Re: Cosmetic hot pixel removal in autostacks?
Posted: Sat Oct 10, 2020 8:38 pm
Hi Tim,
thanks for the additional files – I've been looking into this further and I have found a problem with the hot pixel handling that leads to the dark pixels/trails in the output - this was related to a fix I did some time back where the starburst noise on some of the earlier high resolution CMOS cameras was being interpreted as hot pixels. Unfortunately the fix is stopping hot pixel detection from working correctly for your images. I'm now trying to work out a way to fix both problems at the same time
The bright pixels (coloured pixel/coloured trails) are different. In the most recent set of data I can see that the dark frame was captured at a temperature of -4.9 and the light frames were captured at a temperature of 1.7 C. Looking at one specific hot pixel, in the light frame, the dark frame and the final stacked, I see this:
The Pixel The Light Frame The Dark Frame As you can see I get a pixel readout of about 7000 in the light frame, but only about 4000 in the dark frame. I'm fairly sure that a big chunk of this difference is down to the temperature difference between the dark capture and the light capture, which highlights the need to get the temperature the same between light frames dark frames. Mind you, if I get the hot pixel code working as it should, it will remove this particular blemish – even at the lower temperature the pixel is about 80 standard deviations above the mean!
I will try to get a fix for the hot pixel removal into the next update of SharpCap 3.3.
Cheers, Robin
thanks for the additional files – I've been looking into this further and I have found a problem with the hot pixel handling that leads to the dark pixels/trails in the output - this was related to a fix I did some time back where the starburst noise on some of the earlier high resolution CMOS cameras was being interpreted as hot pixels. Unfortunately the fix is stopping hot pixel detection from working correctly for your images. I'm now trying to work out a way to fix both problems at the same time
The bright pixels (coloured pixel/coloured trails) are different. In the most recent set of data I can see that the dark frame was captured at a temperature of -4.9 and the light frames were captured at a temperature of 1.7 C. Looking at one specific hot pixel, in the light frame, the dark frame and the final stacked, I see this:
The Pixel The Light Frame The Dark Frame As you can see I get a pixel readout of about 7000 in the light frame, but only about 4000 in the dark frame. I'm fairly sure that a big chunk of this difference is down to the temperature difference between the dark capture and the light capture, which highlights the need to get the temperature the same between light frames dark frames. Mind you, if I get the hot pixel code working as it should, it will remove this particular blemish – even at the lower temperature the pixel is about 80 standard deviations above the mean!
I will try to get a fix for the hot pixel removal into the next update of SharpCap 3.3.
Cheers, Robin