Comment by nine_k
13 hours ago
If speeding is not something that happens constantly, then a radar could detect the instances of speeding, and only turn on a camera when a speeding car is nearby. This would keep the majority of passing cars from being recorded, and would record the fewer cars the fewer drivers would be speeding.
How long until the AI estimates how fast you were going based on the time you were tagged at 2 cameras? The system says travel between these cameras should take 3 minutes. You made it in 2:45. No review, just a ticket in the mail.
That's how it actually works in some places (https://www.carwow.co.uk/blog/average-speed-cameras-how-do-t...). You don't need AI because we already know that distance = speed * time. We can calculate to a high degree of certainty - with high school math - that you had to have 1. been speeding or 2. bent space time if you cover the distance between them in too short of time.
Several months later, the state changes the speed limit on the street without changing the ticket check. Some time after, traffic authorities add a bypass road or a fast lane to deal with the new traffic problems and forget to update the timings. Then it's reported that a rural municipality set timing thresholds just below the speed limit to cover budgetary shortfalls.
Eventually the camera manufacturers offer "easy setup" systems that dynamically adjust based on actual journey times. This works fine until the first office holiday, when the routine congestion is gone.
If their clocks are actually synchronized and can show that in court then unless you've got a warp drive and can bend space time it can be pretty definitive that one was speeding between those two points. Especially if you're on something like a toll road which utilizes transponders for billing. I'm almost surprised this isn't already done in those situations.
You don't even need AI for this. Its pretty basic.
It is similar to how air enforced speed limits are done. They just paint two lines on the highway. A plane overhead times your car between the lines.
This is exactly how non-AI assisted speed cameras [1] have worked for almost four decades. You don’t even need video for it.
[1]: https://en.wikipedia.org/wiki/Gatso
If speeding is not something that happens constantly, then a radar could detect the instances of speeding
And if it is, which it generally is, it means the speed limits are not set appropriately. But that point always seems to be overlooked.