a hierarchical rule-based process.
To generate the map, data provided by the SPOT-5, ASTER and IRS LISS 3 (P6) sensors and derived products (e.g. vegetation indices, endmember fractions of shade/moisture, photosynthetic and non-photosynthetic vegetation) were utilised. Prior to classification, each dataset was pre-processed to ensure a high level of geometric, atmospheric and topographic correction. Ancillary data thematic layers representing the extent of infrastructure and water and raster datasets describing elevation, slope and aspect were also integrated to support segmentation and classification of the landscape.
The mapping sequentially differentiated non-vegetated (e.g. urban, rocks) and coastal marginal habitats (e.g. mudflats), bracken, woodlands, hedgerows, coastal vegetation mosaics, agricultural land and remaining lowland and upland habitats (mires, heaths and moors). A range of procedures were developed to assist classification, with these considering aspects such as contrasts in reflectance between habitats, relative homogeneity/heterogeneity, context (e.g. proximity to other habitats or environments) and site (e.g. elevation). For many habitats comprised of mosaics (e.g. heath), mapping of communities (e.g. Calluna, Vaccinium-dominated) was undertaken initially using fuzzy membership functions. A secondary rule-base then assigned objects to Phase 1 habitats classes based on the relative membership of the communities contained. Over 60 Phase 1 classes and 100 sub-habitat communities have been discriminated and mapped.
The study has significantly advanced the use of remote sensing data for mapping habitats across Wales and has application across the UK and also within Europe. The maps are sufficiently detailed and comprehensive to contribute significantly to the assessment of, for example, the conservation, carbon and biodiversity values of the Welsh landscape and ecosystem services. The mapping also has application in the management of the landscape into the future and in supporting scientific investigation into modelling the response of the landscape to future climatic change. The processes developed also pave the way for routine mapping and monitoring of habitats across Wales.
For further information please contact:
Alan Brown, Remote Sensing Manager, Countryside Council for Wales,
a.brown@ccw.gov.uk; http://www.ccw.gov.uk
Professor Richard Lucas, Institute of Geography and Earth Sciences, Aberystwyth University,
rml@aber.ac.uk ; http://www.ies.aber.ac.uk
Steve Keyworth, Director, Environment Systems Ltd.,
steve.keyworth@envsys.co.uk ; http://www.envsys.co.uk
Definiens integration to ESRI ArcGIS Image Server
First prototype presented at ESRI User Conference 2008
With ever increasing volumes of remote sensing imagery, triggered by an increasing number of platforms and sensors, image accessibility has become a major bottleneck. This problem is efficiently addressed by image server technology developed by several GIS and Remote Sensing companies. We are starting to see operational implementations hosting terabyte-sized image mosaics on the Internet. It is likely that the volume will vastly increase as data providers are considering the move towards serving data through image services replacing traditional distribution models.
With the accessibility bottleneck removed, broader and richer applications of imagery will be facilitated as users can exploit the content of imagery integrated into the GIS environment. In addition large image services will also introduce stronger standardization, which, in turn, makes larger-scale image analysis applications possible. This has been demonstrated by Definiens at the ESRI Users Conference in San Diego this August when a prototype integration to ESRI´s ArcGIS Image Server was presented.
This prototype supports two specific workflows:
-Definiens users can access ESRI image services for rule set development in Definiens Developer
-ESRI users can leverage the power of Definiens eCognition® Server to extract information from image services directly within the ArcGIS Desktop environment
Using ArcGIS Image Server in combination with the Definiens Developer, data hosted in image services can be loaded directly into Definiens Developer for rule set development. Consequently, image server users do not need to create local copies of data sets in order to generate rule sets for automatic extraction of image features. In addition the data can be standardized by leveraging the dynamic imagery processing options offered by the ArcGIS Image Server.
Using ArcGIS Image Server in combination with Definiens eCognition® Server allows direct workflow integration of the image analysis component into the GIS workflow. Utilizing a small extension to ArcMap which was developed by Definiens, an area of interest can be selected by the user and sent for analysis to the Definiens eCognition® Server. The selected area is processed and results are sent back as a shape file and added to the display in ArcMap.
In an effort to gather further insight in requirements to a future product integrating image services and Definiens technology, we are currently looking for Definiens customers interested in testing this prototype. If you are a Definiens Developer user and also use ESRI ArcGIS Image Server technology, please mail to earth@definiens.com. For further details on the featured prototype click here.
Source Definiens