VHR land cover maps

Deliverable 5.2 (D5.2), entitled “Habitat Maps from Very High Resolution (VHR) imagery” is a deliverable of WP5 and focuses on the production of Food and Agricultural Organisation (FAO) Land Cover Classification Systems (LCCS) maps from very high resolution (VHR) spaceborne optical remote sensing data, primarily Worldview and Quickbird, and their translation to life forms, surface aspect and physical status. D5.2 consists of two parts and includes: 1. A review of the semantic nets language and descriptors. 2. An overview of VHR spaceborne data available for the BIO_SOS study sites and the approach to data pre-processing. 3. An overview of field data collected to support the classification of FAO LCCS categories and their translation to General Habitat Categories (GHCs). 4. Land Cover class description in 3D world based on semantic networks, with focus initially on a generalized classification to LCCS Level 3 and, subsequently, on more specific classification of lower level classes. 5. Methods for representing and including semantic nets for the classification of LCCS categories, with a detailed knowledge-based example list provided for the Italian site. 6. The approach to classification based on spectral information as well as derived products (e.g., textural measures, vegetation indices). 7. VHR maps of land covers for the test sites in Italy (Le Cesine), Wales (Cors Fochno but also Cors Caron) and the Netherlands (Ginkelse, Eder heide and Wekeromse Zand). 8. Approaches to assessment of classification accuracy. The Deliverable has established that consistent classification to LCCS Level 3 categories can be achieved between the four Natura 2000 sites and surrounds using a combination of spectral data and indices and ancillary data layers (e.g., existing cadastral information). However, classification to lower levels of the LCCS is less consistent between sites because of differences in land covers occurring, the uniqueness of some and the different types and dates of the remote sensing data used. Improvements to the classifications are, however, anticipated with the inclusion of data from multi-temporal VHR, hyperspectral sensors and Light Detection and Ranging (LiDAR). Report on VHR Land Cover maps, BIO_SOS Biodiversity Multisource Monitoring System: from Space TO Species, Projet FP7-SPA-2010-1-263435 (PDF Download Available). Available from: https://www.researchgate.net/publication/298354288_Report_on_VHR_Land_Cover_maps_BIO_SOS_Biodiversity_Multisource_Monitoring_System_from_Space_TO_Species_Projet_FP7-SPA-2010-1-263435?channel=doi&linkId=56e8af8708ae9bcb3e1cda38&showFulltext=true [accessed Jan 16 2018].