LCSB Galaxy Detection

Overview:
The purpose of this operation is to recover low central surface brightness galaxies (and nebular objects ) that might have been missed by the pipeline Galworks processing. All bright stars (SNR > 15) found by PIXPHOT are masked off (from JHK-band images) using a radius large enough to ensure that no flux remains in the coadd from them. All galaxies found earlier are masked off similarly.

The masked-off coadd is then block average, first 2 X 2, then 4 X 4 and finally 8 X 8. A filter is then applied to the blocked image that detects all sources with a peak-pixel flux above an N-sigma threshold, where N is tied to source density; for the low density case, N = 4.0, and sigma refers to the background noise of the blocked/smoothed image.

Detected sources are then processed: additional parameterization, photometry (fixed circular, isophotal, petrosian and flux-growth), and extraction.

Components:

The low central surface brightness galaxy processor consists of two main components: detection and photometry. By definition LCSB objects have a very "soft" profile; that is, their central surface brightness is generally lower than the GALWORKS detection threshold (3-sigma). Consequently, in order to detect/recover LCSB objects, it is necessary to do some kind of block averaging and/or smoothing of the coadd images to reduce the net background noise. Since computational efficiency is crucial to the design, we have chosen simple block-averaging and boxcar smoothing of the images (as opposed to gaussian convolution for example).

The coadd images are first masked of stars and galaxies detected in the primary GALWORKS processor. Note: It is necessary to mask all bright stars (SNR > 15) and detected galaxies, but it also important to not mask out previously detected (but faint) sources -- potential LCSB candidates.

The cleaned images are then block-averaged and the resultant images boxcar smoothed. The modified images are then searched for local maxima sources (described below).

Preliminary testing reveals that three sets of block averaging are optimal: 2 X 2 (for compact objects), 4 X 4, and 8 X 8 (for very extended objects). The processing order begins with the 2 X 2 kernel, from which all objects are detected, analyzed, written to disk and blanked from the coadd image. The 4 X 4 is then processed accordingly, followed by the 8 X 8.

Detection/Analysis Steps:


Source Detection and LSB Selection

Given: the input image has been block averaged and boxcar smoothed. Mean stats are computed: mean, median and sigma. Candidate detections are local maxima (defined by radial area = 3 pixels). Detection threholds:

Photometry and Parameterization

For each LCSB candidate, several parameters are computed. First it is necessary is determine the position of the source peak on the coadd (recall that the sources are detected using a blocked & smoothed modified image). With the source position center determined for the LCSB candidate, the next step is to compute the peak pixel surface brightness and the central (radius=5) surface brightness. Other parameters computed: "shape" or "sh" score (see GALWORKS SDS), integrated signam to noise (circ radius=10) with the inner 10 pixels blanked, and the "super" coadd 2X2, 4X4 and 8X8 blocked/smoothed peak SNR values.

Photometry is performed after neighboring sources (stars and galaxies) are subtracted from the coadd. Fixed aperture photometry is performed using aperture radii = 5, 10, 20, 30, 40, 50, 60, 70, 80. The total signal to noise ratio is computed using the radius = 10 flux and area. A petrosian radius and flux magnitude, isophotal magnitudes and flux-growth magnitudes are all computed in similar fashion to normal galaxies.

LCSB Flowchart