Integration of Remotely Sensed Data and Cadastral Ownership Boundaries for LandInventory Purposes

摘要

The objectives of this study were to assess: the possibilities of using cadastralownership boundaries to extract land cover data from satellite images by means of the object classification; the accuracy of the cadastral land use and remote sensing derived land cover data; and the use of satellite remote sensing for updating the cadastral land use data. For this purpose the digital cadastal boundaries and a Spot and Thematic Mapper image were required for the Lunteren area (The Netherlands). The Lunteren area is a rural area of 3600 ha, consisting of 2110 cadastral parcels. Both satellite images were used for a multitemporal object classification which resulted in one land cover label per cadastral parcel. The underlying assumption of the object classification is that one object (one cadastral parcel) has one land cover type. The multitemporal classification was realized by means of a look up table. The classification results were compared with observations from a field survey in which 289 cadastral parcels were visited. The selection of these parcels was based on a number of criteria to enable comparison between the land use data as stored at the Cadaster (in AKR, an information system that stores administrative information for each cadastral parcel), and the land cover derived from remote sensing.