Raw astronomical images (and even poorly processed images) don’t usually have a uniform ‘sky’ value over their surface prior to processing, see Sky value for a complete definition of the sky value. However, a uniform sky value over the image is vital for further processing. For ground based images (particularly at longer wavelengths) this can be due to actual variations in the atmosphere. Another cause might be systematic biases in the instrument or prior processing. For example stray light in the telescope/camera or bad flat-fielding or bias subtraction. The latter is a major issue in space based images where the atmosphere is no longer a problem.
SubtractSky is a tool to find the sky value and its standard deviation on a grid over the image. The size of the grid will determine how accurately it can account for gradients in the sky value. Such that if the gradient (change of sky value) is too sharp, a smaller grid size has to be chosen. However the results will be most accurate with a larger grid size.
|• Sky value:||Definition of the sky value.|
|• Tiling an image:||Defining a mesh grid on the image.|
|• Mask image:||How to mask pixels for no analysis.|
|• Invoking astsubtractsky:||Options and arguments to SubtractSky.|