License Plate Recognition with flat/printed license plates

[Scruit 0]

Am I the only one who has seen problems using infra-red to read printed license plates (not stamped). I'm using an 850nm emitter and an 850nm ir-pass filter.

 

They appear as blank to me. I've seen this with an Ohio Veteran plate and a stock Indiana plate.

 

So what do the big names like Extreme do with Indiana plates?

[ak357 0]

Am I the only one who has seen problems using infra-red to read printed license plates (not stamped). I'm using an 850nm emitter and an 850nm ir-pass filter.

 

They appear as blank to me. I've seen this with an Ohio Veteran plate and a stock Indiana plate.

 

So what do the big names like Extreme do with Indiana plates?

 

Get camera which has "negative image option""

have a look under IR

results are not too bad

 

[bryceb 0]

Someone else on this forum suggested using a combination of IR cameras and non-IR in order to get both.

[Scruit 0]

Am I the only one who has seen problems using infra-red to read printed license plates (not stamped). I'm using an 850nm emitter and an 850nm ir-pass filter.

 

They appear as blank to me. I've seen this with an Ohio Veteran plate and a stock Indiana plate.

 

So what do the big names like Extreme do with Indiana plates?

 

Get camera which has "negative image option""

have a look under IR

results are not too bad

 

 

 

Wouldn't there need to be some contrast in the original for a negative image to work?

 

What does the non-negative image look like?

[cglaeser 0]

Wouldn't there need to be some contrast in the original for a negative image to work?

 

I would assume so.

 

Best,

Christopher

[rory 0]

What does the non-negative image look like?

 

negative again.

[cglaeser 0]

negative again.

 

I'm guessing the original was nearly identical to that double negative. If that's the case, then the negative adds no forensic information. You might be able to pull something out using a photo editor, but just reversing video does not really help.

 

Best,

Christopher

[rory 0]

negative again.

 

I'm guessing the original was nearly identical to that double negative. If that's the case, then the negative adds no forensic information. You might be able to pull something out using a photo editor, but just reversing video does not really help.

 

Best,

Christopher

Yes if you make an image negative, making it negative again will take it back to the original. and so on and so forth

 

Eg. Original - Negative - Negative

[Scruit 0]

What does the non-negative image look like?

 

negative again.

 

Looks like a stamped plate...

[cglaeser 0]

Looks like a stamped plate...

 

Agreed. The shadows do suggest that.

 

Best,

Christopher

[ak357 0]

Am I the only one who has seen problems using infra-red to read printed license plates (not stamped). I'm using an 850nm emitter and an 850nm ir-pass filter.

 

They appear as blank to me. I've seen this with an Ohio Veteran plate and a stock Indiana plate.

 

So what do the big names like Extreme do with Indiana plates?

 

Get camera which has "negative image option""

have a look under IR

results are not too bad

 

 

 

Well I guess I was not clear with my explanation

lets try again

 

1. get camera which can do "negative: as video output

( you playing with image after its has been capture --wrong )

 

2.look at the image under IR

 

-------------------------------------------------

here is image of plate under IR ( non negative)

 

 

--------------------------------------------------

then my pix in my First post is produce by camera with "negative" output

 

that all

 

lic plate number are different

just lost original

any ???

[cglaeser 0]

Well I guess I was not clear with my explanation

lets try again

 

1. get camera which can do "negative: as video output

 

Why??? Are you suggesting negative video provides more forensic detail than positive video?

 

Best,

Christopher

[ak357 0]

Well I guess I was not clear with my explanation

lets try again

 

1. get camera which can do "negative: as video output

 

Why??? Are you suggesting negative video provides more forensic detail than positive video?

 

Best,

Christopher

 

All I am saying that video with "negative" can see in my case RED numbers much better

 

Christopher IR will not get reflected from RED number

 

that why I show results above

 

by the way all pix are been taken by WDR CNB while back ( 605 in model number sorry don't remember exactly )

since then do not buy or sell CNB my self

[rory 0]

-------------------------------------------------

here is image of plate under IR ( non negative)

 

Can we see the negative of this one, or the original of the other one?

Why do you think their negative would be much much different from doing it digitally?

Just asking as I havent tried it but cant see it being that much of a difference, just less quality.

[bryceb 0]

 

--------------------------------------------------

then my pix in my First post is produce by camera with "negative" output

 

that all

 

First of all, I don't think you understand his problem. Second, each image you provided shows identifiable numbers on the license plate, regardless of whether it was the original or negative.

