The difference in scale between x and y image 
coordinates is clearly not as large as in the case cited 
earlier. This suggests that the value varies significantly 
between various PC based frame grabbing systems and 
so a means of estimating differential scale factor is 
clearly important. The estimated value is close too 
unity and the difference in estimated staff height with 
and without this additional parameter is not large, 
particularly with single frame estimations, (A/C). 
Indeed the estimated staff height is extraordinary close 
to the known value of 1.600m, although reference 
should be made to the low precision (+/-15mm) of 
these particular estimates. The most accurate and 
reliable estimate is represented by Option D in which 
all six frames are used and differential scale is 
included. The estimated plan position of the staff also 
agrees closely with known position represented by 
Option E and the a posteriori variance factor is also 
below unity suggesting that a suitable functional model 
has been selected. 
5. OTHER CONSIDERATIONS 
The relevance of a photogrammetric estimate of a 
persons height can perhaps be questioned, particularly 
for legal purposes. A comparison between an estimated 
height from an incidental photograph and a physical 
measurement of a person standing erect, is not 
necessarily valid. Posture, stance and footwear all 
effect the height of an individual. The linear effects of 
these sources of systematic error surely transcend the 
boundaries of photogrammetric expertise into the 
medical world. The additional expert witnesses’ who 
would become involved with producing a result will 
inevitably confuse a jury already baffled by the science 
of photogrammetry. If the experts have not achieved 
this confusion with their own explanations then 
opposing counsel will gratefully assist and reduce the 
significance of the evidence. 
It should be stated also that the human frame is not a 
fixed and linear object and a persons height is not a 
definitive parameter which can be used for 
identification. This type of evidence is perhaps best 
suited for defence purposes in which it may help to 
exclude an individual. The technique can not be used 
to prove that a particular person committed an alleged 
crime, even if the individual was exceptionally short or 
tall. 
6. CONCLUSION 
It has been demonstrated that with care 
photogrammetrists can determine geometric data from 
single camera security imagery. Security systems based 
upon photographic systems would appear more 
preferable than their video counterparts. 
The value of strong network geometry should not be 
underestimated because such strength leads to an 
estimate which is both internally and externally 
reliable. Reliability can help identify a suitable 
functional model and will ultimately lead to both a 
precise and, perhaps more importantly, an accurate 
result. When network geometry is weak, extreme care 
is necessary at all stages of processing and analysis. 
REFERENCES 
BROWN, D.C. (1976). The bundle adjustment- 
progress and prospects. International Archives of 
Photogrammetry, 21(3) Paper 303:33pp. 
CHANDLER, J.H. & COOPER, M.AR., (1988). 
Monitoring the development of landslides using 
archival photography and analytical photogrammetry. 
Land and Minerals Surveying, 6:576-584. 
EL-HAKIM, S.F., BURNER, A.W. & REAL, RR. 
(1989). Video technology and real-time 
photogrammetry. In: Non-topographic photogrammetry, 
ed. by HM. Karara, (Falls Church:American Society 
of Photogrammetry and Remote Sensing), pp. 279-304. 
GRANSHAW, S.I., (1980). Bundle adjustment 
methods in engineering photogrammetry. 
Photogrammetric Record, 10(56):181-207. 
HOTTIER, Ph. (1976). Accuracy of close range 
analytical solutions. Photogrammetric Engineering and 
Remote Sensing, 42(3):345-375. 
LINNEY, A. & COOMBES, A.M. (1990). Report on 
comparison of photographs and video images. RE: 
Regina v McVicar. 4pp. 
MCGLORE, J.C., (1989). Analytical data reduction 
schemes in non-topographic photogrammetry. In: Non- 
topographic photogrammetry, ed. by H.M. Karara, 
(Falls Church: American Society of Photogrammetry 
and Remote Sensing), pp. 37-57. 
   
   
oa iac I, iN a TL ao 
ub je a N jaa 5 Foy uh Bud Cli