How to Edit Out "Wanderers" Across Multiple Nights

[RabidNelson]

Hi,

I recently imaged the same widefield target on two separate nights, however, a planet was in two separate positions in each night. When I image integrate the two nights, I ended up with odd artifacts in the positions where the planet was. Does anyone have any suggestions for how I could either edit how the planet completely or keep the planet in only one position? Thanks in advance.

Jonathan

[pfile]

probably what's going on is that there's a gaussian shape to the brightness of the planet, so during integration with rejection some pixels are not being rejected, as they are too near the value of real signal. that leaves a funny shape behind.

RickS outlined a technique for "zapping" satellite trails with pixels valued 1.0 - this overwrites any "weak" pixels that look like signal, and pretty much guarantees that the whole thing will be rejected.

i think in this case you could do something similar with clonestamp. first you'd need to make an image that's all white. you can use the "NewImage" process and set the default color to white, or just use the default of black and then immediately invert the image.

next open up CloneStamp. the first thing you need to do is click in the target window. then after that, you need to command-click in the white window to pick a source point - anywhere will do. then go back to the sub with the planet and paint over the planet with the mouse. be sure to click the green check mark in CloneStamp to commit the operation. then repeat as necessary to all the subs, save them out, and integrate them.

could be a pain in the butt if there are a lot of subs, though.

rob

[RickS]

Thanks, Rob.  It had never occurred to me that the source for a CloneStamp operation could be a different image.  I can think of some good applications for that!

Cheers,
Rick.

[RabidNelson]

Rob,

Thanks for your response. I followed your suggestion, despite the large number of subs I had. I found it worked, but not that well. I used a single giant dot of white to cover up Saturn in each of the subs from the first night. When I stacked using Winsorized Sigma Clipping, I ended up with an off color circular patch of sky where Saturn was on the first night. Also, I underestimated how long the diffraction spikes were, so around that patch were tiny spikes. Because the first night was such a large chunk of my data, I decided there wasn't going to be a way to make that patch of sky look "normal". I ended up doing a Linear Fit Clipping which resulted in there being two Saturns in my image. Most people would have no idea what they're looking at unless I pointed it out, and this result is a lot truer to reality than the Clone Stamp attempt or my GIMP clone tool cleanup.

Here's my final image:
http://www.astrobin.com/252010/B/

I'd be willing to entertain any further suggestions or feedback. Thanks again.

Jonathan