Poor IR contrast in new Tennessee license plates

[wittaj]

Cant wait to hopefully hear positive results from this! I am kinda surprised there are not more LED on it.

 

[wtimothyholman]

Cant wait to hopefully hear positive results from this! I am kinda surprised there are not more LED on it.

I learned from experimenting with the 940nm IR flashlight that one good high-wattage LED can be much brighter than you would expect.

 

[wtimothyholman]

I've just completed an hour of various tests with the 740nm IR floodlight, with zero success.

There are only two possible conclusions: either the Dahua -Z12 is almost completely insensitive to 730-740nm infrared light, or both the 730nm bulbs and the 740nm floodlight are complete duds as infrared sources. And yet both of those IR sources light up an interior room for an Amcrest PTZ camera with its own internal IR LEDs turned off.

I actually had a perfect target for my tests. One of my neighbors has new plates, and parked his vehicle across the street right in the field of view of one of my LPR cameras. I've attached two images, one using the internal 850nm IR diodes of the camera, and one with illumination from my 940nm IR flashlight. The Dahua -Z12 has no problem whatsoever getting a good plate reflection (albeit with terrible contrast) with those sources. But the 740nm floodlight might as well have been turned off for all the good it did. I even aimed the floodlight right at the camera, and only saw a small white blob.

At this point I'm stumped. It makes no physical sense that the camera sensor would be sensitive to 850nm IR light, 940nm IR light, and visible light, but not to 730nm or 740nm IR light. Could there be some sort of optical filtering going on? I really have no clue.

 

[wittaj]

That absolutely sucks.

Can you swing your PTZ over and see if it see it like you did with the other.

Have to wonder if the Z12 has some filtering going on?

 

[wtimothyholman]

That absolutely sucks.

Can you swing your PTZ over and see if it see it like you did with the other.

Have to wonder if the Z12 has some filtering going on?

That will be next. I'll try using the PTZ tomorrow night as an LPR camera and see if I get any results. I'm done for this evening.

 

[Flintstone61]

Thanks for your research.

 

[TechieTech]

Two questions: how far is the illuminator from the camera and are they at the same angle? If it's not very close to it and aimed in the same spot, you won't get the IR light bouncing off the plate back into the camera lens.

You can test this by standing right next to a floodlight and looking at a license plate in darkness outside of the main beam of light. You can clearly see the extra light bouncing off the reflective plate. Move farther away from the floodlight and you won't see the bounce as the light reflects back to the light source. I have an extra 850nm illuminator pointing right where one of my cameras is pointing and it doesn't do much for license plates because it's about 15 feet away from the camera and at a hugely different angle (below and off to the side), even though it is pointing at the exact same spot the camera does. There is a reason external illuminators are placed behind and directly above LPR cameras that do not have internal illuminators. Capturing the light bounce in the camera sensor is what is key, not simply having the illumination pointing at the same spot.

Another issue is that those illuminators probably have a 120 degree spill on them. If there is any way to narrow that beam down they will work much better.

 

[wtimothyholman]

Two questions: how far is the illuminator from the camera and are they at the same angle? If it's not very close to it and aimed in the same spot, you won't get the IR light bouncing off the plate back into the camera lens.

You can test this by standing right next to a floodlight and looking at a license plate in darkness outside of the main beam of light. You can clearly see the extra light bouncing off the reflective plate. Move farther away from the floodlight and you won't see the bounce as the light reflects back to the light source. I have an extra 850nm illuminator pointing right where one of my cameras is pointing and it doesn't do much for license plates because it's about 15 feet away from the camera and at a hugely different angle (below and off to the side), even though it is pointing at the exact same spot the camera does. There is a reason external illuminators are placed behind and directly above LPR cameras that do not have internal illuminators. Capturing the light bounce in the camera sensor is what is key, not simply having the illumination pointing at the same spot.

Another issue is that those illuminators probably have a 120 degree spill on them. If there is any way to narrow that beam down they will work much better.

No, I'm certain it's not the angle of reflection. The 940nm IR flashlight can be adjusted to narrow beam or wide beam. In both cases, holding the flashlight parallel to the camera, I had no problem zeroing in on the plate and getting an image. That was my sanity check.

