DORI calculations

[PedroAsani]

I have a system that I have wired and built myself, consisting of 9 Alptop AT-LIR400. The coverage is great, but I'm curious what the numbers are for it.

3.8mm lens, 76 degree angle, 4MP feed. How do I plug all of this into the DORI formula?

Radius = (( Horizontal Res / 100 ppf ) * ( 360 / Angle ))/2*Pi

Radius = (( Horizontal Res / 100 ppf ) * ( 360 / 76 ))/2*3.14

 

[mat200]

I have a system that I have wired and built myself, consisting of 9 Alptop AT-LIR400. The coverage is great, but I'm curious what the numbers are for it.

3.8mm lens, 76 degree angle, 4MP feed. How do I plug all of this into the DORI formula?

Radius = (( Horizontal Res / 100 ppf ) * ( 360 / Angle ))/2*Pi

Radius = (( Horizontal Res / 100 ppf ) * ( 360 / 76 ))/2*3.14

Hi @PedroAsani

What is the H resolution?

 

[PedroAsani]

2560 x 1440 I think.

 

[mat200]

2560 x 1440 I think.

Hi Pedro,

SO which of those numbers is the Horizontal Resolution, and which is the Vertical Resolution?

 

[PedroAsani]

2560. I just remoted in and checked.

 

[mat200]

2560. I just remoted in and checked.

Very good.. now plug that number in and use a calculator to determine the result for the ID distance
( let us know what you get for the answer.. )

 

[PedroAsani]

I have 190.38 which I am assuming is feet.

 

[mat200]

I have 190.38 which I am assuming is feet.

Hi @PedroAsani

Yes as the equation has "pixels per foot" -> thus the unit all cancel ( pixels / pixels and degrees / degrees ) except the per foot.. 1/(1/foot) = foot

Thus the answer will be in feet

However, 190.38 feet seems crazy long distance for this... so recheck your calculations.

 

[PedroAsani]

Wolfram Alpha did the calculation, (( 2560 / 100 ) * ( 360 / 76 ))/2*3.14

 

[mat200]

Wolfram Alpha did the calculation, (( 2560 / 100 ) * ( 360 / 76 ))/2*3.14

Hi @PedroAsani

Double check your inputs... ( hint missing a factor of 10.. )

 

[PedroAsani]

Hi @PedroAsani

Double check your inputs... ( hint missing a factor of 10.. )

I'm not seeing it.

 

[mat200]

I'm not seeing it.

Hi @PedroAsani

Try calculating it by hand then...

 

[PedroAsani]

Hi @PedroAsani

Try calculating it by hand then...

Same answer.

2560/100 = 25.6
360/76 = 4.74
25.6*4.74 = 121.344
121.344/2 = 60.672
60.672*3.14 = 190.51 (up by 0.2 because rounding)

So unless one of the numbers in the formula is incorrect, this is what I have.

 

[mat200]

((2560/100)*(360/76))/6.28=
19.30942004693261816959

hint: order of math operations matters...

 

[PedroAsani]

Ah, order of operations issue. I grabbed it off the wiki page, so it could use a modification and some examples.

Radius = (( Horizontal Res / Pixels per face ) * ( 360 / Angle )) / ( 2*Pi )
Assuming ppf of 5/20/40/100 for DORI
So a 2MP camera with 1600 x 1200 and 87 degree Horizontal angle would have a DORI of 210.8/52.7/26.3/10.5

My AT-LIR400 has 386.1/96.5/48.3/19.3

Does that seem more in line with what is expected?

 

[mat200]

Ah, order of operations issue. I grabbed it off the wiki page, so it could use a modification and some examples.

Radius = (( Horizontal Res / Pixels per face ) * ( 360 / Angle )) / ( 2*Pi )
Assuming ppf of 5/20/40/100 for DORI
So a 2MP camera with 1600 x 1200 and 87 degree Horizontal angle would have a DORI of 210.8/52.7/26.3/10.5

My AT-LIR400 has 386.1/96.5/48.3/19.3

Does that seem more in line with what is expected?

Hi @PedroAsani

ppf = pixel per foot - a face needs about 60-80 pixel across the base iirc for a good ID image thus the 100 ppf guide. Dahua when they calculate DORI they use a smaller number which increases their distances.

I have not calculated the DOR part of DORI - just the ID part

A example of what I calculated before, and note I am using 1080P for a 2MP camera thus 1920x1080, with 1920 = H resolution

Dahua IPC-HDW4231EM-AS starlight fixed lens 1080p 1920x1080 w/2.8mm lens H 110° 10 feet 95.94 sqft
Dahua IPC-HDW4231EM-AS starlight fixed lens 1080p 1920x1080 w/3.6mm lens H 87° 12.65 feet 121.43 sqft
Dahua IPC-HDW4231EM-AS starlight fixed lens 1080p 1920x1080 w/6mm lens H 51° 21.58 feet 207.16 sqft

Remember this is in theory with effective pixels - compression, environmental conditions just reduce the quality of image capture

 

[PedroAsani]

Hi @PedroAsani

ppf = pixel per foot - a face needs about 60-80 pixel across the base iirc for a good ID image thus the 100 ppf guide. Dahua when they calculate DORI they use a smaller number which increases their distances.

