Processing the whole pack!
During the weekend I found out that the dome flats are "consistent with each other" but not very good to correct the image large scale variations and that the twilight flats are totally bad! They really suck! One flat is not consistent with the next one ... none of them present a good correction to large scale variations of the image ... there's no way you can combine them and make them work.Solution to be tried: combine the pix-to-pix variation of the dome flats with large scale variation estimated by the science frames themselves. Let's see if it work.
In that sence I started to process the whole package of images ...
Also some bureaucracy to the projects around here ... and a bike flat tire! :-)
posted by Cris Da Rocha @ 11:40 PM
2 Comments:
There's only one issue: It is possible that some gradients in the science images can be additive, not multiplicative. The SDSS data were bad for several years until we realized that "sky flats" were being contaminated by smooth reflected/scattered light. In principle the additive and multiplicitive components of the mean variations in the science frames can be separated if you have images with a wide variety of exposure times, but in practice this is difficult.
Thanks for your comments about fossil groups! I am going to read those papers.
Unfortunately, in answer to your question, I don't have a reference. We had to figure it out in SDSS by modifying the large-scale flat (not the small-scale, pixel-to-pixel variations) until we got consistent "overlaps" where ccds ended up overlapping each other in the scan strategy. If you have lots of redundancy in your data you might be able to do this. But it's not easy (and it goes beyond what most astronomers consider for reasonable photometry). My only advice is that if you don't do it, beware gradients with position on your detector, such as subtle color gradients.
Post a Comment
<< Home