The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part BI. Beijing 2008 
COST 
-MATCHING IMAGE « 
BREAK UNES 
-HOAR 
HOAR * BREAKUNES 
SO* QUALITY/SO* TIME 
UOAS 
MATCHING IMAGE » BREAK LINES 
TIME 
Figure 5. Triangle for the study with two variables 
COST 
Figure 6. Triangle for the study with three variables 
In this case, considering the three variables jointly, the best 
option is the “pure” LIDAR, closely followed by the LIDAR 
with break lines, the worst option being the correlation with 
break lines. 
5. QUALITY CONTROL 
Quality is defined as “a built-in property of something which 
allows comparison with any other thing of the same kind or 
nature and which fulfils certain requirements For our purpose, 
certain routine practices and controls have to be contributed 
ensuring the integrity, accuracy and completeness of the data, 
i.e. it must be verified that the desired quality has been achieved. 
In the case of the information provided by the photogrammetric 
method, we are dealing with aspects concerning measurement 
redundancy - say observation redundancy. The results of 
photogrammetric triangulation provide quantitative measures of 
result precision. 
• Variance component. 
• Covariance matrix of calculated coordinates. 
• Comparison of values with nominal data. 
Independent measurements to verity precision through control 
point analysis. 
• The coordinates of the photogrammetric points are 
compared to coordinates independently obtained (i.e. 
field GPS). 
Regarding LIDAR information, the control is a procedure 
subsequent to the mission to ensure and check the quality of 
registered data. The control may be made with two criteria. First, 
focusing on the “causes”, i.e. studying the behaviour of the 
elements defining the mathematical adjustment model, where 
“internal causes are considered (flight planning, external 
orientation and calibration) and external causes (flight 
conditions: direction of passes, flight altitude, etc. and type of 
terrain: height, vegetation type, etc.)”. See Habib, Advanced 
Photogrammetric Techniques: 
X PI 
s; 
' 0 ' 
\ 
y p , 
= 
Y 0 
+ &INS 
s y 
+ R m R S 
0 
+ 
e y 
_ Z PI. 
z 0 _ 
s z 
rP. 
) 
fz_ 
where: 
Xpi 9 Yph Z pl 
Xo,Y 0 , Z 0 
Rins 
8*, ô y , 6 Z 
Rm 
Rs 
P 
e x , ty, e z 
coordinates of the laser track on the ground, 
phase centre coordinates of the GPS antenna, 
rotation matrix between terrain system and IMUs. 
displacement between laser unit and phase centre 
GPS antenna. 
rotation matrix between IMUs and laser unit 
(Pao»R-A(P> Rax)- 
rotation matrix between laser unit and laser beam, 
vector range of measurement of the laser system 
(laser beam). 
random errors of components. 
Second, focusing on the “effects”, studying the results provided 
by the scatter plot, where consideration is given to “relative 
internal effects (altimetric control, planimetrie control) and 
absolute external effects (control points, difference between 
DTM and DSM, difference of intensity level, stereoscopic 
checking)”. In most cases the actual quality control of results is 
carried out with this second criterion, while possible corrections 
to those results are carried out on the elements of the equation 
chosen for adjustment. 
A working proposal must take into account the following 
elements: an appropriate mathematical model to calculate the 
LIDAR point coordinates; an algorithm of extraction of linear 
or surface features per flight pass; setting up reciprocal 
relationship between equivalent entities in overlapping zones 
(cross-coverage); an appropriate selection process in the linkage 
entities; applying a least square adjustment to the model. The 
ideal solution is based on a high degree of automation (minimal 
user interaction), use in standard flight missions, minimal 
requirement of additional flights (i.e. cross-passes) and a 
minimal number of ground control points. 
6. CHANGE DETECTION 
After having generated the orthophotography through the 
previous processes, cartographic exploitation comes next. Here 
we raise the possibility of obtaining, more or less automatically, 
linear cartographic objects and features that have changed 
between two images (orthophotos) taken on different dates, 
which is commonly known as change detection. 
The automatic change detection may be carried out according to 
the technique based on the matching of an image pair or by 
extraction of linear features of an image of the current state of 
the area and subsequent comparison of this vector layer with the 
official vector file, thereby updating the existing cartography. 
Anyway, we are dealing with image processing, a set of 
operations that are applied to images in order to enhance or tone 
353