te of 
tting 
aths" 
cates 
arrov 
than 
, the 
This 
on of 
' must 
|^ are 
n and 
ously 
'ibute 
ected 
, the 
7 road 
| map 
they 
tized 
ional 
paper 
: Vas 
“ed by 
aring, 
ine to 
m the 
red. 
on in 
E the 
stical 
  
tional 
IS was 
as the 
curacy 
s are 
rin 
normal 
e, the 
the 
tional 
| m or 
tting, 
yy the 
POSITIONAL ACCURACY TEST 
MONDPLOTT ING 
  
  
  
  
  
  
  
  
  
  
  
17 |= 
18 |- 
45 - 
14H 
13 | 
€ 42 - 
è nr 
s 10 - 
pot 
2 "T 
- sk 
= 
= aL 
£ si 
+} 
3} 
2 + 
1 
0 T 5 1 +5 | #7 ] #e [1011 | 42-14 | 16718 | 22-28 | 30-40 
4-2 3-4 5-8 7e 9-10 11-12 14-18 18-22 28-30 
Distance Cm.) 
POSITIONAL ACCURACY TEST 
PA- 2000 
70 
so} 
so |- 
& 
M 
? 40 
i E 
3 
de 
€ 
e 30 r- 
» 
u 
s 
2 ag |- 
10 [- 
a 2-3 4s | 87 | 88 | 10-11 | 42-14 | 46-18 
+2 5 5-8 7-8 9-10 11-12 14-18 
Distance Cm.) 
Figure 2: Relative frequency line graphs of data 
of digital monoplotting and PA-200. 
Confidence level Monoplotting PA-2000 
m z*SD+Mean Accuracy (m.) z*SD+Mean Accuracy (m.) 
68 0.468*3.8+4.35 6.13 0.468*1.09+1.06 1.57 
80 0.84*3.8--4.35 7.54 0.84*1.09+1.06 1.98 
85 1.044*3.84-4.35 8.32 1.044*1.09+1.06 2.20 
90 1.28*3.8+4.35 9.21 1.28*1.09+1.06 2.46 
95 1.645*3.8+4.35 10.60 1.645*1.09+1.06 2.85 
98 2.055*3.8+4.35 12.16 2.055*1.09+1.06 3.30 
  
  
  
  
  
  
  
Table 4: Positional accuracy 
normal distribution 
model. 
The accuracies with the same 
in table 5. 
estimation using 
statistical 
as 
confidence levels 
were calculated from the sample data and are shown 
  
Confidence level (76) Monoplotting Accuracy (m.) 
PA-2000 Accuracy (m.) 
  
68 5.10 
80 6.96 
85 8.02 
90 921 
95 11.05 
98 13.60 
  
  
  
1.24 
1.62 
1.79 
2.08 
2.52 
3.46 
  
  
Table 5: Positional accuracy, 
sample data. 
calculated from 
497 
Positional accuracy requirements 
Positional accuracy requirements vary according to 
application. For example, according to United 
states national map accuracy standards for 
horizontal accuracy of maps on publication scales 
of 1:20,000 or smaller, no more than 10 percent of 
the points tested can be in error by more than 
1/50 inch (0.5 mm). These accuracy limits apply in 
all cases to positions of well-defined points 
[12]. A line is captured by digitizing both end 
points; thus the accuracy of lines is a function 
of the positional accuracy of the points involved. 
The accuracy specifications of lines can be 
converted to epsilon distances [3]. 
In order to cover a wide spectrum of accuracy 
requirements, epsilon distances (1/2 of the 
epsilon band) ranging from 0.1 mm to 0.25 mm on 
the map were translated to meters on the ground 
for several map scales (see table 6). 
  
  
  
Scale Positional accuracy requirements for spatial data 
0.1 mm 0.15 mm .2 mm 0.25 mm 
1:20000 2m. 3m. 4m. 5m. 
1:25000 2.5 m. 3.75 m. 5m. 6.25 m. 
1:30000 3m. 45m. 6m. 75m. 
1:50000 5m 75m. 10 m. 12.5 m. 
  
  
  
  
  
  
  
Table 6: Positional accuracy requirements of 
spatial database in several scales. 
The estimated positional accuracies of digital 
monoplotting and stereoplotting at the 90% 
confidence level (see table 4), were compared with 
the positional accuracy requirements of the above 
table; the results are summarized in table 7. They 
indicate the expected suitable scales for which 
the digital data produced by digital monoplotting 
and stereoplotting can be used. 
  
  
  
  
  
  
  
  
  
  
  
Scale Positional accuracy of Monoplotting (£9.21 Positional accuracy of PA-2000 (£2.46 m. with 
m. with 90% confidence level) 90% confidence level) 
i.lmm +.15mm i2mm | 25mm | i.imm +.15mm +.2mm +.25mm 
1:20000 - E = - - + + + 
1:25000 - - - - + + + + 
1:30000 - - = - + + + + 
1:50000 - - + + + + + + 
Kemarks + positional accuracy under standards 
  
  
  
  
  
  
  
positional accuracy out of standards 
Table 7: Expected suitable scale of digital data 
produced by monoplotting and PA-200. 
The data collected by  monoplotting from 
photographs at scale 1:30,000 are suitable of 
updating a database at scale 1:50,000 with 0.4 mm 
accuracy at 90% confidence level, while data 
collected by stereoplotting are suitable for 
producing and/or updating a database at scale 
1:20,000 with 0.3 mm at the same confidence level. 
For proper appreciation of the results, the 
following facts should be emphasized. 
. The data used for the positional accuracy 
evaluation were derived from line coverage 
collected in dynamic mode and converted to point 
coverage. 
The data were collected by an operator with 
limited experience. 
The results were derived from only one 
experiment on 1:30,000 scale photographs. 
No magnification was used for the data collected 
by monoplotting, while the CalComp draving board 
vas used for digitizing. 
DTM of only one grid spacing was considered