International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
absorption minimum can only be estimated. Lagrangian 
interpolation (Dawson and Curran, 1998) was used to estimate 
the position of the minimum, using the band closest to the true 
minimum and two adjacent bands. 
Absorption feature parameters are then extracted from the 
continuum removed chlorophyll absorption feature. The 
maximum depth, wavelength of maximum depth and the 
FWHM were found to be useful indicators of subsidence 
features. A relationship has been demonstrated by a shallowing 
in depth of the absorption feature of vegetation in areas affected 
by fissuring. Fischer et al. (2003) also noted a decrease in the 
band depth when vegetation showed any changes caused by 
different hydrological conditions. Further work is required to 
further the understanding of the relationship between these 
parameters and known subsidence features. 
5. RESULTS AND DATA INTEGRATION 
The British National Grid is the reference projection used for all 
data processing. This common reference frame is compatible 
with existing GIS data, including topographic and geological 
mapping, and mine plans. This approach allows the results 
from photogrammetric and hyper spectral processing to be 
compared in a GIS environment. A common reference frame is 
also required for subsequent integration of results. 
5.1 Aerial photography 
Photogrammetric processing and surface matching has 
facilitated the creation of a DEM for each epoch of aerial 
photography. The rigorously controlled contemporary surfaces, 
processed from photography acquired in 2002 and 2003, serve 
as a baseline for future change and as a reference surface to 
assess changes in the historical data sets. Using known 
subsidence occurrences within the test areas it has also been 
possible to develop a priori models of the topographic 
expression of subsidence. Consequently, areas affected by 
subsidence may be identified using aerial photography either as 
a difference in surface elevation between survey dates or 
directly, where the area has a topographic expression which is 
characteristic of subsidence. 
5.2 Hyper spectral imagery 
A relationship has been established between thermal anomalies 
and some subsidence features. However, these anomalies may 
be masked by vegetation cover and are very dependent on the 
weather conditions prior to image acquisition. 
Spectral parameters, describing the red edge and the 
chlorophyll absorption feature (Figure 5) have proven to be 
useful indicators of the presence of subsidence features. 
Further work is required to increase understanding of the affects 
of seasonal changes in vegetation on these parameters. This 
will facilitate change detection and future monitoring of 
subsidence using hyper spectral imagery. 
  
Figure 5. CASI-2 image of chlorophyll absorption continuum 
removed depth for part of the Houghton-le-Spring 
site. The overlay shows the position of fissuring 
mapped during field work. 
5.3 Integrated analysis 
Research is now exploring the use of multi criteria analysis and 
other integration approaches, to combine the results from the 
aerial photography and hyper spectral imagery with other 
relevant data sets. Coal Authority mine abandonment plans 
show the location of pillar and stall workings and can be used to 
identify areas of possible risk which can then be investigated 
using the airborne imagery and techniques outlined above. 
Simple decision rules can be applied to remove spurious 
anomalies and apply a degree of confidence to anomalies. For 
example, red edge and continuum parameters are meaningless if 
no vegetation is present, and photogrammetric measurements 
are unreliable in areas of crops awaiting harvest. Further work 
is required to develop a fully integrated system for subsidence 
monitoring and mapping. 
6. CONCLUSION 
The legacy of mining subsidence in the North East of England 
is difficult to quantify and has economic implications for the 
development of the area. This ongoing research has two aims. 
The first aim was to develop methodologies for the extraction 
of relevant information from aerial photography and hyper 
spectral imagery for subsidence hazard mapping. This has been 
achieved by characterising the topographic and spectral 
response of subsidence features in the study area, and 
developing existing processing techniques to enhance these 
characteristics to enable areas of subsidence to be identified 
above abandoned coal mines. The second aim was to develop a 
methodology for integrating results from the two disciplines 
with ancillary data for subsidence hazard mapping. Multi 
temporal airborne imagery has considerable potential for 
subsidence hazard mapping and there is considerable potential 
in integrating photogrammetric and hyper spectral data sets. 
This research is ongoing and further results will be reported at 
the ISPRS Congress presentation of this work. 
REFERENCES 
Banninger., C., 1991, Phenological changes in the red edge 
shift of Norway spruce needles and their relationship to needle 
chlorophyll content. In: Proceedings of the 5" International 
Colloquium — Physical Measurements and Signatures in 
718 
  
Interi 
  
Remo 
Noor: 
Bell, 
to al 
Engir 
Bell, 
subsk 
exam 
Clark 
Quant 
Appli 
6329- 
Cleve 
Meer, 
and th 
pp. 31 
Daws« 
interpi 
Journ 
Faig, 
an exe 
Congr 
Fische 
Detect 
imagir 
Spectr 
Goulty 
coalfie 
Faulte 
Gray, 
abandc 
longw: 
Littletc 
Guyot, 
spectra 
Procee 
Physic 
ESA-S 
LaScol 
Survey 
US Bui 
Miller, 
charact 
inverte 
Remote 
Mills, J 
Fusion 
Models 
(4), pp. 
Pickeril 
rural a 
Interna 
2433.