Background on images

mask = segmentation_map > 0 # Sextractor's segmentation mask with DETECT_THRESH  =  1. and ANALYSIS_THRESH =  1.
bkgmask = (data == 0.0)  # Mask of the outer parts of the image without data (blank part)
mesh_size=[50, 100, 150, 200, 300, 500]
percent = [1, 7, 10, 15, 25, 26]
sigma_clip = SigmaClip(sigma=3.)
bkg_estimator = MedianBackground()
bkg = Background2D(data, (mesh_size, mesh_size), filter_size=(5, 5),mask=mask,coverage_mask=bkgmask,sigma_clip=sigma_clip,bkg_estimator=bkg_estimator,exclude_percentile=percent,fill_value=0.0)

Left: background_model Middle: Meshes drawn on the background subtracted image Right: Smoothed bkg subtracted image with Gaussian stdev=5 (icl100.png)

Background from boxes (1D)

Background is the mean value from boxes 32 x 32 pix large, inside which all objects are masked. Also, the mean standard deviation is estimated from the boxes. There are several cases plotted in the radial profile plot:

• bkg = image - mean,
• bkg1m = image - (mean-stdev),
• bkg1p = image - (mean+stdev)
• base = image
• iclsub[mesh_size]