 

The problem that he is discussing is that IR cameras can't read a printed page. If you took a magazine cover with white and red print and held it up to an IR camera in complete darkness, you wouldn't be able to tell the difference between it and a blank page.

 

Some states have license plates that are printed and not stamped, and the print comes in a red color that makes it difficult to pick up using ONLY IR imaging.

 

I think the only way to get around this is a non-IR camera with a light source.

[Scruit 0]

ak357,

 

Makes sense to me in that one the image is saved to jpg then the similar colors will be 'locked' in so negative of that image will simply produce a negative (but still difficult to read) image. In fact, a negative of your "original" (difficult to read) is just as difficult to read.

 

But you're saying that with the camera set to give a negative image, it DOES produce enough contrast to read the plate... You can confirm that the ONLY

difference between the two images of that license plate that you produced is that one of them is set to negative *at the camera* and the other is normal... all other things (lens, camera, brt, cont, etc) are the same?

 

If so, then that sounds like a viable option. I was already looking for an excuse to replace my cam with a WDR. I need a WDR with Negative...

 

Thanks,

Grae

[ak357 0]

ak357,

 

Makes sense to me in that one the image is saved to jpg then the similar colors will be 'locked' in so negative of that image will simply produce a negative (but still difficult to read) image. In fact, a negative of your "original" (difficult to read) is just as difficult to read.

 

But you're saying that with the camera set to give a negative image, it DOES produce enough contrast to read the plate... You can confirm that the ONLY

difference between the two images of that license plate that you produced is that one of them

is set to negative *at the camera*

and the other is normal... all other things (lens, camera, brt, cont, etc) are the same?

 

If so, then that sounds like a viable option. I was already looking for an excuse to replace my cam with a WDR. I need a WDR with Negative...

 

Thanks,

Grae

 

Yes ,the same

I will try to do again if I can find cam with"negative image option"

[cglaeser 0]

I will try to do again if I can find cam with"negative image option"

 

If you are saying that sometimes it is easier to find Waldo by reversing video, then yes, I agree. Many thermal and night vision systems have the option to reverse video, and sometimes it's just easier for the human eye to locate objects of interest, and sometimes it's operator preference. But if you are saying that reverse video increases forensic detail, then please post an example of the positive and negative taken of the same plate at the same time.

 

Best,

Christopher

[Scruit 0]

I will try to do again if I can find cam with"negative image option"

 

If you are saying that sometimes it is easier to find Waldo by reversing video, then yes, I agree. Many thermal and night vision systems have the option to reverse video, and sometimes it's just easier for the human eye to locate objects of interest, and sometimes it's operator preference. But if you are saying that reverse video increases forensic detail, then please post an example of the positive and negative taken of the same plate at the same time.

 

Best,

Christopher

 

 

I think we will find that there is a difference between the negative image of visible light and the negative image infra-red light.

[cglaeser 0]

I think we will find that there is a difference between the negative image of visible light and the negative image infra-red light.

 

Well, of course. The reflectivity (and contrast of two adjacent) colors will vary across the spectrum. Two colors may have high contrast at 400 nm and low contrast at 850 nm. I have a black jacket that appears white on an IR camera, and quite similar to a light pair of jeans in IR, significantly less contrast than the visible spectrum. The question we've been discussing is, is there a difference in forensic detail of the positive vs negative at 850 nm.

 

Best,

Christopher

[Scruit 0]

I can't seem to find any box cameras with WDR *and* negative image option for under $400... Does anyone have any suggesitons?

[cglaeser 0]

I can't seem to find any box cameras with WDR *and* negative image option for under $400... Does anyone have any suggesitons?

 

Why do you want negative image option? Have you seen even one comparison example yet? If so, can you please post a link here?

 

Best,

Christopher

[rory 0]

bottom using negative image .. not same as as full image from camera though:

 

thing is, what software are you using to capture/process the plate numbers?

[Scruit 0]

While ak357 is looking for his example I'm trying to figure out if it's worth pursuing from a financial perspecitive because of the cameras are $500 then that is just not going to happen.

 

For me the confirmation will be:

- The plate is illegible on a positive IR image,

AND

- Importing that illegible image into photoshop (etc) and doing a negative image on the PC still results in an equally illegible image

AND

- A plate output using the camera's negative feature IS legible.

 

IF those three are true, and I can get the camera (Just the box cam, I already have the lens/filter/housing) for about $200 then I'll be all over that.

[cglaeser 0]

bottom using negative image ..

 

Thanks for posting these. This is what I would expect. The plate can be identified using either positive or negative. Does someone have an example where the plate is only readable in negative and not readable in positive?

 

Best,

Christopher