status 
if it 
lon to 
a just 
sditing 
art of 
ation), 
-edone. 
r undo, 
>. In 
place 
redraw. 
scribe 
- data 
ice of 
Lt was 
if one 
- data 
-e(main 
, brown 
| area 
brown 
plate, 
parts 
je and 
c, blue 
S are 
>, and 
> each 
simply 
[n run 
1e are 
value 
') and 
quence 
image, 
at the 
re the 
ad and 
int of 
-efore, 
pixels 
ad for 
1coding 
ich of 
an be 
block, 
at the 
sitions 
las an 
lon in 
blocks 
^,:- the 
S cor 
in the 
ais: is 
these 
ita and 
a disk 
> this, 
-ained. 
aborted 
sed for 
or: Or 
of the 
raster 
revised 
map is output to two or three pieces using the non- 
expensive color thermal printer for small format. 
2.5 Film Output and Reproduction 
  
Topographic map data edited by the raster method is 
output on films using laser plotter. In this 
case, one method is to output mask plates and  non- 
mask plates respectively, and to make plates for 
reproduction by photo process. The other method is 
to output plates for reproduction directly after 
combining mask and non-mask plate of same color on 
the computer. The latter is preferable from the 
point of work efficiency, but requires the raser 
plotter with 0.0125 mm output pitch for halftone 
plate. Reproduction process after this is same as 
conventional method. 
3. MERIT OF RASTER BASED MAP REVISION 
When the raster editing method is introduced in the 
routine work, following merits are expected. 
First of all, the instruments required for edit 
change from a light table and scribers to CRT 
screen and a mouse. The efficiency of this method 
is not clear at present because it depends on 
programs. However, interactive editing on EWS 
gives less stress to the operator because different 
from scribing it permits the cancel of edit 
operation. In addition, the conventional method 
allows editing for only one plate simultaneously, 
because photo process intervenes every time after 
revising one plate. On the other hand, in the new 
method all plates except annotation mask can be 
edited simultaneouly, so the process for the map 
revision becomes simple and its control becomes 
easy. 
Second, the raster based map revision causes no 
degradation due to photact process. The raster 
method has a possibility to cause degradation when 
map plates are scanned, but from the experiment so 
far, there is practically no problem if we use such 
map plates as lines on map plates are thick enough 
after photact process. In addition, once map 
plates are scanned, no more scan is required. 
Therefore, the degradation is not accumulated by 
the repeat of revision. Namely, the degradation 
by photact process. cannot be recovered, but does 
not get worse any more. 
Finally, raster data of topographic maps are 
obtained as the by-product of the raster based 
revision. So far, GEOGRAPHICAL SURVEY INSTITUTE 
has been engaged in digitizing maps in vector 
format and has been servicing Digital National Land 
Information and so on. As digital map data in 
raster format are thought to get different demand 
from vector data, we intend to study how to serve 
raster data in future. 
4. CONCLUSION 
Raster based 1/25,000 topographic map revision 
method improves the efficiency of map revision 
while keeping the quality of map, prevents the 
degradation, and produces digital map data in 
raster format. GEOGRAPHICAL SURVEY INSTITUTE in 
Japan regards this method as most promising way of 
map revision, and is endeavoring to make this 
method practical. 
Medium or small scale map is demanded that the 
positions of objects on the map are accurate and 
map representation is easy to see with full use of 
displacement, generalization, break to avoid 
overlap and selection of what to represent. For 
these demand on map compilation, not a vector based 
method which is good at batch processing but a 
raster based method is suitable. In addition, 
different from digital large scale map, raster data 
cannot be used for analysis usually attained in 
vector based GIS, but can be easily displayed on 
the screen as a map easy to see. Moreover, as 
medium or small scale map has been already made in 
many cases, a raster based method has the advantage 
of digitizing map because the digitization is 
easily performed at the time of revision. 
Therefore, a raster based revision method has far 
advantage over a vector based method. Several 
concepts mentioned here are useful when a raster 
based method is applied for the medium scale map 
revision. 
5. REMAINING PROBLEMS 
The development of raster based 1/25,000 
topographic map revision method has not been 
completed yet, and there are still some problems 
remaining. 
First of all, to complete programs for numerical 
plotting and raster editing is most important and 
urgent. Figure 3 shows a screen of raster editing 
program being developed now. In addition, it is 
also urgent to prove this revision method to be 
effective by doing a experiment through all process 
using developed programs. 
Second, the handling of a mouse is not superior to 
a pen. So to use a pen type digitizer instead of a 
mouse should be considered. 
Third, it is necessary to decide how to handle the 
annotation plate in this method. The way we are 
now developing is as follows. The annotaion plate 
is scanned in 0.0125mm pitch, and other plates are 
scanned in 0.025mm pitch. And the annotation plate 
is tranformed into 0.025mm pitch to be handled in 
raster editing program. Moreover, in raster editing 
stage, for each annotation a sequnce of letters, 
position, direction, font type, letter size and 
spacing between letters are input interactively. 
After that, raster data in 0.0125mm pitch is 
produced from these data by a batch processing and 
is merged with raster data in 0.0125mm pitch above 
mentioned. 
Finally, as future plan we think of two ways of 
obtaining changed parts. One is to digitize 
changed parts from existing large scale maps, and 
the other is to scan aerial photographs by a small 
type scanner, to rectify them into orthophotos by 
using DEM, and to get changed parts after 
overlaying orthophotos on raster data of a previous 
version map. As the scale of existing maps used 
for 1/2,500 scale map revision are about ten times 
of the map to be revised, or the scale of aerial 
photographs are about same as that of the map to be 
revised in many cases, it seems that the limit of 
accuracy can be easily cleared for these ways. In 
anycase, study should be done from the aspects of 
both system development and accuracy. 
521