RA scale inverted in some channels

[migueldvb]

This has happened in channel 70 of campaign 102. The mosaic created by k2mosaic seems to have the RA scale inverted as shown in the image below.

I'm not sure why for other channels the scale looks fine. This is a video for the channel 69 mosaic centered on the position of comet 67P that has the righ orientation for RA.

https://vimeo.com/209142702

[barentsen]

The wcs and coordinates that k2mosaic produces have not been well-tested, so I wouldn't be surprised to see a bug like this. The wcs is "scraped" from the pipeline-produced FFIs and is likely correct, so it must be the pixel coordinates that are transposed or inverted when k2mosaic writes the pixel arrays.

Kepler's readout gates are at opposite ends of each CCD, so my best guess is that this bug is present for half of all channels.

Unfortunately I do not have time to investigate or fix this bug in the near future. I'd be very happy to add you as a co-author of k2mosaic if you find the fix?

[migueldvb]

@barentsen I will look at this issue in the next couple of weeks. Testing these channels in other campaigns they have the same bug so I think your guess is right.

[migueldvb]

The inverted RA and Dec scale is also present in the FFI frames for some channels. Each channel has a given orientation depending on the readout but the row index is always Dec and the column index always RA as shown in Fig D-1 in the manual.

I think this issue could be fixed when the mosaics are written to frames by checking the values of the coordinates in the WCS and flipping the array if necessary.