Scaling to real life information and miscellaneous details
The average bumper of a vehicle is roughly 20 inches off the ground. The testing height the kinect was roughly 27 inches off the ground (mounted 7 inches the average bumper).
The first set of red lines represent 3 feet away from camera given 27 inches off the ground with the kinect pointed slightly downward.
The distance between the first set of red lines on the user interface is roughly 40 inches,
The average vehicle width is between 65-80 inches, thus this software is intended for about 1/2 the size of a standard vehicle. There are peripherals that can be added to the Kinect hardware to give the camera a wide-angel lens and this should be considered as a possible improvement to the project.
The area directly within 1.5 feet of the camera is considered "unreliable". The Kinect cannot parse sensory information this close, thus use and testing is under the assumption nothing is within this unreliable zone.
Pictures are saved and sounds are referenced from the local project 'bin' folder. There are no external dependencies.
The scalar multiple for the reverse speedometer is between 1 and 2. The scalar multiple is multiplied by the boundaries to increase the detection distance as the speed of the vehicle increases.
Picture snapshots are timestamped with the date taken.
*Possible improvement: Use the Kinect's skeletal tracking to detect people behind the vehicle and send the user a unique alert indicating the danger.
| senior_cs_project_-_car_cam.pptx |
| car_cam_project_build.zip |