USE OF REMOTE SENSING TECHNIQUES FOR VALIDATING HEIGHTS AND DEPTHS

MIDTERM REPORT FOR SSG 3.184

Philippa Berry

Meetings have been held on an ad-hoc basis at conferences, supplemented by email correspondence.

Work on several different radar based techniques for remotely sensed height determination is being carried out by members of this study group.

A global scale assessment of current Global Digital Elevation Models has been performed using an independently derived height dataset, containing over 100 million points. These data were calculated from satellite altimeter data from the ERS-1 Geodetic Mission. Models evaluated include GLOBE v1., GTOPO30, JGP95E, TerrainBase and ETOPO5. The results confirm the presence of substantial errors common to all models tested. These errors have been attributed to reliance on a common stock of a small number of data sources. The most consistent agreement was obtained where data were sourced from DTED; poor results were generally obtained where the source was Digital Chart of the World or the International Map of the World.

Because these errors were found to be contaminating existing GDEMs very significantly, it was decided to create a new GDEM, ACE (Altimeter Corrected Elevations, produced at De Montfort University) by fusing the best of the available ground truth with a DEM derived from the satellite-based dataset.. The first release of this new GDEM has just been made. Globally, 28% of values have been outright replaced with new altimeter based data, and a further 17% corrected for vertical misplacement. The new model is currently being evaluated by members of this group.

However, this approach does not enable detailed mapping, nor mapping over mountainous regions where altimeters to date have lost lock. Several different techniques and missions have been deployed to create accurate detailed DEM data on a global scale. The recently flown Shuttle Topographic Mission created a huge dataset which is currently being processed. Data post-processing has proved very complex, resulting in some delays to the original schedule. However, results should be released soon, enabling assessment of these vital new DEM data to be carried out within the next two years. Additionally, work on InSAR and stereo SAR derived DEMs continues, forming another increasingly important datastream as the application of these techniques to mapping becomes more widespread.

The launch of Envisat within the next few months should further increase the available data pool, both by providing an additional source of SAR data and by deploying the RA-2, an altimeter configured with an additional tracking mode to allow the instrument to remain in lock over more extreme terrain.

All indications are that within the life of this working group, a considerable advance will have been made in the mapping of the earth’s topographic surfaces.

TABLE OF CONTENTS                                                          TOP