Most literature uses processing of video/cam images. I would need a simpler solution, also need to avoid ethical issues due to videoing people.
Suggestions:
1) Kinect based human tracking: you can use only the 3D data and "virtually" the person will remains anonymous (no photos taken, only 3D point cloud);
2) Thermal camera (passive infrared) still remains a good option (anonymous people);
3) Active methods: use RFID tags (people carrying RFID tags) and strategic placed RFID readers. In this case, people will know they are being tracked, and "authorize" it since they are carrying the RFID tags.
4) Some other creative methods... walking tracking, but you will need a specific walking surface infra-structure.
Have a look at this:
http://www.thermitrack.com/
I stumbled across this a while ago and found it interesting.
Stephen, I take a look over the specifications:
"Thermal images : 16x16 pixels, 256 grey levels" Poor resolution.
http://www.thermitrack.com/prodinfo.html
I understand the resolution of these devices is poor, but if all you need it detection and tracking, and not imaging these devices will do all that for you - passing you coordinates of multiple tracked objects in the field of view.
Another approach may be to use video from a camera but not store the footage directly, instead only store information from the footage you require such as number and position of people in view, speed they move at, how long they stop at certain points etc,,
Triangualtion using High Freq Acoustics and Doppler Shift, computationally much simpler and faster when compared to the image processing.
- Badges
- Science topic