Research Phase 3

The analysis of the algorithms in Phase 2 exposed limiting factors to the reliability of the map correlation. This section discusses the next steps that were taken to make this algorithm more robust.

1. Feature tracking[6] was used to find the corresponding features in the consecutive video frames as seen in Figure 1a and 1b. As the video frames were recorded every 30^th of a second, the features in the frames shifted only a small amount making the tracking algorithm relatively straightforward.

 
           Figure 1a. Extracted Feature Points[11]         Figure 1b. Features Tracked Through Video Frames[11]

2. 

2. 2. The camera has six degrees of freedom (three for rotation and three for translation) see figure 2a. The Crista device (see components section) sampled data (accelerometer and gyro data) that could be used to derive information about the camera's movement (see figure 2b). Therefore this device was used with the camera to assist in making the feature tracking algorithm more robust.

   Figure 2a. Camera Degrees of Freedom[11]                              Figure 2b. Crista Data[11]