Page 172 - Green - Maritime Archaeology: A Technical Handbook. 2nd ed

Chapter 4: Conventional Survey 151
2. Random errors usually caused by differing tension on the tape, cur- rents, lack of precision in measurement, etc.
3. Systematic errors usually caused by a stretched tape.
Thus, one can generalize and say that when a measurement is clearly wrong (a gross error) there will be a large residual and this reading can be ignored. Where the residual is small, it is likely that that the error is either random or a transcription error. It is generally not possible to tell the dif- ference between these two causes and, therefore, it is not possible to com- pensate for them. The result is, therefore, a poor fix.
In the case of the photogrammetric method, errors are more difficult to assess. The two main errors are the camera lens calibration and the point positioning accuracy. It was generally found that there were larger resid- uals in measurements that were close to the edge of the photographic image. This suggested that the calibration is less rigorous at the extremities of the image and that in some cases it was difficult to observe targets from particular views, thus reducing the accuracy of the fix.
With the HPASS the question of accuracy have been discussed in Section VI above and by Green and Souter (2002). It was again found that the depth measurements created problems that required the application of a tidal correction.
The comparison of accuracy is always a complex issue, particularly where there are two different measuring systems. The procedure adopted was to indicate in the distance-measuring system what the average residuals were for the measurements and in the optical system what the average tightness was for the points (points are computed by intersection of light rays and due to measurement errors these rays do not necessarily intersect, thus the tightness is a measure of the largest distance). In addition, because there were various methods used and it is felt that the land-based tape measure- ment will give the best results, it has been assumed that these measurements were the most accurate. Therefore the computed X, Y, and Z coordinates of the various methods have been compared with the land-based results to give an idea of the relative accuracy.
3. Site Surveyor on Land
The tower was set up in the car park of the Maritime Museum (see Figure 4.27). Two teams then measured the intertarget distances using a predrawn measuring matrix or pro forma.This matrix was used both on land and under water and provided a system which ensured that all necessary measurements were recorded. A total of 162 measurements were made between the 18 targets, and this took a total time of 96 minutes. Starting at the first target