Page 176 - Green - Maritime Archaeology: A Technical Handbook. 2nd ed

Chapter 4: Conventional Survey 155
The quality of the measurements in Table II; first the arithmetic mean for the X, Y, and Z coordinates, the standard deviations errors for the X, Y, and Z coordinates, and the standard deviation of all the measurements.
B. CONCLUSIONS
The results of this investigation are very interesting, although it must be remembered that the investigation was based on a small site with a large vertical component. It is clear that the following conclusions will not nec- essarily apply to very large or a very small sites. Additionally, the visibility and sea conditions were good, visibility was about 10 m, and the water was calm and a temperate 20°C, thus making the recording easy and reliable. Therefore, in this situation the general conclusions for underwater work are that PhotoModeler is at least as accurate as underwater hand measurement, but considerably faster, thus saving in time and making the recording much more efficient (an aspect, previously noted by Green et al., 2002), which is particularly relevant for deep-water sites where time is at a premium (see Table III for the comparison of efficiency). In this situation the underwater hand measurements (interpoint) took 86 minutes, compared with Photo- Modeler where the underwater image capture took 10 minutes. Obviously PhotoModeler only works in good visibility and at relatively short ranges (in Turkey, where the water was extremely clear, PhotoModeler was used to survey distances around 20–40 m, (Green et al., 2002), however, it must
Survey type
Interpoint hand measurement land
Interpoint hand measurement under water
Control point hand measurement under water
HPASS
PhotoModeler land PhotoModeler under water
(minutes) 96 0
Table III
Accuracy and Performance Time for the Different Systems.
Approximate
accuracy (mm) 0
20 37
53 14 18
Time to survey (minutes)
Set up
survey
Processing Total
(minutes) (minutes) 30 126
30 116 30 118
20 70 25 35 25 35
86 0
100 10
20 30 10 0 10 0