International Archives of the Photogrammetry, Remote Sensing and Spatial Information Science 
  
  
  
survey has ever been undertaken of the motte. It was 
decided, therefore, to select the tower as the final test site 
for the project where a survey would be undertaken to 
mirror the commercial survey process. 
3.1 The project brief 
The project brief was based on the existing metric survey 
specification and the addendum produced in this project. 
It outlined the extents of the required survey, the 
necessary point density, the control system to be used and 
the deliverables required. 
3.1.1 Point density and measurement precision: In 
laser scanning point density and measurement precision 
are directly linked. Without high density point 
measurement features cannot be accurately described. 
Without accurate measurement they cannot be accurately 
located. The addendum provides two methods for 
defining an appropriate point density. The first is using a 
standard table to suggest an effective point density and 
precision for a particular scale of survey. This method of 
referencing the required survey precision to the scale of 
the required output reflects the method used for 
photogrammetric survey, where required precision is 
based on the line width of the final product. Whilst this 
method is not directly suited to laser scanning, it was 
included to preserve parity with the existing document. 
The second method allows point density to be defined 
based on a minimum feature size. For example, the brief 
may state that the minimum feature discernable in the 
point cloud should be 10 mm in depth (in the direct of the 
measurement) and 10 mm in width and height 
(perpendicular to the plane of the scan). 
In the case of Clifford’s Tower, three major areas were 
defined for survey. Firstly a topographic survey of the 
motte was required at a scale of 1:100 which based on the 
standard table required a point density of 25 mm (with 
point measurements with a lo precision of +/- 25 mm). 
Secondly the exterior of the tower itself was required 
showing features of 10 mm in size or larger. Finally, 
detailed scanning of a window on the South-West lobe 
was required, showing features of 4 mm or larger. 
The addendum's user guide contains an equation that 
allows the appropriateness of the chosen point density to 
be assessed: 
(i 
S 
where Q is the quality of the data, m is the point density 
on the object and s is the minimum feature size. Using 
this formula a value can be obtained to indicate the 
likelihood of a particular feature being detectable. For 
example, a point density of 5 mm on a feature 10 mm in 
size would give a value of 0.5, or a 50% confidence that 
the feature would be visible. This equation leads to very 
high point densities being required in order that the 
features, typically surveyed using photogrammetry, can 
be detected in scan data. Although possible using close 
range scanning, terrestrial laser scanning is clearly 
unsuited to recording very small features and it serves to 
further emphasise the fundamental role of laser scanning 
is in surface measurement. 
968 
3.1.2 Control: The project brief also outlined the 
required control system. In the case of Clifford's Tower, 
as with most of the survey commissioned by English 
Heritage, all scan data was to be registered to the local 
site grid. A pre-established network of control points in 
the site coordinate system around the base of the tower 
was provided by English Heritage. 
The addendum sets out clear requirements for the control 
of scan data using resection calculations: 
° sach scan must contain at least four appropriately 
distributed XYZ control points (these should be 
measured using techniques that allow an accuracy of 
twice the value of the accuracy of a single point) 
e the residuals of the registration process (the process 
of transforming the raw scan data to the site system) 
and the geometric precision of the estimated 
parameters should be noted in the survey report. 
These should be shown to be equal to, or better than 
the accuracy required by the final point cloud. 
The addendum also allows for registration to be 
performed using surface matching techniques, but as all 
survey data of sites and monuments has to be registered 
to a local site grid targeted control points are still 
required. A total of n+3 appropriately distributed targets 
are therefore specified when using surface matching 
techniques, where n equals the number of scans in the 
survey. When scanning is performed from a known 
position, for example when an instrument is located 
above a control point, only 3 XYZ targets are required 
per scan. 
The specification requires that artificial targets are used 
and that they must be carefully placed to avoid obscuring 
important details or being close to depth 
discontinuities. The use of natural control points is not 
recommended by the addendum except where 
unavoidable, and then only points away from edges may 
be used. 
large 
3.1.3 Data voids: The addendum requires that areas 
where data voids are anticipated to be unavoidable should 
be highlighted in the method statement before survey is 
undertaken. 
3.1.4 Supplementary data: The project brief also 
defines the data required in addition to the scan data. In 
this case narrative (i.e. non-metric) photography was 
required at each scanner station to aid interpretation, in 
addition to non-metric photography of relevant features, 
such as mouldings, carvings and existing damage to the 
Tower's fabric. 
3.2 Pre-survey deliverables and certification 
requirements 
In addition to the standard information outlined in the 
current metric survey specification, a description of the 
proposed point density and potential data voids, the 
technical details of the proposed scanning system are also 
required. 
s, Vol XXXV, Part B7. Istanbul 2004