Camera system for HOV (car pool lanes) monitoring?

[mark_kinesense 0]

Hello all,

 

I'm looking for companies that supply camera systems specifically for High Occupancy Vehicle lane monitoring. I need something that can basically count the number of people in the car. Does anyone know if such a thing exists?

 

My google searches aren't finding it.

[BestPriceCctv 0]

Hi

 

There are alot of factors they play into fact with this

 

Depents on how fast the cars are moving

 

You need a good recording system that has good seaching

 

I have setup a few in service bays,street corners,parks,grocery stores

The best system is 3vr if you have a budget for it can do everything your looking for

 

Use axis ip cameras are very good for this job

 

You can also do lpr and counting on cheaper dvrs but it comes down to how much time you want to earch and look up the plates when 10000 cars go by a day

 

And the cheaper dvrs are a mess to get going sometimes

if you need some help pm me ive also done alot of facial reconition and lpr and im a certified installer for 3vr

[mark_kinesense 0]

Thanks a lot for that reply.

 

I am particularly interested in systems which automatically detect single person occupancy for vehicles.

[chaz 0]

you may want to contact 3vr. they have people counting and lpr from the same box. if they see a market for the application they might address it.

[ssnapier 0]

This would be a VERY tall order. Think about all of the factors at play!

 

You would need to solve all of the following:

 

glare on the glass

 

speed of the car

 

intelligence to figure out if objects are actually people

 

sunrise and sunset

 

rain, snow, fog...

 

I simply don't see how a camera, even with the best filters, would be able to accomplish this with any kind of accuracy.

 

Now, if you were to have a live person watching these cameras with the ability to constantly change exposure and other settings, then you might have a slightly better chance.

[mark_kinesense 0]

This would be a VERY tall order. Think about all of the factors at play!

 

You would need to solve all of the following:

 

glare on the glass

 

speed of the car

 

intelligence to figure out if objects are actually people

 

sunrise and sunset

 

rain, snow, fog...

 

I simply don't see how a camera, even with the best filters, would be able to accomplish this with any kind of accuracy.

 

Now, if you were to have a live person watching these cameras with the ability to constantly change exposure and other settings, then you might have a slightly better chance.

 

I totally agree, and I would be skeptical of any commercial offering of such tech... however, I have been given the task of finding if such product offerings exist on the market or not.

[skomo 0]

dont forget about small kids in back seats.

 

if it exists, it would be custom made...

[ssnapier 0]

It would also have to have at least 6 different filters on board and the intelligence (or a full time person making adjustments) to change the filters and adjust the focus, exposure, shutter speed, etc.... nightmare.

[mark_kinesense 0]

It would also have to have at least 6 different filters on board and the intelligence (or a full time person making adjustments) to change the filters and adjust the focus, exposure, shutter speed, etc.... nightmare.

 

Why would it need 6 filters?

[ssnapier 0]

I really just pulled 6 out of thin air, but this is to adjust for the constantly changing lighting conditions. You would need at least an IR cut, polarizing, gradient (maybe 2 or 3) and a general neutral density (tinted) filter.

 

I almost suggested thermal instead, but then I remembered that thermals don't work thru glass.

[mark_kinesense 0]

I really just pulled 6 out of thin air, but this is to adjust for the constantly changing lighting conditions. You would need at least an IR cut, polarizing, gradient (maybe 2 or 3) and a general neutral density (tinted) filter.

 

I almost suggested thermal instead, but then I remembered that thermals don't work thru glass.

 

Can radar go through glass?

[ssnapier 0]

That is too broad of a question, there are many different frequencies and methods of doing radar. I actually do not know if it would go thru glass, but I don't think it would work in the way you are expecting.

[mark_kinesense 0]

That is too broad of a question, there are many different frequencies and methods of doing radar. I actually do not know if it would go thru glass, but I don't think it would work in the way you are expecting.

 

Hmmm...

 

How about using a grid of laser dots similar to how the Microsoft Kinects depth sensor works?

[ssnapier 0]

You want to shoot multiple lasers are people driving cars? No way in the world that gets past the safety nazi.

[Kawboy12R 0]

But just think of the possibilities- you could turn up the kilojoules on cars that're identified as single passenger until they change lanes...

[GrouchoBoucho 0]

The best system is 3vr if you have a budget for it can do everything your looking for

 

...polarizing, gradient (maybe 2 or 3) and a general neutral density (tinted) filter.

don't forget each of those is going to reduce the light transmission (a polarizer by up to two stops, or dropping the light by 75%), which will require a lower shutter, which will increase your motion blur.

 

license plates are one thing because they're highly reflective and you can illuminate and expose for them specifically... illuminating the entire inside of the car is a whole other can of worms.

[mark_kinesense 0]

But just think of the possibilities- you could turn up the kilojoules on cars that're identified as single passenger until they change lanes...

[mark_kinesense 0]

You want to shoot multiple lasers are people driving cars? No way in the world that gets past the safety nazi.

 

The laser, like the Kinnect, would be non visible.

 

Would it work, in theory? You could perhaps model an empty seat as a concave surface. If the seat scan is convex then someone (or something) is there.