Aim 2:
Determining the Effect of Synchronisation Errors on Positional Accuracy

It is important to note that for commercial software such as Photomodeler, high target speeds cannot be recognised by the software unless an extremely fast shutter speed is also used, as otherwise the target appears “blurred” in the image. For example, in images taken outdoors with a shutter speed of between 1/60 and 1/800, it was difficult for Photomodeler to detect targets travelling at faster than 10km/hr. However, the theoretical positional errors for targets speeds of up to 15km/h can be shown in the figure below as there is a linear relationship between the target speed, and its relative displacement over the period of the synchronisation delay.


Graph showing the calculated target movement during the synchronisation errors, at different target speeds.

In projects suitable for commercially available photogrammetric analysis (i.e. less than 10km/h), the average synchronisation error introduces a positional error of as much as 6.5mm, as can be seen in the graph. However, by using cameras that are very closely synchronised (in this case, Cameras 5 & 6, represented by the bottom line in the graph), the error introduced by the synchronisation delay is only 0.16mm at 10km/h (approx. 2.8m/s).

Hence, the synchronisation errors can have a significant impact on the positional accuracy of the measured target. Using closely synchronised cameras and a fast shutter speed, it would theoretically be possible to record targets moving at up to 60km/h while keeping the positional displacement under 1mm.

In essence, the importance of synchronisation delay to an individual project depends on the required accuracy; care should be taken in selecting the set of cameras when higher accuracies are required, particularly in terms of how closely synchronised the cameras are.

mouseover