Mail to :[email protected]
Hesham N Al-Telaihi
Fuad S Al-Aqeel
Survey Department, Kuwait
Markus H Guretzki
Inst. of Photogrammetry and Geo-Information, University of Hannover, Germany
Map data with partially inconsistent elevation information is stored in a CAD data model with limited data analysis capabilities. A Germany based company has converted the data from the CAD data model into a GIS and has generated various products
Kuwait Municipality Survey Department (KMSD) is responsible for the collection, administration and dissemination of topographic base data as part of Kuwait’s Spatial Data Infrastructure (SDI). KMSD fulfils this task by organising aerial photography survey missions and the collection of topographical vector data from air photos in so-called ‘line mapping’ projects. Line mapping is usually contracted to private industry, but quality control and administration of geo-data, as well as dissemination to other departments or private industry, is carried out by KMSD.
Line mapping is based on aerial photography stereopairs and is carried out stereoscopically by human operators. Topographical data is commonly collected in a Computer Aided Design (CAD) data model, often using Bentley Microstation or Intergraph software to store vector data in so-called design (dgn-) files.
To standardise results from various line mapping projects and contractors, KMSD has introduced technical specifications including a list of 70 different point, line, polygon, and text feature codes to describe topographic fea tures. The list consists of codes on 62 different levels, a number of which are subdivided into various colour codes. Technical specifications and feature codes are handed out to contractors in each mapping project. Results are CAD computer files and printed line maps, representing topographical features with points, lines, polygons and text of different colour, size and style (Fig.1).
Line mapping yields standardised and reliable topo-maps, providing a recent, geometrically adequate, and consistent model of the real world to users of Kuwait geo-data. But some applications remain difficult to solve with line mapping and the CAD data model resulting thereof.
- Quality control: Generation of Digital Orthophotos (DOPs, or air photos with a constant scale) is presently not part of the mapping procedure. Vector data is therefore likely to be superimposed on the source imagery only during data collection. Control of completeness and geometrical correctness of the vector data can be carried out against the background of the original air photos only if a stereo analysis tool is used. Selecting the correct stereopair for every subset of a project area from a large number of air photos can be cumbersome. Orthophotos enable easy two-dimensional (2D) measurement of position (X, Y), as well as true distances, angles and areas. If DOPs are stored with the vectors, it is easy in state-of-the-art CAD and GI-Systems to switch DOP imagery on an off as an additional data layer. In the resulting orthophoto map, 2D-data collection errors (omitted objects, geometrical inaccuracies, etc.) can be detected fast and easily by visual control.
- Data analysis: Information implicitly contained in the air photos cannot be fully exploited in the CAD data model. Raster images and vectorised topography data are only one part of a complete SDI. Attribute data linked with map features is equally important. Attributes may describe physical geography (soil type, hydrology, vegetation, etc.), socio-economical geography (demography, land ownership, land value, etc.), or man-made features (landuse, buildings, roads, utility networks, etc.). Linking map features with attribute tables or databases is more genuine to GIS than CAD. Geo-data administration in a GIS or geodatabase data model enhances data analysis capabilities. GI-systems are widespread among geo-data users today. Dissemination of geo-data in GIS formats therefore minimises data conversion and compatibility problems on the clients’ side, and raises the number of potential geo-data applications.
- Map statistics: From line maps it is difficult to calculate landuse statistics – commonly an application of interest to planning departments. In the CAD data model, individual dgn-files often describe geographic features in one map sheet limited by the map sheet boundaries, and polygon boundaries circumscribing areal features are often not closed. This is especially true for features of long linear extent, such as roads that may overlap more than one map sheet. Feature attributes often describe the type of boundary between polygons (‘fence’, ‘wall’, etc.) rather than the landuse of the parcel covered by the polygon (‘road’, ‘parking place’, ‘garden’, etc.). Map topology (or neighbour
Fig.1: Subset of a KMSD line map (orthogonal view)
Fig.2: Subset of line mapping data (oblique view). Blue lines (describing edifices outlines) jump between street and roof levels, resulting in sloped building polygons.
Fig.3: Grid DTM (shaded relief view) with automatically generated contour lines (in m above sea level). Top is north.
Page 1 of 3