1113 - Relate the spatial and spectral characteristics of EO data to the types and proportions of materials found within the scene and within pixel IFOVs to relabel spectral classes as information classes of a classification scheme

Concepts

• [IP3-4-3] Classification schemes (taxonomies)
  Long-term monitoring of land cover and land use are particularly relevant for land ecosystem monitoring. Therefore, baseline datasets are necessary that allow assessing changes of land cover and land use where the class definitions remain consistent over time. Accordingly, classification schemes have been established that adhere to taxonomically correct definitions of classes of information organized according to logical criteria. If hard classification is to be performed (i.e. without fuzzy class boundaries), the classes in the classification system should normally be mutually exclusive, exhaustive, and hierarchical. Mutual exclusive classes have no taxonomic overlap and assign a land cover patch to a single class. An exhaustive classification scheme is able to cover the area of interest comprehensively and leaves no land cover patch unassigned. A hierarchical system allows combining sub-classes into higher-level categories. From a remote sensing classification perspective, it becomes clear that a classification scheme consists of information classes defined by human beings. Conversely, spectral classes are those inherent to EO data. An analyst must identify spectral classes and label them as information classes that satisfy bureaucratic (or scientific requirements). Additionally, the advantage of using established classification schemes is that their use in scientific studies and applications produces results that are comparable to other studies and suitable for sharing of data. Established classification schemes include: CORINE land cover (CLC), Land cover classification system (LCCS), American Planning Association land-based classification standard, United States Geological Survey land-use/land-cover classification system for remote sensor data, U.S. Department of the Interior Fish & Wildlife Service classification of wetland and deep water habitats of the United States, U.S. National Vegetation Classification system (NVCS), International Geosphere-Biosphere Program IGBP Land cover classification system.