Aim 1:
Determining the Time Delay Between “Synchronised” Camera Shutters

The following figure shows the weighted average synchronisation error of each camera, with respect to the reference camera (Camera 2). It is interesting to note that although the cameras were connected with different length cables to the remote, this does not correspond to the differences in delay. For example, Camera 2 was on a medium-length cable, yet was consistently the first camera to record an image.


Graph showing the weighted average synchronisation error of each camera.

One theory which may explain this, is that the internals of each camera are slightly different; that is, depending on current processing capabilities, there may be a delay between receiving and processing the signal to take a photograph, which is slightly different in each camera.

The average synchronisation error was 19.6 milliseconds; which is comparable to the 16.7 milliseconds theoretically incorporated by using two 60-frames-per-second video cameras filming from different angles (Ebrahim, 2010). However the minimum recorded error of 0.4 milliseconds, between cameras 5 and 6, was substantially less than this.

Hence, the average synchronisation error is not the most important consideration when it comes to close range photogrammetry of moving targets. Of more importance is determining the exact synchronisation error between pairs or sets of cameras, so that the delay can be incorporated into the calculation of the results.

.

Ebrahim M. A-B. (2010) “Influence of video frame mis-selection on the accuracy of moving 3D measurements using close-range photogrammetry.” In: Proceedings of the ASPRS 2010 Annual Conference, San Diego, California, 26-30 April.

mouseover