X hits on this document

PDF document

A Prototype Optical Tracking System Investigation and Development - page 47 / 170

393 views

0 shares

0 downloads

0 comments

47 / 170

3.5 Summary

33

.

0.5m

3m

Figure 3.8 The maximum tracking volume (shaded) surrounding one beacon (dot) is determined by the minimum and maximum distances that the camera modules can observe markers, assuming the hub has a FOV that covers all angles.

a point in space, the markers can be imaged at up to ±90 degrees from their optical axes, and a pose estimate can be generated if the beacon is visible at any of these angles. ssume that the hub has a FOV in all directions. The tracking volume under these assumptions is a hemisphere with a radius of 3 m minus a hemisphere with a radius of 0.5 m (a 2D diagram

is shown in Figure 3.8). This corresponds to a tracking volume of V = 56 m3.

2 3

π33

2 3

π0.53 =

possible design based around using beacons consisting of 4 markers is considered. Fig- ure 3.9a shows a cube of dimensions 3 m × 3 m × 3 m that matches the tracking volume given in the system specification. Beacons are placed in the centre of each face of the cube except for the bottom face. This gives a total of 5 beacons and 20 markers. The tracking volumes arising from the side beacons are shown overlaid on top of each other in Figure 3.9b using the assumptions above. This illustration shows better tracking coverage in the centre of the cube with poorer coverage in the corners of the cube.

In reality the hub has a limited FOV and multiple beacons are needed to allow pose estima- tion to occur in all orientations throughout the tracking volume. It is difficult to estimate the number of beacons required to do this as it is a function of the FOV of the hub, the placement of the beacons, the number of markers required to produce an accurate pose, and the angles at which the markers may be imaged effectively. Using a number of camera modules it is believed that the configuration shown in Figure 3.9a can be used to provide tracking within most of the cube in many orientations, however, further development is required to test this.

3.5

Summary

The design behind the main functions of the system is analysed. Marker images are lo- cated by searching for bright pixels in an image frame. Regions of Interest (ROIs) are

placed around the bright pixels and surround the marker images. The centroid of the pix-

Document info
Document views393
Page views393
Page last viewedWed Dec 07 15:30:59 UTC 2016
Pages170
Paragraphs6307
Words54996

Comments