Per-field classification of land use using the forthcoming very fine spatial resolution satellite sensors: problems and potential solutions

摘要

This chapter reports research undertaken as part of a project within the Applications Demonstration Programme,led by the British national Space Centre and the Ordnance Survey.The objective of the project was to develop an automated operational system for classifying land use with fine spatial resolution satellite sensor imagery at the local scale,that could later be extended to the national scale.Land use classes were selected according to the specifications of the National Land Use Stock System (NLUSS) and classification was performed on a per-field basis (field refers to a parcel of land including semi-natural areas,agricultural fields and urban roads,buildingsand gardens) by utilizing digital vector data and geographical information systems (GIS).Per-field,as opposed to per-pixel,classification provided a vehicle within which the spatial variability and texture inherent in fine spatial resolution imagery could be utilized.In performing the classification several sources of misclassification were identified and addressed.As a result per-field classification was 8