The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
sets in whole or part. The Type II specification for the DSM 
creates a level of detail (lm RMSE vertical accuracy, 5 meter 
sample spacing) intermediate between lidar or 
photogrammetrically-produced products on the one hand, and 
SRTM or SPOT5 products on the other. The associated ORI 
carries a resolution of 1.25 m and horizontal accuracy less than 
2m RMSE. 
NEXTMap Britain was implemented in 2002/2003 (England 
and Wales) and subsequently extended to include Scotland (for 
a description, see Mercer, 2004). On the basis of the success of 
that project, as well as lessons learned, the decision was made 
to proceed with a NEXTMapUSA project with the current goal 
of 2009 completion. As of May, 2008 about 65% of the 8 
million km 2 in the USA (lower 48 states) had been acquired and 
over 1/4 of these data had been interferometrically processed, 
edited, QC’d and delivered to the data base repository. An 
example is shown in Figure 3 of the DSM of the State of 
California. 
Figure 3. NEXTMAP USA Example - California DSM, 
validated vertical accuracy (1430 check points) 0.76 m RMSE 
NextMAP Europe, a major trans-national program, was initiated 
in 2006 and currently includes eighteen countries in a single 
block comprising 2.2 million km 2 combined area. The data 
acquisition phase is now complete and as the various stages of 
the processing flow are completed, the database repository will 
be progressively populated with early 2009 scheduled for 
overall completion. 
Examples from two application areas which should benefit from 
the availability of NEXTMap coverage are shown here. In 
Figure 4 a flood risk application is exemplified, while in Figure 
5 a visualization example is presented. In both cases the DSM is 
draped by a high resolution colour air photo. Because of the 
availability of the 1.25 m ORI it is possible to easily ortho- 
rectify the air photos using rational functions procedures (or 
similar), making the co-registration of air-photo to DSM 
relatively simple. The flood risk application involves 3 rd -party 
models for which the DSM product is an important input 
component. The visualization example relates to many 
applications and markets ranging from recreation to automotive 
safety. In this instance it is a scene extracted from a fly-through 
(Eye-Tour). 
4. SINGLE-PASS L-BAND POLINSAR SYSTEM 
4.1 Introduction 
As noted earlier, X-Band and C-Band InSAR, which are the 
major sources of DEM data from airborne and satellite 
platforms respectively, are usually measuring the elevation of 
the upper part of the canopy, not the ground below. At longer 
wavelengths (L-Band and P-Band), attenuation is less but can 
still be appreciable (Bessette and Ayasli, 2001). The received 
backscatter signal comprises both canopy and ground 
components and the interferometric phase difference receives 
contributions from both sources, thus implying that the apparent 
phase center will be somewhere above the ground. Elegant 
methods have been created (e.g. Treuhaft and Siqueira, 2000; 
Papathanassiou and Cloude, 2001) to separate the ground and 
canopy contributions at L-Band using combined polarimetric 
and interferometric (PolInSAR) data. While considerable 
success has been demonstrated using PolInSAR, most of the 
validation effort relates to the extraction of tree height rather 
than DEM information (Zhang, et. al., 2008, and references 
therein). Moreover, all L-Band PolInSAR efforts to date have 
used repeat-pass data. This results in two significant problems: 
(1) temporal decorrelation, and (2) uncompensated residual 
sensor motions. The former degrades both tree height and DTM 
extraction accuracy while the latter generates systematic errors 
of the DEM in the along-track direction. Recent attempts 
(Reigber, et. al., 2006) to apply repeat-pass E-SAR data 
acquired during the INDREX-II campaign illustrate some of the 
problems. 
26 Teat Flood 
Figure 4. Example from NEXTMap Britain flood risk 
application 
Figure 5. Example from NEXTMap USA - California scene 
extracted from a fly-through near San Francisco 
143