NGC 6412 - Arp 38


NGC6412, SBc galaxy in Draco, 11.6 mag, 2.4' diameter, North up, East on the right.

Subimages: 101 frames x 300s (8.5 hrs) SD Mask combined after full calibration. 20 other frames were discarded due to changes in focus.
Scope: C11XLT, EFL=2936mm (f/10.5 effective)
Mount: Losmandy G11 with Gemini L4
Camera: ST7EI (0.63"/pixel) set to -15C and -10C (click for sample cropped frames to observe the difference 5C makes).
Guiding: SX716 camera on a Taurus Tracker III OAG
Acquisition and Processing: MaxIm DL/CCD v4.5

Galaxy imaged with over three weeks, 2006/07/01, 07/15, 07/18 and 07/19. This object is listed as peculiar (low surface brightness companion on arm) by Arp due to what appears to be a ring in the spiral arm on the North side. Perhaps the bright knot is what is left of the nucleus of a captured and devoured galaxy. Under extremely high stretching an outer halo to the north-east is apparent, I wonder if this is the remnant of the captured companion's spiral arms.

Even though the field of view is only 8'x6' the low surface brightness of the galactic disk (it is only ~50 counts above bgd) required me to flatten the background. Somewhat arbitrarily I decided on 07/19 to continue imaging the same target as the previous night, this was a lucky break since the extra thirty useful exposures made a tremendous difference after the image was deconvolved (see below for a comparison).

I wrote a small VB program to script MaxIm to calculate the FWHM of a star for each image and then output a sorted list. Creating a histogram helped determine the rejection criteria, which turned out to be FWHM>=2.6".

The deconvolution settings were adjusted to balance the amount of sharpening and the creation of artifacts e.g. Gibbs Rings seen as darks haloes around bright stars. I used the Lucy-Richardson method with 11 iterations. The gaussian PSF option was used, but I reduced the automatically retrieved value from 1.75 to 1.2 (~30% reduction). This eliminated the haloes at the cost of a slightly less sharpening (FWHM 1.5" versus 1.36" for unreduced PSF). The next processing step was to apply a gamma stretch to bring out the detail in the spiral arms and preserve the galactic core. To do this I used the DDP command without a low-pass filter. After the deconvolution and the aggressive non-linear stretch some artifacts that were hidden at low intensities became visible around the brighter stars. In addition the huge difference in intensity between the brightest stars and the galaxy nucleus (more than 100x!) imply an aggressive contrast setting which leaves the stars looking like, as someone else once put it, cutoff tree stumps. So I wrote a MaxIm plug-in to i) find the stars in the image, ii) round off the edges of the stumps near a cutoff value and iii) optionally erase the halo with the local background.


DigitalSkySurvey image (48inch Schmidt camera) scaled to match my image

Arp catalog image (Palomar 200inch)

Without the "fixed" stumps and tainted slightly yellow for kicks.



Strangely, the 72 subimage composite show little difference from deconvolution. However deconvolving using identical settings (same star for PSF generation, same number of iterations) produce noticeably better results witht the 101 subimage composite.

Pre-convolution Post-convolution
Images FWHM (") SNR FWHM (") SNR
72 2.0 5600 2.1 6600
101 2.2 6500 1.4 7100

Download the calibrated, combined (101 subimages) and background flattened pre-convolution FITS data file.

72 subimages (5min each) Lucy-Richardson deconvolved
101 subimages (5min each) Lucy-Richardson deconvolved