Aim 4:
Investigating the Impact of Various Target Speeds on the Positional Accuracy of Close Range Photogrammetry

Due to the limitations of PhotoModeler’s internal software, it was not possible to capture targets travelling at speeds above 10km/h, so a range of speeds between 0.3 and 6.2 km/h (10 – 170 m/s) were used. Within this range, no significant trend was determined between the target speed and the circular fit error (a measure of the software’s target recognition), or between the target speed and the residual.

The absence of this particular result does not indicate that such a trend does not exist; by either manipulating existing software or creating new programs, it may be possible to image and reference faster-moving targets, and to hence extract trends over a larger velocity rage.

It is also worth noting that although no trend was determined between the target speed and the residual, the average residual of the moving targets was consistently higher than that of the static targets, as can be seen in the figure below. This indicates that there may be a trend between target speed and residual, however due to the constraints of this Thesis Project, it is not possible to determine what type of relationship this may be.


Average Target Residuals (Moving v. Static). The higher average for the moving targets indicates a trend between target speed and residuals may exist.

Target Sizes

Another result worth mentioning, unrelated to the four specific Aims, is that the circular fit errors increased as the target size decreased. As can be seen in the figure below, the average circular fit error of the small targets (left) were consistently larger than the medium and large targets (middle and right, respectively). This is true across variations in both shutter and target speeds.

This result is consistent with the recommendations by PhotoModeler on which target sizes produce optimal results for the particular experiment design. Particular elements of the design used in PhotoModeler’s calculations include the average distance from camera to target, as well as the scale of the objects in the cameras’ fields of view.


Average Circular Fit Error of Targets increases as Target Size decreases.

mouseover