EVALUATION AND INTEGRATION OF URBAN HEIGHT INFORMATION INTO A DATABASE FOR RADIO PLANNING PURPOSES

摘要

With more than 10 million subscribers Mannesmann Mobilfunk is the leading GSM network operator in Germany. Due to the increasing number of subscribers in the D2 network and the resulting high traffic load, the network must be continuously adjusted. For planning radio networks topographic data is used. This includes demographic data, traffic data, information about land use and terrain height data. For urban areas, where the network is planned using very large scales, details of the location and height of buildings is necessary. Up to now the building structure of more than 50 major cities has been derived from aerial image interpretation using photogrammetric procedures. In total, an area of more than 6000 sq km of building structure data has been identified and is available in a high spatial precision. Since the production of such 3D-city models with manual photogrammetric methods and subsequent building-modelling is very time-consuming and costly, it is only applied to selected cities and up to now without modelling additional objects like trees. However, also for smaller cities and the surroundings of metropolitan areas where information about the location and height of buildings and trees is currently still missing, data with improved accuracy will be needed for network planning in the future. For this reason alternative methods with a higher extent of automation are gaining in importance. The acquisition of Digital Surface Models (DSM), i.e. precise information about the location and height of objects like buildings and trees, with methods using radar and by correlation techniques will be considered in this paper. The objective of this paper is the detection of buildings and trees in Digital Surface Models (DSM) with methods of digital image processing. A radarinterferometrically derived DSM and a DSM derived from correlation methods applied to aerial images of the Duesseldorf area serve as a source for the object detection. Different filtering techniques are used to detect objects in the DSMs. The classification of the detected objects into buildings and trees is realized by the application of special methods on the according magnitude image of the radar data and the orthophoto. In addition, information from ATKIS (Authorative Topographic and Cartographic Information System) is also used. The comparison of the results of the radarinterferometric data on the one hand and the correlated aerial image data on the other hand leads to an evaluation of the potential of radarinterferometric data for the application in radio network planning. For accuracy purposes the two approaches will also be compared with Mannesmann Mobilfunk's existing 3D-building data of that area. Finally, the exemplary integration of the detected objects (buildings and trees) into the German ATKIS is shown. The ATKIS data will be used at Mannesmann Mobilfunk for several purposes, e.g. for controlling and enhancing the network quality. ATKIS which is still in development offers nationwide detailed land use information. In the first stage of realization, urban areas are represented by polygons. ATKIS does not yet contain single buildings or trees. Hence, a way is shown how information on buildings and vegetation can be added to ATKIS in the Mannesmann Mobilfunk environment in order to increase the information content of this data.