5. THE ORIENTATION OF PHODOGRAPUS
The orientation of the photograpns i i
several steps. Starting points are g J
and the control points being present in a cart
coordinate system. The looal zero points X , 7 G
according to typical features prespecified for the image
material to be restituted (aerial photographs, terrestrial
photonraphs, image format etc.). Within the scope of t
interior orientation tne matrix Gy is calculated from fi-
ducial mark measurements for the correction of image errors;
this matrix Gy is used for calculating xj, 75, M; dg and
the matrix K in (4.0) for all segments.
Subsequent to this the determination of matrix À is performed
within the scope of exterior orientation. The matrices Gr
and A are again used for calculating xj, Ye» Je: 55 and the
matrix X. In the case of aerial photographs Ug and Ugzg can
on principle be taken into account for the influence of the
earths curvature and refraction.
If restitution is not to be made in the control point coordi-
nate system x, y,2, then by measurements in tne model those
quantities must now be determined which are characteristic
of the relationships between the coordinate systems. In the
example of Section 4.4.1 this is the angle œ , for the resti-
tution in a cylinder coordinate system {Section 4.4.2.2.) it
is X4, % and r. The segments in the coordinate system
X, 7, Z have to be fixed by suitable prescriptions. From
their zero points we now obtain the final transformation
parameters.
Thus, the adaption of the algorithm to different problems is
accomplished by the calculation of the matrices Ug and Ugg
for each segment and their multiplication by the matrix F-.
G2
x]
t
(9
; tj
roD
®
6. OUTLOOK
In Sections 3 and 4 we explained the algorithm for an analytical
plotter using central perspective images as an example.
It is easy to see, that the subdivision of the model into seg-
ments also provides the possibility of allocating other orien-
tation elements to each segment. Thus, the restitution of
visualised scanner recordings in which other data of exterior
orientation apply to each image point or each image line ere
in principle possible. ilathematically one may proceed so that
the imaging equations (2.1) are extended by the variable
orientation elements. The time dependence is represented as a
function of the flight path in the coordinate system (x, y, z).
Hence the changes of the orientation elements are likewise
functions of x , y , z and of dx, dy, dz. The basic equations
(3.8) and (4.3) then hold just as with orientation elements
which are constant for one image.
Tor the restitution of recordings taxen with opto-mechanical
scanners, TV-systems or amateur cameras use is frequently made
of polynomials of the 2nd or 3rd degrees, which is also
possible with the described algorithm. In (2.2) the matrix &
then contains the partial derivations o? these polynomials with
respect. to x, y, Z2.
- 511 -