Friday, January 1, 2010

Astrometry 101 - Pixel Scale

Every astrometry observer and measurer needs to know the pixel scale of their images. This is the angular width and height of a pixel. Pixel scale tells how much of the sky is covered by one pixel.  Many CCD cameras have square pixels (equal height and width). If the camera pixels are not square, then both dimensions musts be considered. Normally width is denoted by x and height y.

To calculate pixel scale, one needs telescope focal length and pixel width or height in microns.

pixel scale or sampling in arcseconds = (206.265) * (pixel size in microns) / (focal length in mm)


focal length = (206.265) * (pixel size in microns) /  pixel scale

The constant 206.265 converts microns to mm and radians to arc seconds.

A micrometer or micron is one millionth of a meter, or equivalently one thousandth of a millimeter (mm)

[360 deg / (2 pi radians)] x (3600 arc sec / deg) x ( 1 mm / 1000 microns) = 206.26480624709635515647335733078 ~= 206.265

or just about 206 as many use as rule of thumb.

This is a straightforward calculation if you know the two input parameters.

Focal ratio is the ratio of the focal length to telescope aperture

Focal ratio = focal length / aperture

For example, if focal length = 3048 mm and  aperture = 30.48 mm, Focal ratio = 10 or telescope is at f /10.
If you have 9 micron square CCD pixels, then your pixel scale is  206.265 x 9 / 3048  or 0.6 arc sec per pixel.

This has meaning when compared to your "seeing" measured in arc secs of FHWM (Full Width Half Maximum) of the stars in your images and is referred to as "sampling". You have near optimal sampling when your pixel scale is about half of the value of your seeing. If your pixel scale is less than optimal, the sampling is termed "under sampled". If your pixel scale is larger than optimal, the sampling is termed "over sampled".

Focal reducers reduce the effective focal length of your imaging setup and increase your pixel scale.
Binning your CCD camera imaging chip 2x2 (versus high resolution 1x1) doubles the size of each pixel and increases your pixel scale.

Barlow lenses increase the effective focal length of your imaging setup and decrease your pixel scale.

Astrometrica calculates these parameters for you if you can interpret the values provided after doing a plate solution. The default values of its configuration settings must be initially modified by a new user for their imaging setup. Thus, one must be able to calculate or estimate a focal length as an input parameter for this program.

This is  a copy of several lines in a recent Astrometrica.log file from my own astrometry:

The column labeled FHWM reports the seeing of the stars in the image and shows that sthe seeing is about 5 to 6 arc seconds. The pixel size is calculated at 2.58 arc sec or about half the seeing. I was using 2x2 binning on a SBIG ST-8XME CCD camera. There is a f/6.3 focal reducer a filter wheel and an SBIG AO-7 unit in the imaging setup, but due to the spacing of the components resulting in an effective focal ratio of near 4.7. This also gives a nice size field of view to capture more stars which is better for accurate astrometry. The SNR of the stars measured for astrometry is also given. It is important to have sufficient SNR for an accurate centroid calculation of stars and the object to be measured. The Fit RMS is the most important columns because it shows the bottom line of the astrometric calculation and the goal is to have fractional arc second accuracy is all measurements..

Pixel scale is used to determine how long of an exposure one can take of a moving object. That will be the subject of another topic.

The larger your pixel scale the longer you can expose your target per image.

The larger your pixel scale, you provide allow more photons from your target to fall on a pixel which improves the SNR of your target, improves the calculation of the target centroid in the image, improves astrometric precision, and results in smaller residuals.(only up to optimal)

Too high a pixel scale beyond optimal however, is not good for astrometry because stars will be undersampled and are not spread out over one or more pixels. It would be like trying to accurately measure the position of a fly with a ruler with only inch or centimeter tick marks. and is of no astrometric value no matter how reference stars you have imaged..


IAU Minor Planet Center Guide to Minor Body Astrometry:

1 comment:

  1. A good blog always comes up with a new or excited information.I really appreciate about all the information that you have been mentioned above.Good to read about the pixel scale and all the factors that you elaborated here.To calculate dpi of screen or screen density also the pixel per inch you can use this PPI Calculator,it shows the instant results.