A 3d City Model of Kuwait: Data Processing and Possible Applications



Objectives
Objectives of this paper are
  • to explain how KMSD topographic base map data is processed in order to generate a DTBM, and
  • to show examples of how DTBMs are used in technical applications and planning, e.g. by the telecomm industry, in environmental studies on the propagation of air pollutants or noise, and in CAAD.
Material & Data
Test data from the KMSD topographic base map is used to demonstrate data processing for the generation of a 3D city model (DTBM) of Kuwait. DTBMs and projects from other cities, selected from the PHOENICS GmbH archive, are used to demonstrate model applications.

As test data, one sheet of the KMSD topographic base map, covering 1 km by 1 km of urban residential area in flat terrain, is selected from a recent line mapping project. The following data is used:
  • digital scans of 14 mm resolution, of three colour aerial photographs (two stereo-models), taken consecutively in east-western direction at an image scale of 1: 6.000 with an overlap of 60%, using an aerial survey camera with a 302.98 mm focal length lens. Scans were stored in compressed TIFF format, including image pyramids, with a radiometric resolution of 24 bits (8 bits per colour channel),
  • calibration data describing the camera and lens, including co-ordinate information of eight fiducial marks,
  • orientation parameters of the air photos (from aerial triangulation calculation),
  • vectorised topographical data from stereoscopic line mapping, stored in dgn file format (CAD data model),
  • the standard KMSD line mapping feature code list.
DTBMs of the German Cities of Bremerhaven and Dachau are selected from the PHOENICS archive to demonstrate possible DTBM applications.

Methodology
DTM: A DTM of the Kuwait test area (excluding buildings, structures and vegetation) is generated. To derive the DTM, the KMSD topographic map data is transfered from the CAD system into a GIS environment (ArcGIS by ESRI). Features containing elevation information are selected from the vector data. A Triangulated Irregular Network (TIN) is generated from cell (i.e., point) features (spot heights, traffic light base heights, high tension tower base heights, and manhole heights), and line features (5 m and 1 m contour lines). A regular grid DTM of 2 m cell size is derived from the TIN.

3D City Model: To generate the Kuwait test area DTBM, the air photo stereopairs and image orientation data are imported into stereo analysis software. Using stereo glasses controlled by an infrared emitter for stereoscopic viewing, roof top elevation is measured for every building according to the following specifications, which are developed by PHONICS GmbH in co-operation with vodafone D2 GmbH for telecomm applications in Germany:
  • Roofs composed of sections larger than 5 m2 are separated when the elevation difference between sections is 2 m or more. One elevation value is stored with every roof section.
  • If more than 50% of a roof is covered with roof top assemblies (water tanks, elevator shafts, etc.), the elevation of the assemblies is measured and is assigned to the entire roof.
Values are stored as absolute (above sea level) and relative (above terrain) elevation.

A Digital Orthophoto (DOP), an orthophoto map and a landuse map of the Kuwait test area are also produced. The applied methodologies and results are described in Hermsmeyer et al. (2005).

3D-flythrough animations: Flythroughs are generated from the orthophoto mosaic, landuse map, and DTBM using animation software. Results are stored in Audio Video Interleave (avi) file format.

Several animations of increasing photo-realistic appearance are generated, using (i) the DTBM and landuse map, (ii) the DTBM and orthophoto-mosaic, and (iii) the DTBM, orthophoto-mosaic and terrestrial photographs of building facades and vegetation. To visualise facades photo-realistically, terrestrial photos of facades are rendered randomly on the vertical ‘walls’ of the DTBM.

Various mathematical algorhithms and software products are used to compute line-of-sight, visibility, mobile phone signal strength, air pollutant distribution, or noise propagation with use of DTBMs. Visualised results of such computations are shown in this presentation to demonstrate possible DTBM applications, but computational algorithms are not discussed. Information on the methods applied is available through PHOENICS GmbH upon request.

Results & Discussion

Kuwait test area results:
DTM: The Kuwait DTM grid (cell size 2.0 m, 500 rows x 500 columns, Xmin: 514,000.00, Xmax: 515,000.00; Ymin: 205,000.00, Ymax: 206,000.00) is shown in Fig.2. The terrain is almost flat, with some road embankments visible in the northern part. The highest point is at 37.63 and the lowest point is at 24.00 m above sea level; average elevation is 30.44 m. We list these values to demonstrate possibilities of numerical applications, such as those described below.

