Not sure why this is being downvoted, it is an actual thing:
“Several companies operate independent, non-law enforcement ALPR databases, contracting with drivers to put cameras on private vehicles to collect the information.”
“DRN is a private surveillance system crowdsourced by hundreds of repo men who have installed cameras that passively scan, capture, and upload the license plates of every car they drive by to DRN's database. DRN stretches coast to coast and is available to private individuals and companies focused on tracking and locating people or vehicles.”
To further this, I know someone who was a public defender a decade ago in the US. They would regularly try to get intersection camera footage to attempt to prove their clients not guilty. Often it was impossible to track down who owned the cameras, or even if they were municiple or private. I imagine this has only gotten worse as cameras have gotten cheaper.
So many systems are legacy and entirely unused, or best effort - ie. if the camera is working, it'll be used, but nobody will be sent to repair it if it's broken.
Is the location of your state-registered vehicle vehicle on publicly owned roads personal information in your state? Despite the phrasing, I'm honestly curious.
No, but likely his face biometrics are off he happens to have been captured in the video. So if they are keeping the raw video, not just the processed license plate output, they may fall under this jurisdiction.
Assuming you're talking about the Illinois Biometric Information Privacy Act, that only requires consent for facial recognition databases based on facial geometry. The law explicitly allows storing photographs of people's faces without consent.
It's probably not, which is the problem. I feel like the wilful processing of such details into parse-able, privacy-violating form should be regulated.
What state is that, California? I'd love to have a way to block ubiquitous surveillance, also places taking your picture when you walk into a drugstore or restaurant, uploading it. If it hasn't happened yet, we will be tracked all the time.
I didn't downvote it, but $50 for cameras is not even a blip in the cost of setting operation like that, and op's comment make it sound, like that's all you need.
You'd need to place those, maintain them, prevent vandalism, route power to them, pay for network connectivity and weather-proof them. The price of the cameras is not the problem.
You'd likely lose a lot of money trying to do that, you're vastly underestimating the cost of maintaining the data network for that kind of installation alone, even before looking at the cost of physical installation.
It absolutely is a bargain. I don't have the money or time to risk but I strongly recommend that you start a company, build a website and advertise your service to the world. I'm assuming you'll also deal with any regulatory and data protection issues, get permission from local authorities etc... . You've found an astonishing gap in the market, it seems.
SLA: 75% of the cars that pass will be captured. The 25% covers equipment failure/theft/service outage/etc. This applies to the whole set, not per camera.
We mutually agree the roads - but it's unlikely I'll disagree unless you want to place four cameras per country worldwide to maximize logistics difficulties.
Half upfront, half when the service has been operating 3 months. Full refund if SLA not met. If you can show you're very solvent, you can pay in arrears instead.
1 year minimum (my breakeven point is 8 months optimistically, 12 months more realistically, and the profit comes from either you continuing the contract past 12 months or from someone else contracting me for the same cameras).
Fair point, you can almost read number plates on some of the cameras that are close to roads. Presumably the cameras are capturing data at a higher resolution than those public images?
I'd bet with a bit of data processing on that video feed (subpixel aligning and stacking all the frames, using the prior knowledge that you know the numberplate isn't changing as the car drives down the road), you'd be able to read them.
Not even close. Cars on a freeway are doing 10 to 100mph. To capture a plate on a moving car you need less than 1" of movement during the exposure time.
At 10mph a car is moving at 176 inches per second, and to be able to read the plate with less than 1" of movement you need a camera that has a shutter speed of 176th of a second, or rounded up at 1/200th . To capture at 100mph a vehicle is moving at 1760 inches per second. To read the plate you'll need a camera that can capture at 1760th of a second, or rounded up to 1/2000.
I don't know of any android phones that can capture both at 1/2000th and video on a second camera at the same time to know when to fire the first camera.
You assumed the camera was capturing a photo of the car viewed from the side where the full speed is apparent (and ironically where the plate will not even be visible).
Most cameras point down the length of the road and the "speed" that the camera sees is only a fraction of that. You can record a video with a merely-ok phone and probably see most plate numbers assuming the lighting isn't terrible. Good luck getting a phone camera to work at night with an LED flash though
(oh and also, this assumes you want to catch people speeding, to capture every plate number you would just put the camera near a slow area like a bend or stop sign)
I assure you the motion blur, even looking down the same plane (parallel) to the car will not be able to capture the plate, even in decent light on video alone. Go try to take video yourself of a car driving by in full daylight.
The obvious approach is to just train your numberplate recognition algorithm with blurred plates. Since the blur is almost equal across the whole plate, and nearly all cars are moving in the same direction, you aren't really losing much information. Sure, it might be hard for a human to read, but for a deep learning algorithm I don't think it's actually any harder.
But there are other approaches too - like putting a 99 cent novelty zoom lens on the front of your camera to capture more light for your region of interest, allowing you to use shorter exposure times. Or an infrared strobe light that flashes once per frame (most numberplates are retroreflective, so IR strobes work really well).
'Something you could do anyway' with something that would actually cost a huge amount of money and effort to build (those $50 phones probably wouldn't have adequate cameras, and would still need to be supplied power by solar panels or something, not to mention being subject to removal).
did you know that the city of Menlo Park california requires companies to have cameras which can see for miles and track and report all vehicles that drive by them....
But this way you don't have to pay for the cameras - someone else has done it for you.