(1)
ina-
ina-
the
| Ay
atic
the
[ al,
ara-
he i
(4)
mate
rect-
ame-
ional
10t0-
Ogra-
00h50 Dd yn v:
The structure of the D matrix reflects the photo-variation
of the additional parameters. In case we use photo-variant
additional parameters then D = D; and y = y;. Therefore
the switching from photo-variant (non-metric photos) to
photo-invariant (metric photos) is an algorithmically trivial
matter.
The least squares criterion is
m
= vi P; v; = min. (6)
i=l
where P;= Q;! is the weight matrix pertaining to observa-
tions on the i th photograph
The system of normals is then
N NN, |[, à
2T .9 .. ie = i.
N NN; Xx |=] u (7)
. .. u
NI NI N, y
or in detail
N |N, N, RIN NS, N {{* u
NT N, 0 0 Ny, 0 0 X ü,
NO. RK 0 Ny, 0 X, ü,
Reo 0 .. {0 0.2% IIz, [*]1&, (8)
e. oe. u
Ny, Ny, 6 … 9IN 0. 0 Yi y
° ve u
Ny, 0 Ny, ... 0 0 N,, ... 0 Y? y?
Ny 0 «9 RS [0 0 Ny. | | 7n un
where
m m
. . HM S .
N-YN-XATBÀ
i=1 i=1
m m T
a=) =) A} Pb; (9)
i=1 i=1
Ni = AT Pi ‘À; N; = AT Pi ‘A; (10)
N,= DTP,D; N,=ATP,D; N,=A;MD; CD
and
=A} Pb uy, = Di P;b; . (12)
2.2 Definition of the reference frame
Since additional parameters are not related to the refe-
rence frame, they can be eliminated and thus the final sys-
tem is
z^ a3
R-N-N,Ny;NI r-ü-N,Nyu,
R-N-N,NyNI rf-uü-Ny,N;'u, (14)
In order to define the reference frame the minimum requi-
red constraints can be introduced with the help of relation-
ships, of the general form
Hx+Hx=z (15)
where H = [H, H, oh Hal. Besides this general form, the
constraints may refer only to e.o. parameters Hx-z (H
= 0), or refer only to ground control points H x - z (H
0).
In the most common case, in order to define a reference
frame we can simply constraint same coordinates of con-
trol points or some e.o. parameters. In this simple case the
system of normal equations is easily solved by elimination
of the corresponding rows and columns of matrices R,R
and R and the corresponding vectors r and ®. This techni-
que is valid both for minimal and redundant number of
known parameters (i.e. coordinates of control points, e.o.
parameters).
2.3 The sequential photo-wise adjustment
The following matrices can be computed for each photo-
graph using the general form (15) of the minimum con-
straints relationship
. MUR . -Ix;T e d TK —l
Ri=N;— Ny, Ny, Ny, ri =u; — Ny, Ny; uy,
~ — e IST eo Tee
Ri=N;- Ny, NI NJ + HT H; (16)
m
R-R-RR-{ART-Y (R,- R,RHRT)
j=l
— 99 in — oe
r=u-R Ri =), (i;-R; R71; )
i=1