998
photographs. The figure was calculated from a test
area in one sub-district.
6.9 Training of interpreters.
To carry out such a large and complex survey
requires a large team of airphoto interpreters.
Only two members of NHA staff had been trained in
this field at the ITC Urban Survey. For efficient
photo interpretation, semi-professional staff was
required. They were recruited from newly graduated
or 4th year Geography students as they had some
previous training in this area. NHA, through CHHSS,
organized a five day exercise with the aim to help a
team of interpreters to achieve a good standard of
recognizing, identifying and classifying. They were
also trained in technique of housing count, housing
unit estimate, area measurement and data recording.
7 AIR PHOTO INTERPRETATION PROCEDURES.
For comparing sequential aerial photographs and
making a complete interpretation of each year, it is
preferable to start with the most recent one. Among
24 districts of BMA, each student was responsible
for the interpretation of one or more districts
depending on their local knowledge. Procedure of
interpretation is described below. Upon completion
of the interpretation of the enlarged SFAP 1984, the
procedure was repeated for the airphotos 1974 for
comparison of information as required by terms of
reference. Six interpreters were required for the
BMR study for six months. Interpretation cost was
Bht 90,000 (US$ 3,300).
Steps of procedure.
Seq. Description of work Remarks
1. Preparation of
district boundarymap
of BMA (24 district)
2. Construction of un
controlled mosaic of
enlarged SFAP 1984
scale 1:15,000 for
each district
boundary
3. Airphotos were
scanned stereosco-
pically
4. Interpretation of
residential area by
type of project (by
delineation on en
larged SFAP)
5. Overlay transpar
ency on photocopy
uncontrolled mosaic.
6. Transfer delineation
from enlarged SFAP
to the overlay of
each district
7. Cross-check (sampling)
8. Housing count of
each polygon by type
of housing, counting
was done on enlarged
SFAP, and fieldcheck
for multi-storey
buildings.
9. Area measurement
Based on BMA administra
tion map scale 1:50,000
From photocopy of en
larged SFAP
To get an overiew
picture of the area
Interpretation was done
monoscopically, in case
of doubtful area,
stereoscopic viewing
was used for checking.
Boundary of district
and main spatial fea
tures were marked.
Locational distribut
ion of the residential
areas was recorded on
overlay by making
annotation of differ
ent classes in
colours.
By supervisors
Data of each area unit
were recorded in each
polygon on the overlay
in codes, i.e. d=50,
where ’ d* refers to de
tached house of 50 units,
*sd’= semi-detached
house, *r’= rowhouse and
so on.
For each polygon by dot
grid. (1 dot=0.5625 ha)
10. Estimating housing
density for informal
settlement
11. Calculate the number
of houses for each
type of settlement
12. Transferred data into
table form for
analysis
Making a one hectare
block and counting
number of houses in
block at random by type
of settlements.
Area x density
Data included: resident
ial area, type of
housing project of 1974,
1984 and number of
houses by type of pro
ject and type of build
ing.
8 RESULTS
Data from photo interpretation revealed that
during the decade from 1974 to 1984, the number of
dwellings in the BMA increased dramatically both in
termsof quality and quantity as shown in table 1.
Thus, the supply of housing increased at an annual
rate of about 6.4 percent which is virtually identi
cal to the rate of increase of households (census
data). This is primarily attributable to the rapid
expansion of private residential development (map
no.1).
The supply of housing units, by private housing
projects increased from 4.2 to 14.9 percent of the
housing stock in decade ending in 1984. The increase
in this category corresponded roughly to the
decrease in the small scale development tabulated in
the individual category. This may indicate an in
creasing formalization of the housing market, an
impression is also enhanced by the increase in the
market share of public housing from 2.1 percent to
9.1 percent. The low income informal sector includ
ing Klong houses and slum and squatter settlements,
declined from 28.6 percent of the stock to 21.1
percent. Map no. 2 indicates the spatial distribu
tion of the slum areas in Bangkok. The composition
of the housing stock by type of building is also
changing in significant ways. Table 2 shows an
increasing tendency towards the relatively high
density forms, especially apartments, at the expense
of detached house units.
The airphotos used for the housing count were
further applied in the analysis of trends concerning
the physical development of the BMR area up to 2001
and for the selection of potential housing areas for
the 6th plan. Results of this analysis are shown in
map no.3 taken from the final draft report on the
BMR study (NESDB, 1986)
CONCLUSION
The result of the Bangkok survey described in this
paper can be considered as a particular example
among similar studies for the following reasons:
1. Large area (5,000 sq.km) covered by vertical
Small Format Aerial Photography despite SFAP
being generally recommended for small areas.
2. Flight mission was carried out by a professional
aerial photographer and navigator, previous
experiments were generally done by planners or
amateur photographers.
3. The study was undertaken for real work, so the
result of the study has been incorporated in the
6th 5-Year Plan.
For this reason, relatively high accuracy of data
was expected.
4. In principle every house was counted either from
airphoto or a map except only in unplanned areas
where housi:
previous st
only done i;
teristic of
5. Classificat
designed fo
tion was do:
months to c
accounted f
larged SFAP
6. Enlarged SF
of sufficie
houses, fur
in the info
Although the
on the applica
photographs ha
local institut
like map updat
tion and infra
No data acqu
cheap but SFAP
is an approach
used elsewhere
are a primary
future where h
delivered time
tions, SFAP wi
detailed infon
est where largi
