210 
Lingua, Rinaudo 
The data acquired on the statue of Ramsete II were regularised and the outliers and gross errors rejected, using the above mentioned 
procedure, with a regular grid of 1 cm regularly spaced points: only the meshes with more than 5 original points were treated. The 
resulting DDTM (Dense DTM) had some empty meshes which were defined using a standard median filtering. 
In order to test the quality of the obtained DDTM a traditional photogrammetric survey of a DTM with the same grid was performed. 
Table 2 shows the distribution of the discrepancies, AZ, computed between the photogrammetric and the original DTMs and be 
tween the photogrammetric and the DTM produced following the previously described procedure. Figure 8 shows the same results 
using raster images. 
AZ > 8 mm 
(green) 
8mm<AZ<16 mm 
(yellow) 
16 mm < AZ < 24 mm 
(red) 
AZ > 24 mm 
(black) 
Photogr. - original DTMs 
38% 
25% 
21% 
16% 
Photogr. - cleaned DTMs 
89% 
11% 
- 
- 
Tab. 2: Distribution of the discrepancies 
Classification: | |AZ| < 8 mm [^] 8 mm < |AZ| < 16 mm ^ 16 mm < |AZ| < 24 mm ^ |AZ| > 24 mm 
(a) (b) (c) 
Fig. 8: Photogrammetric DTM (a) - Discrepancies of the original DTM (b) - Discrepancies of the cleaned DTM (c) 
4. Integration of the laser scanner data and photogrammetry 
The cleaned data can be used to derive metric information, as shown in the previous paragraph, or integrated with other survey tech 
niques. 
Two possible integrations with digital photogrammetry are now shown: the production of 3D orthophotos and 3D navigation. 
3D orthophoto is a well known product: the photographic images are orthoprojected onto a DTM and the resulting representation can 
be used to produce planar views, axonometric and perspective images of the surveyed object, with photographic contents, which are 
useful for a better understanding and a complete recording. The required DTM is usually produced with the photogrammetric ap 
proach using matching algorithms. The level of success of the matching procedures depends on the shape and radiometry of the ob 
ject: usually at least 70% of the points surveyed with automatic procedures can be accepted, but the others must be recorded manu 
ally by the operator. The previously mentioned percentages drop drastically if the surface of the object does not have a good contrast 
level (as in the case of the statue of Ramsete II). 
As explained before, laser scanner devices can produce effective dense DTMs after automatic cleaning procedures: a DTM can be 
directly used for 3D orthophoto production. Figure 8 shows a view of the 3D orthophoto produced using STEREOVIEW (a digital 
photogrammetric package produced by MENCI and distributed by NIKON). 
The dense DTM derived from laser scanner data can offer another powerful contribution to the 3D navigation of the object with the 
stereophotomap. 
The stereophotomap is a new digital photogrammetry product that was presented by our research group in the last Cl PA Sumposium 
[Albery, 2000]. All the oriented digital images are managed in such a way that the final user can continuously explore all the blocks 
that describe the object. The operator controls a floating mark, and the coordinates of all of the points that refer to a unique reference 
system can be red in real time. A special function of the software used to explore the stereophotomap allows the automatic control of 
the floating mark in order to avoid the stereoscopic collimation of the user. This function requires a DTM of the object: the operator 
has simply to move the floating mark in the XY plane, the software interpolates the Z-coordinate from the DTM and furnishes the 
correct stereo-viewing to the operator. This results in a 3D navigation of the object in a correct metric system and the possibility of 
the user of reading coordinates, measuring distances, areas and volumes without performing any stereoscopic collimation.