Rum multidrizzle/astrodrizzle

Choosing the drizzle.scale Parameter

In general, the full-width at half-maximum of a PSF in the final image should be around 2.5 pixels for a well-sampled output image. In most cases choosing an output pixel size equal to 0.6 or 0.5 the input pixel works well with HST instruments.

For observers who generally have three to four dither pointings with 0.5 pixel shift increments, the drop size should be larger than the output pixel size. (for example, for WFPC2, scale=0.05 arcsec and pixfrac=0.6) - remember, the scale is now given in arcseconds while the pixfrac is relative to the input pixel.

Choosing the drizzle.pixfrac Parameter

One must therefore choose a drop size that is small enough to avoid degrading the image, but large enough so that after all images are drizzled the coverage is reasonably uniform. The drop size is controlled by a user-adjustable parameter called pixfrac, which is simply the ratio of the linear size of the drop to the input pixel (before any adjustment due to the geometric distortion of the camera).

One test of the choice of drizzle.scale and drizzle.pixfrac is to measure the r.m.s of the weight image away from the edges of the output image (preferably under a region of sky). The r.m.s. of the final drizzle weight image should be less than 20% to 30% of the median (use imstat to get these numbers). A larger r.m.s. suggests that the weights are so variable that the ignoring the weight image when doing photometry will add noticeably to the final photometric errors.