3D City Model: The Kuwait DTBM is shown in Fig.3. A cell size of 0.5 m is used to conserve sufficient building details for DTBM applications. The model fulfils criteria as required by the telecomm industry.

In Fig.4, a subset of the DTBM is shown, superimposed with building polygons from the vector data (manually subdivided and with corrected elevation). Dark yellow polygons represent low, and bright polygons represent high buildings. Lengths of building shades correspond to building elevation.

The DTM and DTBM of the Kuwait test area fulfil telecommunication industry criteria as listed in the methodology section of this presentation. In the following section we will demonstrate examples how comparable elevation models of other cities are used in a number of different applications. The Kuwait DTM and DTBM can be used in similar applications.


Fig.2: Grid DTM (shaded relief view) of the Kuwait test area with automatically generated contour lines (values in m above sea level). Top is north.



Fig.3: DTBM (shaded relief view) of the Kuwait test area.



Fig.4: Enlarged subset of the DTBM (grey-scale shaded relief view), superimposed with building polygons.

Examples of possible DTM and DTBM applications:
GSM base station visibility: Planning GSM (Global System for Mobile communications) and comparable mobile communication antenna networks is an important application of 3D city models. Telecomm companies realise the bulk of their business and turnover in highly developed urban areas where high rising buildings are common. Mobile phones never communicate directly with each other, but through the closest base station of an antenna network. GSM signals travel best along visible lines between the closest base station and the mobile phone, but signals are absorbed and reflected by building materials. The aim of antenna network planning is to minimise areas with no reception, i.e. areas shaded by edifices from base station visibility, while minimising areas of signal overlap to reduce network operation and maintance costs. An example from the City of Bremerhaven in Germany is shown in Fig.5 (a and b). Proposed base station locations are in the centres of the two circles, 1 and 2 (Fig.5a).


Fig.5, a: Modelling GSM base station visibility in a DTBM (2D view, image by PHOENICS GmbH).



Fig.5, b: Modelling GSM base station visibility in a DTBM (3D view, image by PHOENICS GmbH).

Base station network planning is carried out with use of DTBMs. Co-ordinates (X, Y, and Z) of the proposed base station location are selected in the 3D city model, and a circular buffer is defined around the location to describe the area covered by the station. Spatial trigonometry is applied to determine DTBM raster cells inside the buffer visible from the proposed location. A colour code is assigned to all visible cells. An offset of the base station over the roof top at its proposed (X,Y) location can be described by increasing the Z-co-ordinate, simulating the effect of a base station mast.

More advanced models of GSM signal propagation take into account the flux density of the signal field strength (indicated in Watts per meters square). Signal strength does not break down rapidly when a buffer area around the base station is left (as is suggested in Fig.5), but rather it decreases continuously as the distance from the base station increases. In combination with signal absorption and reflection on building materials in urban areas, this results in a geographical distribution of signal reception which is more complex than suggested by base station visibility alone. An example from the City of Dachau near Munich in Southern Germany is shown in Fig.6.


Fig.6: Modelling GSM signal strength in a DTBM (image courtesy of enorm GmbH).

Distribution of air pollutant concentration and noise: Car traffic is often dense in urban areas, and causes emissions of noise and air pollutants (dust particles, carbon-particulate matter, nitrogen oxides, etc.). Objectives of planning an urban road system should be to direct car traffic efficiently (i.e., to minimise traffic congestions) while keeping through-traffic away from residential areas and densely developed inner cities. For new settlements and extensions of the road system the amount of pollutant emissions can be estimated from expected traffic densities (number of cars or trucks per hour). For existing roads, emissions can be estimated from traffic census. Numerical algorithms and computer software are available to calculate pollutant concentrations and noise levels under various atmospheric and wind conditions. Pollutants have a tendency to accumulate in “urban canyons” along narrow streets between buildings. Therefore DTBMs need to be included in computations of concentration distributions. An example is shown in Fig.7.


Fig.7: Modelling air pollution caused by car traffic using a DTBM (image courtesy of GeoNet Environmental Consulting GmbH).

Page 2 of 3
| Previous | Next |