| 
ESSE 
  
Lo 15] 
was 
ues 
the 
let 
lts 
oth 
in 
sly 
lts 
ts, 
ore 
ons 
mm. 
hat 
ore 
ued 
the 
ere 
soft 
out 
ror 
ues 
ega 
9m. 
for 
ent 
is 
the 
ent 
I 
the 
gly 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
5 | >| SPRONSGAE :9 
= M N i 3 2 wd 
2c |8853888388 :3 
Su it LIU LE 
0 
D |2z DSENSNNDSS ‘ES = 
312152885385858. :85]$ 
éjalsges qétar 3943 
&gl,|BE88552959 :83 |B8 
4 | N 1 ODOM NOS << n 
P5| |B333808283:82|3 
0.0 
ed BSSSEESERN:RS|R 
£95 $|RSEREESSDE:AS E 
D p INNNNAPPON:NA[N 
pil 33383333838 :838 138 
i PRRENRESASM:BM|S 
AE ERE EEE 
e AIT Nl 
b ed ipio ipipioio io iD «io | uo 
u Ds OMe—AaFT = M + 
a > NOS Sn aid n 
i|z|888888888:88|]8| . 
ES 8 
ü 530 MNOL -ND |M 
REALE gl | cs 
3|c18868088-8 :28/8 5 
cU te) 9999970990 490 © p 
o —L-— E 
S DSRSSBOMSS:xO]|G = 
ÿ4 |©BÉSHSE493:53815| | 
0. crt BF 
ws GENHIGENIT 18 |S 9 
48 © IN MIN — 0 n 
35]* RRRRSRRAS (R&R) 7 
X . > 
> EL 3 P 
C E 
5 SANNGRERTS (RR |S Ë 
t ~ vi a vi 
ef} RA HE i| € 
e BB88B3883888 88118 2 
D 
pe [n] 
5 |N SSShHRRHSNGE S be 
| fdas i888 +543 | 
28 [995833R3%% nT [© 
á 7 | E oo-coooóoodoódols 
NAMESEEEFEFEEEEERERE 
54 | €|S588338333338 | 
?z| |&98852298zR83 [P 
GcliEloo-sooóoóoóooooocol|oóe 
252% TFT SE SE SE SE SE SE SE SE ST g 
: 
c9 ANMTNOON=—NDO0OGOOAN > 
oO — I 
low when one considers that a standard 
topographic stereo-plotter will regularly achieve 
height accuracies of 1/5,000 to 1/10,000 of 
the flying height - equivalent to 0.2%. to O0.1%. 
of H. 
The results show a large spread in the 
recorded values of the orientation parameters. The 
r.m.s.e. for the X value of the left projection 
centre has the smallest value (5.7m) with the Y 
and Z values considerably larger at t10.6 and 
16.5m respectively. On the right photograph, the 
r.m.s.e. of *27.0m for the Y value is much larger 
than the corresponding values for X (14.1m) and 
Z (*9.9m). These figures show the very poor 
repeatability of the orientation. 
Coordinates measured by the APY digitizer will have 
a limited accuracy and the operator is likely 
to make some observational errors during the 
Measurement of these control points. These errors 
will affect the accuracy of the whole solution. 
As a result, the r.m.s.e. values will almost 
certainly be larger than the quoted or expected 
values, which are given in the APY literature and 
appear to be based only on the accuracy of the 
digitizer itself. 
In order to orient a stereomodel, the APY 
instrument needs at least four control points at 
which x and y parallaxes have to be eliminated. 
However, the operator will make observational 
errors in the elimination of these parallaxes and 
with the observations of the planimetric positions 
of the control points. These are the two most 
probable errors affecting the r.m.s.e. values 
obtained in a test. 
The conclusion must be that the quoted r.m.s.e. 
value mpl for planimetry (called md in the APY 
literature) of 1:10,000 of the map scale number, 
equivalent to 0.1mm on the table, is not 
realistic for practical purposes. A value of at 
least +0.15mm., and probably larger, would be 
more reasonable. 
Orientation tests were carried out on the four 
other models listed in Table 1, using the same 
procedures as described above. Detailed results 
are given in Turker (1991); only a summary is 
presented here. 
5.3. Summary of Results of the Accuracy 
and Orientation Tests 
The results from the various models have been 
summarized and are presented in a series of tables 
below. 
  
  
  
| mX | mY | mPL | mZ 
Kelvingrove |0.39|0.22| 0.48/1.02 
Llandudno (2) |0.27|0.25| O.37{1.37 
Greystoke Q.70|[0.56]-0.9012. 70 
Greenock  (1)[O.32[O.40| 0.32/[0.82 
Greenock (2) 10.3310.441 0.5611.197 
Rorbas 0.4310.171 0.4611.31 
  
  
  
Table 3. Summary Table of the Absolute 
R.M.S.E. Values in metres for mx, my, mpl and mz 
at the Control points. 
Table 3 gives a summary of the absolute values in 
metres of the r.m.s.e. values for mx, my, mpl and 
mz. The planimetric values (mpl) for the 
Kelvingrove, Llandudno and Greenock models all 
lie within the range *0.57 to O0.52m for a 
common map scale of 1:2,500. In the case of 
Greystoke, the mpl value is t0.90m which is only 
twice greater than the previous three models, 
although the photo scale (1:18,000) and map scale 
(1:10,000) are both four times greater. 
In the case of the three models 
Kelvingrove, Llandudno and Greystoke with 
fairly flat terrain, the my values are all 
substantially smaller than the mx values. Only 
in the model with hilly terrain - Greenock - 
is this situation reversed. In the case of the 
absolute height values, the values for the three 
larger scale models (Kelvingrove, Llandudno and 
Greenock) are lying in the range 10.82 to 1.37m. 
487