Page 170 - Green - Maritime Archaeology: A Technical Handbook. 2nd ed

Chapter 4: Conventional Survey 149
s=
 (x - x 2 ) n
2. Survey Rationale
The survey required a system that allowed measurements to be made on land and under water and that the results could be easily compared. The photo tower was selected for the project because it was rigid, easily col- lapsed for transportation, and had calibrated targets on the surface of the framework. As different systems would be used both on land and under water, it was also necessary that the system could be transported from the terrestrial to the marine environment without undue difficulty. Because the targets on the photo tower were suitable for measurement using the different systems, it was then relatively easy to compare these data. The tower was also appropriate because it had considerable vertical elevation so that an assessment could be made of both the horizontal and vertical measurements and the relating accuracy. Bathers Bay was selected for the underwater work as it was close to the Department of Maritime Archaeol- ogy (about 500 m), reasonably sheltered, and with a hard sandy bottom that gave good visibility while divers were working on the survey.
In the tape-measuring system (trilateration), there are two basic methods of conducting a survey: either by measuring the interpoint distances or by using control points and making measurements from the control points to points of interest. In the interpoint measuring system, where the measure- ments are made without control points, there is the problem that, as the number of points increases, the number of measurements rapidly become unmanageable (Table I illustrates the problem). Duncan in Atkinson et al. (1988) suggested that in these types of calculation (least-squares) it is more useful to measure the absolute minimum necessary to define the station coordinates. If only the minimum number of distances have been measured, the procedure will still be able to calculate the coordinates. The minimum number of distances required to define the coordinates of k stations is 2k— three for a two-dimensional fix. However, for a three-dimensional fix Holt (personal communication)suggested that it is 3k, ignoring redundancy. Thus it is possible to reduce the number of measurements considerably. In the second system control points are used. This system is commonly used in archaeological survey work, particularly excavations, where there are usually a very large number of points of interest that become uncovered and therefore need to be surveyed. Because the control points remain in place during the survey and generally are rigid points outside the excava- tion, each position is located by a series of measurements equal to the