1:50000 SCALE TOPOGRAPHIC MAPPING IN WEST CHINA USING SPOT 5 DATA

摘要

On the one hand, the successful launch of the SPOT 5 satellite in 2002 can be seen as a milestone for DEM extraction and mapping, because its payload includes HRS, a stereoscopic imaging instrument devoted to collection of large areas. Operating in panchromatic mode, the HRS instrument has two telescopes, one pointing forward and one aft of the satellite. HRS is thus able to cover an area of 600 km×120 km in a single pass (ie 72 000 km2 stereoscopic strips). Moreover, SPOT 5 is able to collect stereoscopic triplets, through a simultaneous activation of HRG sensors in vertical mode, which will overcome the difficulty of terrain shadow in DEM matching. Meanwhile the extremely precise ancillary metadata of SPOT 5 allows accurate geometric processing with sparse ground control points. Thus SPOT 5 stereoscopic imagery becomes one of main satellite data sources for accurate DEM extraction and quickly mapping large area. The current status and coverage of stereoscopic HRS imagery as well as the metadata of SPOT 5 will be briefly exposed.One the second hand, about 2 million km2 in west China never has been covered by 1:50,000 scale topographic maps for several decades. The area contains the famous Qingzang plateau with average 5000m altitude and boundless Talimu basin desert. Mapping this area with traditional methods had unimaginable difficulties and was impossible to complete at all. With the high technology advancing, especially the great progresses in geomatics, it is the time to start to map the area with new geo-information technologies. So in 2006, China decided to initiate a large project, named 1:50,000 scale topographic mapping of west China (ab. west China topographic mapping project), in order to get the 1:50,000 scale topographic maps (DEM, DOM and DLG) of that area within next 5 years. As we know the weather is extremely atrocious and a lot of sites are out of reach in that area, mapping from aero-born and satellite remote sensing imagery with spares ground control points was determined to be the main solution for the project. Within the several satellite remote sensing data used in the project, SPOT 5 data played an important role with its distinguished characteristics. The paper will present the methodology of mapping from SPOT 5 data developed and implemented by CASM. The methodology emphasizes to use as spares ground control points as possible to facilitate the data processing. A simple general geometric RPC sensor model was established instead of the complex rigorous SPOT 5 sensor model. The model has been validated in several test sites using SPOT 5 HRS trips. The results proved that the model has the same accuracy as the rigorous model and just needs sparse ground control points to carry out the triangulation. Then, High precise matching algorithms considering feature points and feature lines were used to extract DEM from HRS stereo images or triplets. Comparing with some results get from other satellite images shows that the DEM has very fine textures. Contour lines were experimented to be interpolated automatically by DEM and DOM products were made finally. Sanjiangyuan block is one of the 8 mapping blocks of the project. It locates on the north of Qingzang plateau with average 4500m altitude and 120,000 km2 area. In 2006, 250 sheets of maps of the block were produced using the methodology. Over 100 ground control points within 6 maps were accurately measured in field with GPS to assess the accuracy of DEM, DOM and DLG products. The presentation will detail the accuracy assessment results, and give some comparison results of contour lines obtained by automatic interpolation with DEM and by manual stereo measurements. In conclusion, the methodology of large area mapping with SPOT 5 data and spares ground control points is successfully established in the west China mapping project. It was proved that the accuracy satisfied our standard demand of 1:50,000 scale map. Furthermore the west China mapping project can be an example for other mapping projects in difficult mapping areas considering weather, altitude and transport conditions, etc.