Linear defect problem

cbradshaw

Active member
I took about 180 1-minute images of the Cat's Eye nebula with my Canon T3 and have a linear artifact in my green and red pixels. I tried linear defect detection and subtraction, but it had no effect.

catsEye_158m_drz.jpg


The crosshatch in the background is from my attempts of removing horizontal and vertical defects. Any suggestions?

Regards,

Charlie
 
from the look of it the background pattern seems like it's leftover hot pixels after calibration + no dithering, or unguided imaging.

if that is the case, you can inspect the calibrated subs and see if they have hot pixels; one reason for the leftover hot pixels is dark optimization during light calibration. so you can try without optimization and see if the hot pixels still remain.

another strategy is to tune the pixel rejection parameters to try to get rid of the residual hot pixels during integration.

and another would be to use CosmeticCorrection to try to clean up the hot pixels in the calibrated frames before registering them.

rob
 
from the look of it the background pattern seems like it's leftover hot pixels after calibration + no dithering, or unguided imaging.

if that is the case, you can inspect the calibrated subs and see if they have hot pixels; one reason for the leftover hot pixels is dark optimization during light calibration. so you can try without optimization and see if the hot pixels still remain.

another strategy is to tune the pixel rejection parameters to try to get rid of the residual hot pixels during integration.

and another would be to use CosmeticCorrection to try to clean up the hot pixels in the calibrated frames before registering them.

rob
I dithered every second frame and guiding was good. I've tried various correction strategies, but so far no good. There are hot pixels, which were not removed by my first calibration run. These, I believe were from some kind of bleed over in the debayer routine. I tried with super-pixel and the hot pixels did not affect my registration, but the linear defects still show up. I only see the defect in the green and red channels.
 
well there could still be differential flexure and/or field rotation due to polar alignment error leading to the curved pattern of hot pixels in the registered subs.

did you try calibration without dark optimization? are you also talking about the horizontal lines in the image? those are probably some other problem.

rob
 
There is a slight field rotation seen in the star alignment data. ~0.01 or less degree between images. I'm referring to the single large blue line, which is an artifact I've never seen before. I did try without dark optimization and the red channel was really poor. I'm currently trying a different and new process flow as an experiment, which is about the 5th time I've tried different calibration process flows. The week before with the same setup calibrated fine. Frustration!!

Charlie
 
Cats_eye_LRGB_small.png


This version looks better. I took the calibrated and debayered files, separated into RGB components and registered them all against a selected green frame. I then integrated all as a luminance and the separate RGB frames to create a synthetic LRGB image. It's still far from my best image, but the artifact is gone. I noticed that some of the frames failed registration, but have not tracked down why.

Thanks,
Charlie
 
this is a tough target... that is for sure. anyway good that you sorted it out but that artifact is weird; i have never seen such a strong banding before in a canon camera.
 
Back
Top