calculating pixel dimensions

If you know how many pixels there are on a CCD or CMOS and the resolution
of the output, can you determine the pixel dimensions?

On Tue, 20 Feb 2007 01:13:57 GMT, "TMA" wrote:

If you know how many pixels there are on a CCD or CMOS and the resolution
of the output, can you determine the pixel dimensions?

How do you define "the resolution of the output"? Unless this is some
odd spec provided by the camera manufacturer, then no, you can't
determine anything about the pixel spacing or pixel size from the output
alone.

You can determine the size of the sensor and the spacing of the pixels
if you know the focal length of the imaging optics, and have a reference
image of known size (such as the Moon, or a set of stars with known
spacings).

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

On Feb 19, 6:13 pm, "TMA" wrote:
If you know how many pixels there are on a CCD or CMOS and the resolution
of the output, can you determine the pixel dimensions?

Normally the physical linear dimensions of an individual pixel are
found by looking by the detector or camera in manufacturer's
specifications and reference is not made to the angular size of
objects resolved by the chip. If you are unable to find a
manufacturer's spec table on the internet, Berry and Burnell's HAIP
book (sec. 1.4.2) describes the following procedure. We assume you
know the image size in pixels. It is usually reported in by your image
capture software. Assuming the image size is the pixel count (e.g.
640x480), given the pixel count and taking a physical measurement of
the chip size with a ruler or micrometer, you can compute the pixel
size. The following is the HAIP example:

d_pixel = d_ccd / N

For a webcam with a 4mm x 3mm chip and 640x480 pixels, the width of a
pixel is:

d_pixel = 4mm / 640 px = 0.00625 mm/px = 6.25 microns.

The height of a pixel is:

d_pixel = 3mm / 480 px = 0.00625 mm/px = 6.25 microns.

The example chip has square pixels. Some cameras have rectangular
pixels - their aspect ratio (height/width) is not 1.0. For example, I
use a DSI II. It has rectangular pixels.

For some chips, pixels are binned togther in groups. In the example
chip, 4 pixels might be binned together - using hardware or software -
to form 25 micron squares.

- Canopus56

You can determine the size of the sensor and the spacing of the pixels
if you know the focal length of the imaging optics, and have a reference
image of known size (such as the Moon, or a set of stars with known
spacings).


what do you mean reference image of known size? The moon is 1/2 '

On Tue, 20 Feb 2007 13:50:08 GMT, "TMA" wrote:

what do you mean reference image of known size? The moon is 1/2 '

Right. So if the 1800 arcsecond Moon covers 582 pixels, and you are
using a 400 mm focal length lens, you can calculate the pixel spacing p
(in millimeters):

p = ((1800 arcsec)(400 mm)) / (206265 * 582 pixels) = 0.006 mm.

Keep in mind that the Moon varies in size, and it can be hard to
precisely measure its width in pixels. If you are after a precise
measurement, use a pair of stars of known separation. Also, note that
the pixel spacing isn't the same as the pixel size. Pixel size can't be
determined from anything other than chip specifications. Finally, be
aware that many sensors have different pixel spacings horizontally and
vertically, so you should measure both.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

"Chris L Peterson" wrote in message
...
On Tue, 20 Feb 2007 13:50:08 GMT, "TMA" wrote:

what do you mean reference image of known size? The moon is 1/2 '

Right. So if the 1800 arcsecond Moon covers 582 pixels, and you are
using a 400 mm focal length lens, you can calculate the pixel spacing p
(in millimeters):

p = ((1800 arcsec)(400 mm)) / (206265 * 582 pixels) = 0.006 mm.


Where does the 582 pixel value come from?

On Thu, 22 Feb 2007 17:49:27 GMT, "TMA" wrote:

Where does the 582 pixel value come from?

It's just an assumption in this example that the image of the Moon is
582 pixels in diameter. In general, the pixel spacing is given by

p = a * F / 206265 / d

where

p = pixel spacing
a = angle between reference objects in arcseconds
F = focal length of the optics (in the same units as p)
d = number of pixels between reference objects

BTW, 206265 is the number of arcseconds in a radian.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

"Chris L Peterson" wrote in message
...
On Thu, 22 Feb 2007 17:49:27 GMT, "TMA" wrote:

Where does the 582 pixel value come from?

It's just an assumption in this example that the image of the Moon is
582 pixels in diameter. In general, the pixel spacing is given by

p = a * F / 206265 / d

where

p = pixel spacing
a = angle between reference objects in arcseconds
F = focal length of the optics (in the same units as p)
d = number of pixels between reference objects

BTW, 206265 is the number of arcseconds in a radian.


Grand!

On Feb 22, 10:49 am, "TMA" wrote:
Where does the 582 pixel value come from?

When you take an image, how many pixels are on each side in your
picture. Most images report their pixel size with a right click in
Windows. Or open the image up in MS-Paints and look at "Image size".
- Canopus56

"canopus56" wrote in message
ups.com...
On Feb 22, 10:49 am, "TMA" wrote:
Where does the 582 pixel value come from?

When you take an image, how many pixels are on each side in your
picture. Most images report their pixel size with a right click in
Windows. Or open the image up in MS-Paints and look at "Image size".
- Canopus56


YOu Da MAn!!...