Earthshine blog

Earthshine blog

"Earthshine blog"

A blog about a system to determine terrestrial albedo by earthshine observations. Feasible thanks to sheer determination.

Moon halo in R, G and B

Exploring the PSFPosted by Daddy-o Aug 28, 2013 02:42PM
We have indications that the halo seen in B and V images are somewhat similar, but that VE2 images of the Moon have decidedly different halo profiles. Because the filters used in B, V and VE2 are fabricated (and function) differently we may have the situation that this causes the differences in halo profiles. B and V are 'coloured glass' filters while VE2 is a thin-film filter. In a DSLR detector the R,G and B filters are essentially bits of coloured semiconductor material. As a test we inspect the halo seen around the Moon in images taken with a DSLR camera. In such images the R,G and B channels are obtained at the same time at precisely the same observing conditions.

On the internet (http://bit.ly/13YZIOU) we have found a very detailed large-scale JPG (i.e. 8 bit only) 10-min exposure image of the sky near Orion containing the Moon. We submitted the image to nova.astrometry.net and received a solution back, including image scale. We took the resulting WCS-equipped image and extracted the profile of the lunar halo in R, G and B and plot these against radius from the estimated disk centre. We have subtracted a sky level for each colour, estimated by eye, and show the profile in the upper panel and repeat it in the lower panel with straight lines fitted:

The profiles are saturated out to about half a degree but after that they follow a remarkably similar shape in this log-log plot. Fitting power law functions (1/r^alfa), we may even see some sort of 'straight line behaviour' between radius 0.6 and just short of 2 degrees, with another linear trend taking over out until the sky-noise is reached at 4 degrees or so. The slopes are fitted-by-eye only but are -2.9 (near the canonical 'can-never-be-steeper-than-3') value, and -1.3.

Before assuming that 'DSLR RGB imaging' will solve all our filter problems, let us recall that the VE2 vs other profile differences we have found are subtle; that the above is based on an 8-bit image; that nothing is known about the image treatment performed by the author of that image, and that we do not have data similar to VE2 here - the R channel is not VE2 [ ... but read more here link !]. Let us instead try to do something similar with 14-bit RAW images.

Note that the above image was obtained by tracking the sky. Apparently, tracking allowed a long exposure that gave us a wide halo to study. In our own DSLR-On-the-sky wide-field images we have failed to get results similar to the above - exposures were limited to several seconds to avoid overexposure and trailing. Our own MLO telescope images are restricted to the 1/2 degree reach from disc centre.

Note that the above essentially speaks very well for the sort of DSLR optics used - the amount of scattered light near sources must be low or we would not see so steep a PSF! Again, this may be easier to do with a wide-field lens such as used for the above image - trhings may be different with a tele-lens that allowed a closeup of the Moon. The 'core' of the PSF should be better investigated, if we can get some of our own images of e.g. Jupiter or bright stars - the above image is VERY wide-field and contains zillions of crowded stars, not likely to give us good point-source PSFs.












  • Comments(1)

Fill in only if you are not real





The following XHTML tags are allowed: <b>, <br/>, <em>, <i>, <strong>, <u>. CSS styles and Javascript are not permitted.
Posted by Chris Flynn Sep 02, 2013 05:10AM

Great stuff!
I will try this on the weekend with a deeper bit camera -- I'll have access to it only on t he weekend. Really interesting to see that to first order the profiles are so similar. Deeper bits on the CCD and use of raw images should test this out very well... interesting if this truly simultaneous imaging option works much better in terms of the PSF profile matching...