I have not calculated the DOR part of DORI - just the ID part

A example of what I calculated before, and note I am using 1080P for a 2MP camera thus 1920x1080, with 1920 = H resolution

Dahua IPC-HDW4231EM-AS starlight fixed lens 1080p 1920x1080 w/2.8mm lens H 110° 10 feet 95.94 sqft
Dahua IPC-HDW4231EM-AS starlight fixed lens 1080p 1920x1080 w/3.6mm lens H 87° 12.65 feet 121.43 sqft
Dahua IPC-HDW4231EM-AS starlight fixed lens 1080p 1920x1080 w/6mm lens H 51° 21.58 feet 207.16 sqft

Remember this is in theory with effective pixels - compression, environmental conditions just reduce the quality of image capture

The wiki page makes three separate references to ppf meaning pixels per face, none of ppf meaning pixels per foot. Typo?

 

[mat200]

The wiki page makes three separate references to ppf meaning pixels per face, none of ppf meaning pixels per foot. Typo?

pixel per foot and pixels per face are 2 different measures, just like miles per hour and miles per gallon

In general, When referring to 100 ppf = we're referring to pixels per foot.

Could also be pixels per face - then the question is are we measuring horizontal pixels, vertical pixels, total pixels...

So, is the wiki incorrect? It's been awhile since I proof read it - when I did I made notes of any corrections that were needed.

Depends on the context of when pixel per foot is use and when pixels per face is used.

If using pixels per face - then you can not use that to calculate distance without some part of the equation having a face to measurement conversion of some sort.

 

[PedroAsani]

Yeah, it looks like the Cliff Notes on the wiki needs a correction. Here is the original text:

Challenging conditions: Situations with very varying or weak lighting. People, objects and vehicles are seen from an angle where details are in shade, or facing away from the camera. It could also occur in situations where people, objects and vehicles are moving at very high speed through an area. More often occurring in outdoor situations without additional lighting, or indoor situations during very dark conditions.

The above chart translates into the following Pixels Per Face (ppf) numbers for an average size face:
Identification (challenging conditions): 150 ppf [80 pixels across the horizontal of an average face]
Identification (good conditions): 75.6 ppf [40 pixels across the horizontal of an average face]

Inaxsys recommended the below DORI parameters in this document (Unknown Date - dead link)

Detect (8 ppf)
Observe (19 ppf)
Recognize (38 ppf)
Identify (76 ppf)

Given the Identification parameters above of 75.6 – 150 Pixels Per Face (ppf) noted above, 100 ppf represents a number in the middle, and makes the calculations easier. The DORI specs from Dahua seem to use a calculation other than 100 ppf for the Identification distance, so it may be best to make your own calculations. To achieve a 100 ppf identification goal, adjust the focus and placement of cameras used for
this purpose. Mount the cameras <8 feet to ensure a good angle for identification purposes. Measure the distance between the potential camera locations and the target area to determine if you are within the ID distance, and help in selecting the proper camera.

The equation to calculate the Radius (the distance in feet) for identification for 100 ppf as discussed above: Radius = (( Horizontal Res / 100 ppf ) * ( 360 / Angle ))/2*Pi

Here is what I would change it to:

Challenging conditions: Situations with very varying or weak lighting. People, objects and vehicles are seen from an angle where details are in shade, or facing away from the camera. It could also occur in situations where people, objects and vehicles are moving at very high speed through an area. More often occurring in outdoor situations without additional lighting, or indoor situations during very dark conditions.

The above chart translates into the following Pixels Per Foot (ppf) numbers for an average size face:
Detect: 7.2 ppf
Recognize: 38.4 ppf
Identification (good conditions): 75.6 ppf [40 pixels across the horizontal of an average face]
Identification (challenging conditions): 150 ppf [80 pixels across the horizontal of an average face]

Inaxsys recommended the below DORI parameters in this document (Unknown Date - dead link)

Detect (8 ppf)
Observe (19 ppf)
Recognize (38 ppf)
Identify (76 ppf)

Given the Identification parameters above of 75.6 – 150 Pixels Per Foot (ppf) noted above, 100 ppf represents a number in the middle, and makes the calculations easier. The DORI specs from Dahua seem to use a calculation other than 100 ppf for the Identification distance, so it may be best to make your own calculations. To achieve a 100 ppf identification goal, adjust the focus and placement of cameras used for
this purpose. Mount the cameras less than 8 feet high to ensure a good angle for identification purposes. Measure the distance between the potential camera locations and the target area to determine if you are within the ID distance, and help in selecting the proper camera.

7.5/20/40/100 is an easier DORI matrix to calculate.

The equation to calculate the Radius (the distance in feet) for DORI as discussed above: Radius = (( Horizontal Resolution / PPF ) * ( 360 / Angle ))/(2*Pi)

So for a 2MP camera with 1920x1080 resolution and an 87 degree horizontal angle, you would have a DORI of 168.68/63.25/31.63/12.65 feet

And if you can add a link to this calculator, that would likely help too. TheCalculation is already plugged in, so they can just edit HR, PPF and Angle.

 

[mat200]

Yeah, it looks like the Cliff Notes on the wiki needs a correction. Here is the original text:



Here is what I would change it to:



And if you can add a link to this calculator, that would likely help too. TheCalculation is already plugged in, so they can just edit HR, PPF and Angle.

HI @PedroAsani

Pixel Per Face I believe is correct in that context...