Poor IR contrast in new Tennessee license plates

To give everyone an idea of how it will work, below is a photo made tonight from the illuminator mounted below my camera, with the camera operating in daytime mode.

For this test the illuminator is running at continuous output, which means it is too bright. Fortunately I now have a current mode LED driver that can be pulsed at a frequency up to 10 kHz, with a rise time of less than 10 microseconds. I am connecting that to an NE555 circuit which will generate 1 millisecond pulses at a frequency of approximately 100 Hz, resulting in a duty cycle of ~ 10%. To the eye, the light will look much dimmer, while the camera should capture multiple frames each second at full illumination - or at least that's what I'm hoping.

I've got a few more parts arriving this week, and then I'll give it the acid test.

Screenshot 2023-10-09 at 7.35.59 PM.png
 
To give everyone an idea of how it will work, below is a photo made tonight from the illuminator mounted below my camera, with the camera operating in daytime mode.

For this test the illuminator is running at continuous output, which means it is too bright. Fortunately I now have a current mode LED driver that can be pulsed at a frequency up to 10 kHz, with a rise time of less than 10 microseconds. I am connecting that to an NE555 circuit which will generate 1 millisecond pulses at a frequency of approximately 100 Hz, resulting in a duty cycle of ~ 10%. To the eye, the light will look much dimmer, while the camera should capture multiple frames each second at full illumination - or at least that's what I'm hoping.

I've got a few more parts arriving this week, and then I'll give it the acid test.

View attachment 174391

Looks promising. How does reading a front plate work out at night with the headlights shining into the camera? Is the HLC enough to take care of that situation?
 
Reflective coatings won’t help one bit against an LPR camera with white light illumination. If it is visible in the daytime, then the LPR camera will read it.
The problem I am seeing more and more is not a spray coating, but folks are putting smoked plastic plate covers on. It makes them hard to see even in daylight, and even by eye when following behind them.
 
The problem I am seeing more and more is not a spray coating, but folks are putting smoked plastic plate covers on. It makes them hard to see even in daylight, and even by eye when following behind them.
You are correct, but it's a self-defeating tactic. Any type of cover over a license plate is strictly illegal where I live. So these people who think they're being clever, and outwitting the "man" by obscuring their plates, are doing nothing more than giving the police all the justification needed to perform a traffic stop at any time. Any time the police want to put an end to smoked plates, all they have to do is start ticketing everyone who has one.

As I've told a couple of people: "The police don't care about tracking you. But the person who steals your car will really appreciate the effort you made to make your car harder to spot."
 
Different countries have differnet rules, but most of the world can't hide or cover the plate, will have big punishment at HONGKONG.
 
Fortunately I now have a current mode LED driver that can be pulsed at a frequency up to 10 kHz, with a rise time of less than 10 microseconds. I am connecting that to an NE555 circuit which will generate 1 millisecond pulses at a frequency of approximately 100 Hz, resulting in a duty cycle of ~ 10%. To the eye, the light will look much dimmer, while the camera should capture multiple frames each second at full illumination - or at least that's what I'm hoping.

I have always been surprised that security cameras don’t pulse their lights in sync with the shutter. Why make them visually more noticeable and waste POE power when the shutter is not open? Also, it used to be that LEDs would produce much more power when they are pulsed due to heat dissipation.

This should be especially true for LPRs Where the shutter is only open for ~1 mS. Strobe the light once per frame. If the camera were operating at 20 Hz, the light would be on only 20 mS per second. The light would appear 98% dimmer, be 98% cooler, and save 98% of the energy. All while providing the same image. Shorter pulse lengths for better freezing motion would save even more.
 
I have always been surprised that security cameras don’t pulse their lights in sync with the shutter. Why make them visually more noticeable and waste POE power when the shutter is not open? Also, it used to be that LEDs would produce much more power when they are pulsed due to heat dissipation.

This should be especially true for LPRs Where the shutter is only open for ~1 mS. Strobe the light once per frame. If the camera were operating at 20 Hz, the light would be on only 20 mS per second. The light would appear 98% dimmer, be 98% cooler, and save 98% of the energy. All while providing the same image. Shorter pulse lengths for better freezing motion would save even more.
Not only would this feature save power, it would also provide much better nighttime images with reduced (or eliminated) motion blurring. Even discrete LEDs can easily be strobed at tens of kilohertz, well beyond what would be needed for a security camera. It would be a very simple and obvious upgrade.

The one drawback I can't get around with my design is that there's no way to synchronize the shutter of my Dahua -Z12 camera with the illuminator. Instead, I plan to run it at some frequency that is not a multiple or fundamental of the camera frame rate, i.e. for a 60 Hz frame rate, the illuminator will operate at 83 Hz, with an "on" time approximately twice the shutter time. That will ensure sufficient overlap between the shutter opening and the illuminator being on.

So two features that every security camera ought to have (but doesn't):

(1) A strobed illuminator that is synchronized to the camera frame rate and shutter time.

(2) A way to enter the latitude and longitude of the camera into the web interface so that the camera can switch between day and night modes without the use of external software or processors. It is ridiculous that modern security cameras don't have this feature, and that we have to run day-night switching utilities to accomplish it. I've got programmable light switches that are 15 years old that have this built in. The amount of memory and processing power required are trivial by modern standards.

The fact that such obvious features are missing tells me that the people who are making the design decisions do not use their own products, or they would have added these features long ago.
 
I cannot find the post, but someone mentioned maybe with more IR you will get better contrast. That is exactly what I am finding out with the plates here in California. I have added more IR and it does make the hard to read plates easier to read by providing more contrast. Not sure how that would work with the Tennessee plates. Our hard to read California plates have a black background and a yellow/orange lettering. The image comes back showing the colors switched at night.
 
  • Love
Reactions: tech101
Wow, those are impressive snapshots for a camera tuned for LPR. Seems pretty much usable for regular captures too. What are the exposure/gain settings? Is there just a ton of light there?
 
I don't know if this has already been mentioned but Dahua has these two ANPR models with 730 nm IR:
DHI-ITC413-PW4D-Z1 (2.7mm to 12 mm), Dual Illuminators (730 nm IR/warm light)
DHI-ITC413-PW4D-Z3 Dual Illuminators (8mm to 32mm)(730 nm IR/warm light)
As it turns out, a fellow resident here in Nashville has purchased two -Z3 cameras to set up a custom LPR system for his neighborhood. He has the budget and know-how to install them, and I've been providing some advice to him. He has promised to provide me with some images once they're up and running.
 
As it turns out, a fellow resident here in Nashville has purchased two -Z3 cameras to set up a custom LPR system for his neighborhood. He has the budget and know-how to install them, and I've been providing some advice to him. He has promised to provide me with some images once they're up and running.

Oh you are so fast. I deleted the message after I found out somebody else already mentioned it months ago lol.

I will be purchasing one but with the 850nm IR as they don't have any 730nm plates in Texas yet.
 
  • Like
Reactions: Parley
Oh you are so fast. I deleted the message after I found out somebody else already mentioned it months ago lol.

I will be purchasing one but with the 850nm IR as they don't have any 730nm plates in Texas yet.
"Yet" being the operative word. I think the use of 3M printed plates is a growing trend in a lot of states. Even if Texas plates are still readable at 850nm, you'll still find yourself missing plates from other states.

I think that 730/740nm IR, or warm white light illumination, is the wave of the future for all LPR cameras. State legislators simply don't care if license plates can't be read by 850nm LPR cameras. All they care about is the revenue they can collect by offering hundreds of different styles of vanity plates.