field of view for Hikvision cameras

[rak313 0]

Hikvision specs for the DS-2CD2532 say '4mm@F2.0, angle of view: 75.8° (2.8mm, 6mm optional)' One assumes that is 75.8° for the 4 mm lens. But that does not seem to be consistent with the spec: '1/3" Progressive Scan CMOS'. 1/3" sensors are typically 4-4.5 mm wide.

 

A 4.5 mm wide sensor with a 4 mm lens should yield a FOV of about 59 deg.

 

Does anyone know the actual FOV with the DS-2CD2532 (or DS-2CD2032-I) for the various lens choices?

 

Thanks

[ssnapier 0]

It may be that they used the 2.8mm FOV on accident. Your 59 degree figure is pretty much spot on.

[MaxIcon 0]

The actual sensor sizes vary a bit these days, and the image circle of the lens will affect the FOV for a given sensor size. One 4mm lens will have a different image circle size than a different 4mm lens. Many Chinese cam manufacturers don't give accurate FOV data in their specs.

 

Hik also doesn't specify which resolution this is for (turns out it's 1080p). Most cams, 3MP would give a wider FOV, but on the Hik 2xxx series, they sample the output at 3MP and give a narrower horizontal FOV at 3MP.

 

I've measured the actual FOV on my 4mm Hiks, and here's what it is:

 

Hikvision DS-2CD2032-I Resolutions:

3MP = 2048 x 1536

1080P = 1920 x 1080

 

OEM 4.0mm lens - Hik's original lens

3MP = 69 degrees H

1080P = 77 degrees H

[rak313 0]

The actual sensor sizes vary a bit these days, and the image circle of the lens will affect the FOV for a given sensor size. One 4mm lens will have a different image circle size than a different 4mm lens. Many Chinese cam manufacturers don't give accurate FOV data in their specs.

 

Hik also doesn't specify which resolution this is for (turns out it's 1080p). Most cams, 3MP would give a wider FOV, but on the Hik 2xxx series, they sample the output at 3MP and give a narrower horizontal FOV at 3MP.

 

I've measured the actual FOV on my 4mm Hiks, and here's what it is:

 

Hikvision DS-2CD2032-I Resolutions:

3MP = 2048 x 1536

1080P = 1920 x 1080

 

OEM 4.0mm lens - Hik's original lens

3MP = 69 degrees H

1080P = 77 degrees H

Yes, there is no exact size for a 1/3" class sensor, but it will be less than 5 mm - or they would be claiming 1/2.5" sensor.

 

I believe that the specified FOV and/or lens focal length is not accurate. If you are getting 77 deg FOV with the 4 mm lens, this implies a 6.4 mm wide sensor - which would clearly be a 1/2" class sensor. FOV=2*atan((sensor_width/2)/focal_length). I doubt they would have a 1/2" sensor and not hype it, since it's twice the area of a 1/3" sensor.

 

Now if it were the 2.8mm focal length, with a 4.4 mm sensor width:

FOV=2*atan((4.4/2)/2.8 ) = 76.3 deg, which is pretty close to the claimed 75.8 deg FOV.

 

If you are saying that the image is cropped differently in 1080p vs the full 3 MP, so the FOV is corresponding different in the 2 resolutions - that makes sense.

 

I don't follow about different 4 mm lenses have different images circles (I certainly believe not all 4mm FL lenses are 4mm FL). Not even sure what you mean, but if 2 lens are both 4mm focal length, then the angle spanned by a given sensor size will be the same for both lenses. Maybe you are saying one lens can't image the entire sensor area. Even if that is true you are going in the wrong direction here. In order to get 76 degree FOV, you either need a much smaller focal length than 4 mm or a much larger sensor width than 4.4 mm.

 

I'm not doubting your measurements, I'm just trying to figure out whether to get the 4mm or 6mm camera.

 

Thanks for the data.

[MaxIcon 0]

Here's my post on how I measure FOV. The gear is inexpensive and easy to use:

http://www.cam-it.org/index.php?topic=5424.0

 

The image size issue isn't cropping, as you can't crop a directly sampled 1080p image into a smaller FOV 3MP image; you have to re-sample. On the Hik, either the 1080p or the 3MP image is re-sampled, and I've pulled images that indicate it's the 3MP image, but it's hard to know exactly what's going on at the sensor level.

 

The image circle is the size of the image projected when the lens is screwed in to its focus point. This will be different for different lenses, depending on the design.

 

It's more complicated for M12 lenses, as the back focal length varies from lens to lens, so the distance from the rear of the lens to the sensor (which directly affects the image circle size) will vary from lens to lens. The fact that Chinese lens specs are loose to nonexistent doesn't help to figure this out.

