Originally posted by: TridenT
Originally posted by: irishScott
Originally posted by: TridenT
Originally posted by: irishScott
Originally posted by: TridenT
I have always wondered... What if the police used ALL the video from everywhere? Why don't they now? I mean really.. You could basically track anyone almost anywhere and solve tons of crimes if they resourced all that video...
Oh, so if I rob a gas station in Massachusettes and then shop at Sears in California, tracking me is simply a matter of running unreliable facial recognition software over millions of net hours on the tapes recorded by hundreds of thousands of closed circuit cameras, each with it's own format and specs, assuming I faced the right direction for the cameras in the first place.
Why would you be looking at millions of hours instead of maybe hundreds if that? You just follow the guy... "He turns left a 162nd... Ok, cameras are following his car.. Ok now he is turning right at 12th NE... Camera 12 on 12TH ne.." I don't really get how you are going through MILLIONS...
In a city that might be doable, but in the suburbs it'd be ridiculously hard, and in rural areas next to impossible. I live in Fairfax County VA, with a population of well over 1 million. I know plenty of back-roads without lights. Even assuming that every light has a camera, you'd have to check every intersection within a several light radius to pick me up again, and by the time you got there I could easily be at least a country away.
Yes. I am talking a city, but I am also assuming that maybe you resource all the video from gas-stations, mini-marts, etc. Like it all is recorded and you can access it.
I am assuming that these are all digital recordings... All in easy formats like .avi or something. Assuming this, there is no need for billions of dollars in new hardware and shit. You just tell the corporations you need just a few bits of info from them about their recordings and where they are located... Then you build some software that gathers all that video into a nice stream for the end-user and you just process it.
I imagine a database where it is kind of like google-maps, and maybe even streetview, but on steroids. You can easily just follow the cameras as you progress through the map, maybe even live with enough bandwidth and processor power. You just kind of do it that way.