Issue. Existing GetBrightestObject() routine was unable to return the coordinates of Jupiter in Locate Image causing BrightStarLocate code in OB.CenterTarget() to not correctly centre Jupiter and caused the 1/4 frame 1x1 images to fail to capture Jupiter, wasting the opportunity and wasting time.

Description

In the past pictures of bright planets like Jupiter, Mars etc are taken by visually viewing images and instructing the telescope to make the necessary jogs to centre the planet prior to imaging.

With automated imaging the user isn't normally in a position to make the necessary jogs to centre the target.  A normal locate image used for derving a plate solution is typically of 5 -10s exposure. But this is much to long for a very bright target like Jupiter which only requires a locate image of 0.2s long to avoid strong blooming and subsequent ghosting in following images. However the exposure is then too short to capture any stars that are necessary for performing a plate solution.

One option is to simply slew to the target and not centre and to then take full frame 1x1 images knowing that Jupiter will be very probably caught someone in the middle 60% of the image.   This is wasteful in terms of the image file sizes and long delays between frames (1x1 full frame requires 8s download from SBIG ST-10XME camera)

First set of images of Jupiter was captured without centering and using half-frame pictures  (that reduce file size and download times by 25%).  However only 80-90% of planet disk was captured. One part of the disk was cropped off.

The next option is to use existing CentreOnBrightestObject method.   This calls GetBrightestObject() routine to search the Inventory arrays (produced when CCDSoft's Image.FindInventory() method is called) and return the XY coordinates of the brightest object.  This normally gives good results when centering a very bright star like Vega.

A second set of images of Jupiter were attempted using the CentreOnBrightestObject method and using quarter-frame pictures (that reduce file size and download times considerably).  However these failed to capture the planet disk at all.

Analysis
Investigating showed that whilst the GetBrightestObject() seemingly gave correct coordinates of Jupiter from first locate image, it didn't give correct coordinates for the second locate image, giving the coordinates of Jupiter's brightest moon instead.   Examing the second locate image we can see Jupiter with 3 of its moons visible (4th moon lay outside field of view)

Find Inventory show the following,   Jupiter and its 3 moons are found along with dimmer pseudo-star associated with the glare from Jupiter.



Examing the instrumental magntiudes (more negative numbers are brighter) we find that the brightest of Jupiter's moon has an instrumental magnitude that is brighter than the Jupiter, this is most likely because the aperatue used for the SExtractor analysis is set to a size suitable for finding & measuring point sources like stars and it not really suited for analysing large extended objects like the disk of Jupiter (50" diameter).

A larger apeture could potentially be used. Changing Source Extraction setup to aperture diameter of 60 pixels instead of 5 pixels changes instrumental mag of brightest moon from -14.5 to -14.9, but changes that of Jupiter from -13.7 to -17.8.   So Jupiter clearly become the brightest object.

Seeing from the observatory site is fairly variable. But on a good night of seeing the typical FWHM is around 2.8" .
Aperture diameter show be 4-5 times this in order to capture 99% of star flux (Basic Photometry :  https://community.dur.ac.uk/physics.astrolab/photometry_theory.html )

Let's take 4 times the FWHM.  So we need an aperture of 2.8 x 4.0 = 11.2". At plate scale of 1.40"/px (3x3 binned images) this corresponds to aperture diameter of 8 pixels.

Taking aperture of 8 and re-preforming Find Inventory changes instrumental magnitude of brightest moon changes from -14.5 to -14.8, whilst Jupiter changes from -13.7 to -15.0.  So Jupiter becomes brightest object but not by much.  

We need to use a larger aperture than 8 pixels.  With an apparent angular diameter of 50" at time of images then using an aperature of 160 pixels  (50 x 4. 5 / 1.40) seems overkill. We just need to find the brightest object not measure its total flux.  Taking instead the disk diameter for the aperture (50") plus the seeing related size (2.8" * 4.5) gives us  62.6" or 45 pixels (at 1.4"/px plate scale).  Using aperture of 45 pixels the instrumental magnitude of brightest moon changes from -14.5 to -14.9, whilst Jupiter changes from -13.7 to -18.3.  So Jupiter clearly becomes the brightest object and LocateBrightestObject centering code would therefore work reliably.

Actions

Change standard photometry aperture size to 8 pixels (from 5).   This might result in an offset between star magnitude estimates made pre- and post-change.
Create a routine that can change SExtractor setting for aperture size.  Implemented using new routine objConsole.SetDetectAperture()
Modify CentreBrightestObject to select aperture size of 45 pixels before centering Jupiter (or to an aperture diamter of 23 pixels before centering Saturn or Mars).  And to ensure that setting is returned to the standard size (8 px) afterwards.  OB.CentreTarget modified to achieve this.