 

So, one 4mm lens will have a larger image circle than another, and a given sensor will capture a different portion of the image circle.

 

Even with CS/C mount lenses, which have a fixed back focal length, the image circle size varies between lenses. There are image circle size databases on the web for people who use certain types of camera systems.

 

What you'll see is if you put 4 different 4mm lenses from different manufacturers on the same cam, you'll get 4 different FOVs.

 

The only way to know for sure is to measure them.

[rak313 0]

Here's my post on how I measure FOV. The gear is inexpensive and easy to use:

http://www.cam-it.org/index.php?topic=5424.0

 

The image size issue isn't cropping, as you can't crop a directly sampled 1080p image into a smaller FOV 3MP image; you have to re-sample. On the Hik, either the 1080p or the 3MP image is re-sampled, and I've pulled images that indicate it's the 3MP image, but it's hard to know exactly what's going on at the sensor level.

Thanks for the link and method of measuring. Do you have any info on the "6mm" lens?

 

When the 3 MP sensor image is scaled to 1080p, they undoubtedly don't use the same set of pixels as when using the 2048x1536 mode - and this is what must be responsible for the change in field of view (thus me saying different cropping of the sensor). In fact they must be using a larger sensor width (versus 2048x1536 mode) in the 1080p mode to get a larger FOV. (I wouldn't be surprised if both modes were rescaled).

...

The only way to know for sure is to measure them.

I agree - and the reason for my original post.

Thanks for sharing your measurements.

...Original Post...

Does anyone know the actual FOV with the DS-2CD2532 (or DS-2CD2032-I) for the various lens choices?

Thanks

[MaxIcon 0]

I have a 6mm lens mounted now on a 2032, but didn't have time to measure it before putting it up a few weeks ago to watch a bird's nest. I took it down this weekend, and will measure it this week.

 

This is a random ebay 3MP 6mm lens, and I haven't done any testing on it to see if it's any good, so the results may not be very useful. Chinese no-name specs - who knows what they mean? This one didn't screw in as far to focus, and ended up hitting the glass when I re-assembled it. There was barely enough room to get the sensor board screws in well enough to hold it in place.

 

My other 6mm lens, from M12lenses.com, is still on the Dahua 2100, and I won't be taking that one down anytime soon, but the measurement post has the Dahua results.

 

Changing lenses on these compact cams, and M12 cams in general, is always a gamble. The manufacturers should sell different size replacement lenses for their cams, giving the actual FOV and such, but few of them do that.

[L4HikOEM 0]

See Attached Chart

[MaxIcon 0]

Generic charts like this are useful for rough estimates, but are not very precise. They assume the sensor is exactly 1/3" class (which is about 1/4" diagonal) and that the lens image circle is exactly the same size as the sensor, which is rarely the case.

 

In addition, this one appears to be for analog cams. An IP cam will have a different FOV at 1080p and 3MP, and sometimes at 720p and 1MP as well, depending on how the cam's firmware handles it.

 

Many cams maintain the same FOV at all the 4:3 resolutions and another one at all the 16:9 resolutions (because customers complain otherwise), but you can't have the same FOV at 1080p and 3MP without distorting one or the other.

 

Really, the vendors need to provide this level of detail directly on the cam's spec sheet. Hik provides it in a technical bulletin on their web site, but not in their cam specific web pages, last time I checked.

[bearclaw 0]

Ok, so the images may not be tuned at all, but it may give you a bit of an idea of the field of view. I have 3 Hikvision DS-2CD2532F-IWS one 2.8mm, one 4mm and one 6mm. I set them up very basically and held them in roughly the same spot, give or take a few inches and grabbed a capture from them. Its not the best picture or shot, but hopefully it better illustrates the differences in FOV for these cameras.

 

 

 

 

 

 

 

Hope this helps.

 

Cheers

[dkresge 0]

The dmesg output of my DS-2CD2032 reports "MT9T002 sensor ID is 0x2604". According to the datasheet, this is the former name of the Aptina Imaging AR0330.

 

Rev D. 11/1/10 "Changed part number from MT9T002 to AC0330"

Rev E. 12/17/10 "Changed part number from AC0330 to AR0330"

 

The sensor has a raw pixel array of 2304Hx1536V but basically operates in two modes (from which other resolutions can be derived):

 

2304x1296 at 30 fps (16:9 sHD)

2048x1536 at 25 fps (4:3, but -20% for EIS)

 

So it looks like the Hik FOV change from 1080P to 3MP could just be a function of which pixels are active + cropping and not due to any resampling on their part.

 

that said, it would be nice to have access to the full 2048x1536 mode as well since I generally don't go off-roading with my security camera

 

https://www.aptina.com/products/image_sensors/ar0330/.