194
Newsham et al.
The photogrammetry was undertaken using an Intergraph ImageStation, a dualfunction image analysis and digital photogrammetric workstation. It is estimated that
individual DTM points may have an RMSE height accuracy better than ± 0.20 m. A
ground-truth survey using a Nikon Total Station is currently being undertaken and should
provide more quantitative information about co-ordinate accuracy. Lines representing the
cliff-top and bottom were digitised manually into the file from the image. These
features are placed into a Microstation file. The x,y,z position of the features are loaded
into Terrain Analyst to create a digital terrain model of the area. As each different
feature is loaded into Terrain Analyst it is assigned an attribute relating to the type of
feature it is. For example, the lines representing the cliff-top and base are loaded as a
breakline; this prevents the model from rounding off the cliff-top and base. Further
details of the photogrammetric techniques used can be found in Balson et al. (1996).
3D DTM - from bathymetry
Water depth data was used to construct a DTM for the nearshore area. Data was
captured for an area of sea floor out to approximately 2.5 kilometres from the low
water mark where a prominent break in slope is assumed to represent the depth beyond
which shoreface erosion becomes insignificant. Spot heights were loaded into
Microstation and the elevations were corrected to Ordnance Datum to enable the final
model to be merged with the onshore DTM to create a single DTM of the eroding cliffs
and shoreface (Figure 2).
Cliff Geology
Horizontal sections which show the lateral and vertical distribution of the geological
formations exposed in the Holderness cliffs were drawn by Bisat (Catt and Madgett,
1981). The original hand-drawn sections, housed in the archives at BGS, were
digitised using Microstation. The section was rescaled in segments, between points
with known Eastings and Northings, to reduce the effect of horizontal scale errors in
the original drafting. Each geological formation was captured as an individual
polygon. The vertical axis was converted to heights relative to Ordnance Datum. The
digitised section had a vertical exaggeration of x10, to assist in the visualisation of the
polygons. Each polygon was marked with a code to identify the specific geological
formation. Bisat’s sections show the formations exposed only above beach level. For
the area below beach level and on the shoreface the geology must be assumed. For
simplicity it has been assumed that only Basement Till is exposed below beach level. In
the future it is hoped that further study will improve knowledge of the distribution of
Quaternary formations of the Holderness Coast.
To assess the accuracy of the cliff-top topography in the original horizontal
section produced by Bisat, a line was placed in Microstation to follow the current clifftop on the DTM. Terrain Modeler created a profile across the model and placed it
graphically into the Microstation file. The original profile was then overlain and
registered using known points on each profile. Slight discrepancies between the
original profile and the profile generated from the DTM are probably due to the
Newsham et al.
The photogrammetry was undertaken using an Intergraph ImageStation, a dualfunction image analysis and digital photogrammetric workstation. It is estimated that
individual DTM points may have an RMSE height accuracy better than ± 0.20 m. A
ground-truth survey using a Nikon Total Station is currently being undertaken and should
provide more quantitative information about co-ordinate accuracy. Lines representing the
cliff-top and bottom were digitised manually into the file from the image. These
features are placed into a Microstation file. The x,y,z position of the features are loaded
into Terrain Analyst to create a digital terrain model of the area. As each different
feature is loaded into Terrain Analyst it is assigned an attribute relating to the type of
feature it is. For example, the lines representing the cliff-top and base are loaded as a
breakline; this prevents the model from rounding off the cliff-top and base. Further
details of the photogrammetric techniques used can be found in Balson et al. (1996).
3D DTM - from bathymetry
Water depth data was used to construct a DTM for the nearshore area. Data was
captured for an area of sea floor out to approximately 2.5 kilometres from the low
water mark where a prominent break in slope is assumed to represent the depth beyond
which shoreface erosion becomes insignificant. Spot heights were loaded into
Microstation and the elevations were corrected to Ordnance Datum to enable the final
model to be merged with the onshore DTM to create a single DTM of the eroding cliffs
and shoreface (Figure 2).
Cliff Geology
Horizontal sections which show the lateral and vertical distribution of the geological
formations exposed in the Holderness cliffs were drawn by Bisat (Catt and Madgett,
1981). The original hand-drawn sections, housed in the archives at BGS, were
digitised using Microstation. The section was rescaled in segments, between points
with known Eastings and Northings, to reduce the effect of horizontal scale errors in
the original drafting. Each geological formation was captured as an individual
polygon. The vertical axis was converted to heights relative to Ordnance Datum. The
digitised section had a vertical exaggeration of x10, to assist in the visualisation of the
polygons. Each polygon was marked with a code to identify the specific geological
formation. Bisat’s sections show the formations exposed only above beach level. For
the area below beach level and on the shoreface the geology must be assumed. For
simplicity it has been assumed that only Basement Till is exposed below beach level. In
the future it is hoped that further study will improve knowledge of the distribution of
Quaternary formations of the Holderness Coast.
To assess the accuracy of the cliff-top topography in the original horizontal
section produced by Bisat, a line was placed in Microstation to follow the current clifftop on the DTM. Terrain Modeler created a profile across the model and placed it
graphically into the Microstation file. The original profile was then overlain and
registered using known points on each profile. Slight discrepancies between the
original profile and the profile generated from the DTM are probably due to the