Re: Measuring angles in an image.
- From: ImageAnalyst <imageanalyst@xxxxxxxxxxxxxx>
- Date: Tue, 14 Apr 2009 08:26:40 -0700 (PDT)
On Apr 14, 7:50 am, "Stewart DIBBS" <s...@xxxxxxxxx> wrote:
"Geobird" <a1chan...@xxxxxxxxx> wrote in message
news:d34d04a4-fe76-4366-b06c-5889f5d725a1@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> Given that you have the distance from camera to the centre of the
image you are going to capture ( principle point) , is it possible to
measure the angle subtended by any element on the image with respect
to camera axis.
Depends entirely on the lens. You can probably work it out from the lens
specification, but the simplest method is set up a reference grid/ruler at
(say) 1m, set the lens to focus at infinity and see how many pixels the lens
subtends. This is OK with a single focus lens, you'd need to do multiple
measurements for different settings on a zoom lens.
After this "calibration" process, its simple to work out the angular
relationship of the image center pixel with any other pixel.
--
regards,
Stewart DIBBS
==========================================
PiXCL 8: The Advanced Image Analysis Suite
is affordable and available now.
==========================================www.pixcl.com
Stewart:
I don't see it. I can see how the spatial calibration depends on the
lens but I don't think the angle does. If you're looking at a flag
pole that is 100 meters away from your camera's lens, and the flag
pole is 10 meters high, then the angle is arctan(10/100) = 5.7
degrees. Ignoring things like optical distortions, how does that
depend on the lens? Also, no matter how complicated the lens is, the
angle on the field side of the lens will be the same as the angle on
the focus side of the lens, except in the case where the index of
refraction is different (for example, scene is underwater and sensor
is in air inside the camera). Now if your sensor was 2048 pixels
high, then it would be 1024 pixels from center to top, and simple trig
will get you the angle (which will be the same on both sides of the
lens) for any pixel in between. This would be
theta = arctan((p*maxY)/(maxP*d)),
where p is the pixel distance from the center to the point in
question, maxP is the number of pixels from the center to the top of
the image, maxY is the distance in the field from the optic axis to
the edge of the field, and d is the distance of the scene from the
camera. Changing the focal length of the lens will change the size of
the image and the distance of the image from the back vertex of the
lens but it will not change the angle. I'm pretty sure all you need
are the parameters I listed.
Regards,
ImageAnalyst
.
- Follow-Ups:
- Re: Measuring angles in an image.
- From: Stewart DIBBS
- Re: Measuring angles in an image.
- References:
- Measuring angles in an image.
- From: Geobird
- Re: Measuring angles in an image.
- From: Stewart DIBBS
- Measuring angles in an image.
- Prev by Date: Re: Measuring angles in an image.
- Next by Date: recognize small image inside larger images?
- Previous by thread: Re: Measuring angles in an image.
- Next by thread: Re: Measuring angles in an image.
- Index(es):
Relevant Pages
|