On Sat, 2010-09-11 at 19:35 +0100, Andrew Morgan wrote: > On 11/09/10 16:38, Peter Merchant wrote: > > After reading this, and thinking about the cars speeding up our road, I > > wondered how hard it would be to write a speed camera program for a > > webcam. > > > > I believe there is a phrase about asking "can I do this" before asking > "should I do this". > > Sounds like this would be something similar to the 'community speed > watch' scams. If so the results will be: > · Most people aren't speeding. > · The ones who are are your neighbors. > · You'll get bored of it quickly when you realise that you can't > actually /do/ anything. > > Now lets get rid of all this spying technology and start treating each > other like humans. > > (Also a company has had this idea already, they're called (or, were > bought by) RedSpeed. Their devices are little better than two or three > web cams and a computer running DOS.) > > -- Can/Should/Will - I was thinking of it as an academic exercise. I am not interested in buying kit. Simono has written the programs. How would one go about measuring velocity of things passing? Take three shots at known time intervals and measure the change in width/height of the moving object? Try and measure distance from a known fixed object such as a lamp-post? It might not be in the same plane. Use position relative to (vertical) fence posts instead of markings on the road? It all calls for powerful ability to pick out items form pictures.
Peter -- Next meeting: Bournemouth? TBD, Wednesday 2010-10-06 20:00 Meets, Mailing list, IRC, LinkedIn, ... http://dorset.lug.org.uk/ How to Report Bugs Effectively: http://bit.ly/4sACa
