Table 2. Number and size 
of sites in the 
Table 4. 1980 classification scheme. 
detailed 
Dudley area. 
Addition; 
1980 Dudley Derelict Land 
ability 
1971 Dudley Derelict 
detectic 
sites 
hectare 
% 
mean 
areas/ba: 
hectares. 
no. sites. 
(1983) . 
spoil heaps 
11 
13.2 
2.9 
1.2 
Increc 
0-0.5 
2 
excavations/pits 
15 
39.3 
8.7 
2.6 
do not 
0.51-1 
18 
military 
0 
0 
0 
0 
provides 
1.01-1.5 
19 
disused rail land 
4 
6.6 
1.4 
1.6 
monitorii 
1.51-2 
14 
disused sewage works 1 
0.8 
0.2 
0.8 
cover mu 
2.01-3 
17 
disused waterways 
3 
5.6 
1.2 
3.1 
state the 
3.01-4 
9 
neglected land 
85 
260.6 
57.6 
3.1 
cause cl< 
4.01-5 
6 
other 
2 
3.9 
0.9 
2 
from mat 
5+ 
32 
than res 
ground truth 
64 
122.5 
27.1 
1.9 
Baumann 
total 117 
accuracy. 
total 
185 
425.2 
100 
and Forst 
1980 Dudley Derelict 
the prob 
It is apparent 
that the air survey 
methods 
urban la 
hectares. 
no. sites 
0-0.5 
0.51-1 
1.01-1. 
51-2 
01-3 
01-4 
01-5 
1, 
-2, 
3, 
4 , 
5+ 
42 
47 
27 
11 
21 
12 
5 
20 
total 185 
The number of sites in the 1-5 ha range 
seems to have remained reasonably stable, 
with a reduction of 11 sites over the period 
being evident. However, with reference to 
sites over 5 ha a significant reduction in 
number has occurred from 32 to 20 sites. 
Larger sites were generally of the neglected 
waste land category, the type of land that is 
relatively easy to reclaim. It appears that 
much of the land has been returned to private 
development and recreational use. The move 
towards these uses seems reasonable 
considering the Countryside Review Committee 
(1976) statement concerning reclaimed land 
that 'the original agricultural quality is 
hardly ever recovered'. 
Tables 3 and 4 reveal the classification 
results for 1971 and 1980, these are complete 
for the 1971 survey. The absence of 16 air 
photographs means that 64 sites on the 1980 
map have not been identified with regard to 
type of dereliction, these are indicated as 
'ground truth' in the classification and the 
aerial data has been abstracted from the WMCC 
1980 ground survey. The author is currently 
performing categorization of these sites, 
therefore, at the present time, it would be 
fatuous to attempt any comparison between the 
data sets and they are presented merely as a 
guide. 
Table 3. 1971 classification scheme. 
19.7.1 D.udl£yL..-D.£xelict Land 
sites hectare 
spoil heaps 
excavations/pits 
military 
disused rail land 
disused sewage works 
disused waterways 
neglected land 
other 
11 
14 
0 
6 
0 
4 
82 
0 
29.4 
137.4 
0 
14.7 
0 
8 
382.1 
0 
5. 
24 
0 
2, 
0 
1, 
66, 
0 
outlined in this paper are efficient and 
accurate, with the study results reflecting 
county and district changes. It has been 
demonstrated by Collins and Gibson (1980) 
that aerial surveys have substantial 
advantages over traditional ground based 
surveys, which take up to 12 times as long, 
cost 4.5-8.5 times as much and do not locate 
as many sites. 
Satellite monitoring in urban areas 
total 
117 
571.6 100 
The previous section has demonstrated the 
accuracy and benefits of using aerial 
photography in the assessment of specific 
urban land covers. One of the main problems 
with such surveys is that councils obtain 
coverage on an aperiodic basis only. In 
considering reclamation, Bullard (1983) 
explains the philosophy behind this, 'The 
cost of monitoring may be a small percentage 
of the total cost of reclamation but like any 
'service' they are only fully appreciated 
when they can show significant changes taking 
place in one or more of the conditions'. It 
is unlikely that the mentality of local 
authorities will be changed and therefore 
another source of monitoring data needs to be 
utilized. Satellite imagery is readily 
available and being multitemporal is suited 
to the needs of local authorities. The use 
of such data is well documented for 
agricultural and other land uses, however 
interest is increasing in using high spatial 
and spectral resolution systems, such as SPOT 
1 and Landsat 5 TM in urban studies. 
Necessarily there are problems with the new 
data types and what follows is a brief 
evaluation of their suitability for 
qualitative and quantitative urban 
monitoring. 
Landsat MSS imagery has been used in the 
urban environment, but only with limited 
success. Wang (1984) notes that studies in 
the urban environment concentrate on those 
areas requiring less resolution, such as 
boundary changes, and that few have realised 
the harder task of identifying urban land use 
categories. 
The TM appears to have advantages over MSS 
data in terms of data quality that may figure 
significantly in urban area analysis. Anuta 
(1984) reveals that using the same clustering 
and merging sequences, TM data exhibits 42 
classes as opposed to 21 from MSS. 
Additional evidence of the increased 
dimensionality in TM opposed to MSS is 
recognised when Principal Component (PC) 
images are examined. Quattrochi (1982 and 
1985) has recognised that using PC images in 
ratio and photographic forms facilitates 
take the 
pixel diJ 
ref lect ii 
surround 
readings. 
such cor 
(1979) v 
condition 
as sharp] 
land use' 
Bounda 
the urbai 
infrequer 
boundary 
effects. 
that up 1 
scenes ar 
that wit] 
do not lc 
being c: 
int roduc 
specific 
effect, < 
in a nei< 
responsit 
A f i 
discrimir 
regions, 
(PSF) . A 
seriousl; 
classes 
Accordin 
represen 
deconvol 
estimate, 
' spatial 
improper] 
order ti 
sampling 
The views 
of sense 
improved 
suggests 
increasi 
accuracy 
variabil: 
and b) a 
24%) enh 
diminish 
should 
classif ii 
(1979) i: 
Simulati 
Clark a] 
decrease 
increase' 
nature ol 
Final 
higher r< 
1. Reduce 
2. Aid cc