With both the 730nm bulb and the 740nm illuminator parallel to the camera, I saw nothing. I even opened up the iris and gain settings to maximum, and increased the shutter time to the point where I could see the rear of the vehicle, and still saw nothing. And in my previous tests with my sample license plates, I could (with my own eyes) even see the reddish glow of the 730nm bulb reflecting off the plate, and the camera picked up nothing.

My cameras are about 60 to 70 feet away from the center of the roadway, so I'm not pushing the limits of the camera or the illuminators. As I said, I could actually see the reddish reflection of the 730nm bulbs from that distance. Logically, the camera should be picking up something, but it's not.

Tonight I'm going to try one final test, where I'll hold a sample plate a few feet in front of the camera and shine the illuminator on it. If I don't at least see something on the camera, my only conclusion will be that the camera itself isn't responding.

Of course, there is always the possibility that these cheap IR sources that I've purchased direct from China simply aren't generating significant light in the 730nm to 740nm range. I could buy a custom illuminator from Axton for $300 to be completely certain, but I hesitate to spend that much money unless I have some confidence it will really work.

 

[wtimothyholman]

Latest results: tonight I took a test tag and held it up about three feet away from one of my LPR cameras with its internal LEDs turned off. I then shined an external 850nm illuminator onto the plate. The glow of the plate was very bright and very visible.

Next, I used the 940nm flashlight, with the same angle of reflection. Same result - nice, bright plate image.

Then I tried the 740nm illuminator. The plate was actually glowing red from the visible portion of the LED's spectrum. And what did I see on the camera? Nothing, absolutely nothing.

This has to be an issue with the camera. For whatever reason, it is insensitive to the 730 to 740nm IR range.

 

[wtimothyholman]

@EMPIRETECANDY, in case you're reading this thread: can you check with the engineers at Dahua and find out why the Dahua IPC-HFW5231E-Z12 is not sensitive to infrared light in the range of 730nm to 740nm?

 

[EMPIRETECANDY]

ok, I will check it. It's the old one 5231 right?

 

[wittaj]

Did you try yet swinging the PTZ over to the street to see if this NM made the plate easier to read?

 

[wtimothyholman]

Did you try yet swinging the PTZ over to the street to see if this NM made the plate easier to read?

Not yet, but based on the preliminary experiments I performed so far, I think the PTZ is similarly insensitive to these wavelengths. I'll do a more formal experiment later this week.

Hopefully @EMPIRETECANDY can get more information directly from the Dahua engineers. I assume that this was a deliberate design choice, and some sort of filtering is taking place in the camera.

 

[wtimothyholman]

ok, I will check it. It's the old one 5231 right?

Correct, it is the older 5231 camera.

 

[wtimothyholman]

@EMPIRETECANDY, you should also tell the Dahua engineers that newer license plates in the USA may require 730nm to 750nm infrared light in order to be read properly at night, and that they might want to consider that as an option in future camera models.

 

[EMPIRETECANDY]

oh, i see, thanks buddy, i will tell them this.

 

[wtimothyholman]

I did some indoor testing of a Dahua camera (not a varifocal, but fixed lens) with a test plate, and now I'm beginning to again wonder if much of the problem may be that my cheap no-name illuminators are generating very little infrared light at 730 nm or 740 nm.

I would hate to spend $300 on an Axton illuminator and find out that it won't work with my Dahua -Z12 camera, but I may have no choice if I really want to figure out what's going on. I do have a Dahua -Z12E that I can set up indoors for testing.

 

[wittaj]

That certainly is a possibility for the shutter speeds we run?

I wonder if Axton could produce it but also have filters one could put over for various nm so if it doesn't work in the 730 range you could add a filter to make it 850 and use for another camera?

 

[biggen]

Its too bad you can't test with a 5241 since the 5231 isn't being made any more is it? One would think Dahua could also test this as well.

 

[wtimothyholman]

Its too bad you can't test with a 5241 since the 5231 isn't being made any more is it? One would think Dahua could also test this as well.

The 5231 is all I have. If I had some assurance that the 5241 worked better, I would consider buying one. But even then, I'd still need a decent 730-750nm IR illuminator to be sure it was working.

I am making some inquiries to see if I can borrow a good 740nm illuminator. Fingers crossed on that.

It would be nice if Dahua could provide some test results at different IR wavelengths, but they may not see the point. As far as they would be concerned, the camera includes 850nm IR LEDs, so that's all that matters.