Computer Vision For The Blind
HomeIntroductionPhase 1Phase 2Phase 3Background
ConclusionsFuture WorkApplicationsReferences








Phase 1

Research Phase 1

In this phase a study was performed to examine the feasibility of the hypothesis.

 

1. A video camera was mounted to a piece of wood that was secured onto a bicycle helmet see figure 1

Helmet
Figure 1. Testing Helmet
[11]


2. A test subject wearing this helmet walked through a designated room 

3. Video was recorded as the subject navigated through the room

4. A simple map of the room was then constructed by hand see figure 2

Map
Figure 2. Map of Residential Building[11]

5. The video was imported into MatLab where selected  object features where picked out by hand that corresponded to identical features  from the map see figure 3(see colored markers)

Image Features

Figure 3. Image Features[11]

6. Using the feature’s angles(see components section) relative to the camera’s boresight and the dimension’s from the map the camera’s position could be calculated see figure 4

Trajectory Tracking

Figure 4. Map Correlation[11]

7. From the sequential camera positions the trajectory could be estimated

 

Copyright © Christopher Nielsen, 2009. All rights reserved.
Contact: collectorchris@shaw.ca

a