The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Articles | Volume XXXVIII-4/W19
https://doi.org/10.5194/isprsarchives-XXXVIII-4-W19-369-2011
https://doi.org/10.5194/isprsarchives-XXXVIII-4-W19-369-2011
10 Sep 2012
 | 10 Sep 2012

QUANTITATIVE ASSESSMENT OF DIGITAL TOPOGRAPHIC DATA FROM DIFFERENT SOURCES

N. Yastikli, F. Esirtgen, and U. G. Sefercik

Keywords: Accuracy, Quality, DEM, Quality, Photogrammetry, Remote Sensing, SAR, SPOT

Abstract. This study includes the quantitative assessment of digital topographic data of the terrain from photogrammetric methods, satellite imaging and RADAR techniques in test side in Istanbul. The side area covers 10 × 10 km2 and includes five different land cover types. The 3 m sampled DEM, which was produced by 1/1000 scaled photogrammetric maps is used as reference and tested against the 94 Ground Control Points (GCP). The 5 m and 30 spacing DEMs derived from 1/5000 scaled photogrammetric maps are used as test data sets as well as 20 m spacing SPOT DEM and 80 m spacing ASTER DEM. Besides, 90 m spacing DEM generated from InSAR technique with SRTM mission was also used as test data. The quantitative assessment of all DEMs were performed and reported based on FEMA and NDEP specifications. The quantitative assessment was carried out for open terrain (%19), forest (%15), built-up areas (%32), scrub and bushes (%26), and rough terrain (%8). The accuracy assessment and quality analyses had been conducted for each terrain classes to reflect the quality with more suitable approach. The 3 m spacing reference DEM is tested against the 94 GCP and obtained Root Mean Square Error (RMSE) for Z was in the range of 0.85 m (open terrain) to 1.94 m (rough terrain) for different terrain classes and 1.45 m using all terrain classes. The more than 20 GCP was available for each terrain classes for quantitative assessment of DEMs. As a result of quantitative assessment, RMSE Z of tested DEMs from photogrammetric methods, satellite imaging and RADAR techniques were computed for each land cover types based on international standards. It is no surprise that, the DEM from photogrammetry were more accurate than the DEM from satellite imaging (SPOT, ASTER) and RADAR (SRTM). As expected, open terrains have better results than the other classes like forest, scrub, built-up and rough terrain for all DEMs. The accomplished results of the quantitative assessment demonstrated the importance of the data source, resolution, and production methods of DEMs. The terrain class is important factor and can not be disregarded in quantitative assessment of digital topographic data.