Bright Stars in the MSX Field

T. Jarrett, IPAC

The msx field presents a particular problem for bright star blanking because there are so many of them! The galactic coords for the first coadd in the field (scan 041) is glong = 32.8, glat = 1.06., and for the last coadd in the scan, glong = 38.1, glat = 3.8 deg.

The following images demonstrate the overhelming shear numbers of bright stars. Three different cases are considered for star blanking: (1) the nominal method, used for low stellar density scans, (2) threshold for blanking increased by 1.0 mag, and (3) threshold for blanking increased by 1.5 mag. What we mean by the latter two cases is -- effectively -- changing the flux of each star by 1.0 to 1.5 mag fainter; thus we blank less stars as well as relax the blanking radii, etc. Our motivation for doing case (2) or case (3) is that with the nominal method (case (1)) so many stars are blanked that the coadd is pretty much useless thereafter for any kind of extended source extraction. On the other hand, we do not want to leave ourselves vulnerable to false detections due to bright stars.

Bright Star blanking operations include:

The horizontal stripe blanking is performed only when the star is very bright, except for the "middle" strip which seems to appear for any bright star. The number of persistence ghosts that are blanked (including along adjacent coadds) is controlled by the input flux of the star.

For more details of bright star blanking, see GALWORKS Bright Star Cleansing


The following images show four panels, the first panel is the raw coadd image, the second is the coadd after bright star blanking using the "nominal" criteria (case 1, above), the third is the coadd after bright star blanking where we have made the stars fainter by 1.0 mag (case 2, above), and the final panel is the coadd after bright star blanking where we have made the stars fainter by 1.5 mag (case 3, above). Each image is about 300Kb.

First coadd, glong = 32.8, glat = 1.1

Although the field is full of bright stars, K < 9, J < 10, with a 5th mag star near the center right edge of the coadd, the diffraction spike and ghost features associated with bright stars are rather difficult to discern form the image. The reason being the confusion noise has swamped this diffuse light.

For the nominal case in which we apply blanking criteria tuned for low stellar density regions (in which the diffuse light components are easily detected) it can be seen that nearly all of the coadd is blanked away! (although things are not so bad at J, due to the extinction). If however, we change the thresholds for blanking, or effectively make each star appear fainter, than the situation gets much better. At least from these images, it would seem that we should follow suit and modify the thresholds when the stellar density is very high.

Last coadd in scan 041, glong = 38.1, glat = 3.8

Comments?