This system complies with laser safty class II, 
because a visible semiconductor laser is used. 
With reference to the data in table 2-1 the 
ranging performance can be calculated. Assuming 
distances to the targets' surfaces from the 4D- 
LM's output of 0.5 m and 1.0 m respectively the 
maximum possible standard deviation of a single 
shot measurement as a function of the reflectivity 
of the target is plotted in fig. 2-4. It was 
presumed that the target is a Lambertian scat- 
terer. 
Fig. 2-3 shows that the 4D-LM fulfills the 
requirements of machine-tool and molding industry. 
Remarkable is the high sampling rate of the 4D-LM. 
One measurement lasts only 0.5 ms. A range image 
of e.g. 200x200 pixels takes only 20 s. As already 
mentioned each pixel contains not only the Carte- 
sian coordinates derived from the slant range R 
and scan angles a, and a, but also the intensity 
the user gets real three dimensional image from 
the target's surface. The intensity information 
can be used to extract addtional information e.g. 
labels on the workpiece. 
In the following chapters it will be shown by 
practical examples, how 4D-LM 's data can be 
processed so that the geometry of the object's 
surface is described. It will be shown that with 
the 4D-LM in combination with the new software 
module a very powerful tool exists which closes 
the gap for digitizing objects for later CNC- 
manufacturing. 
distance | 1 m 
0.30 
0.20 
g 1.00 
© 0.90 
À 
0.80 
5 
0 
m 
d 0.60 
À 
2 0:50 
a 
s 040 
HI 
e 
kel 
a 
a 
4 
un 
0.10 
distance | 0.5 m 
  
0.00 
0.00 20 00 40.00 60 00 80.00 100.00 
Reflection in % 
Fig. 2-4: Theoretical Accuracy of 4D-LM 
3. INTERNAL COMPUTER MODEL GENERATION 
After the digitizing of freeform surfaces by the 
4D-LM measurement data are present in absolute 
Cartesian coordinates. They are transformed in 
object coordinates by a postprocessor. Now, the 
coordinates can be subsequently processed as 
either NC-commands for dublicating milling 
machines or in data format for the sculptured 
surface modeller. 
3.1 Postprocessing of Objects Coordinates for 
Dublicating Milling Machines 
Analyzing  NC-programs of  dublicating milling 
machines we found out that the coordinates of the 
cutter center point (x Yo: Z4) represent a set of 
cp’ 
loop contours. Normally the determination of loop 
using 
contours is very time consuming. However, 
the 4D-LM these special data sets can be deter- 
mined very effectively by straight forward calcu- 
lations. 
As the object raw data are still distored by noise 
and therefore not useful for determination of 
smooth contours, they are filtered first by using 
best fit algorithms (Olomski, 1989) which can be 
interactively selected by the user. In a second 
processing step the loop contours are computed by 
defining virtual planes which are perpendicular to 
  
Fig. 
3.1-1: 
Contours and 
Axis 
Definition of the Cutter 
  
  
  
  
Fig. 3.1-3: Perspective View of the Workpiece's 
Contours