New Script: DBXtract

rhusseingalindo

rhusseingalindo

Active member
Dual Band eXtract
The script allows you to extract the SII, HA and OIII signal from dual band filters to compose a SHO image

DBXtract_v02.PNG


Download script: https://dbxtract.astrocitas.com/

In color cameras, it's increasingly popular to use dual-band filters like the well-known Optolong L-Ultimate and Askar ColorMagic. However, extracting the specific signal from each emission band can be challenging.

This script enables the extraction of the emission bands of Ha, SII, and OIII, taking into account the contributions of each band in the RGB channels. This is done by considering the quantum efficiency (QE) parameters of different Sony sensors (IMX). Inspired in Dark Sky Geek youtube channel (thanks)

For example, the following graph represents the QE of the IMX 571 sensor. By drawing a line for each of the emission bands captured by the most common dual-band filters, we obtain that each RGB channel is represented by a percentage of transmission. Therefore, it is not correct to assign the Ha or SII signal natively to the R channel, nor the OIII to the G or B channels.

From these formulas, it's possible to isolate each of the Ha, SII, and OIII signals as part of an RGB combination. For example, HA = (R - 0,03*O3) / 0,80 or S2 = (R - 0,03*O3) / 0,73 to IMX 533 sensor.

formula.PNG

Finally, the OIII signal should be calculated after isolating Ha and SII in their respective B and R channels using this formulas:
formula_dev.png


An example of Askar ColorMagic C1 and C2, ZWO ASI 533mc-pro of NGC 7635NGC 7635, Bubble Nebula by Fran Masip

Figure 1. HO and SO images
dbx1.jpg


Figure 2. SII, OIII and Ha extracted signal
dbx2.jpg


Figure 3. SHO combination
sho.jpg


Figure 4. Final image
final.jpg


Download script: https://dbxtract.astrocitas.com/
Youtube video (spanish version):

DBXTract TO-DO List
v0.5 - New sensors imx094, imx455, imx662, imx462
v0.6 - Manual QE sensor
v0.7 - Astrometic copy
v0.8 - Integration SII/OIII method selection (avg, max)
v0.9 - Antlia SII y Hbeta
v1.0 - Pedestal correction
v1.1 - Icon and container execution
 
Last edited:
Hi Raúl,

Thanks a lot for this useful script. I was using process containers with pixelmath, but this is much better.

I'm a bit surprised that the factors for IMX533 and IMX571 are not the same, but really hadn't checked them. Good that you noticed that.

Regards,

Francisco
 
Thanks a lot Francisco :). The QE is very important to know the RGB contribution on an emission band, but it has some troubles in the OIII composition. Dark Sky Geek solved them manually for 533. Im working on solving it for any camera sensor and any object.
 
Very interesting! I will have to try it out this week. Would the L-Ultimate have any SII contribution? I always assumed it was effectively HOO.
 
TurtleCat said:
Very interesting! I will have to try it out this week. Would the L-Ultimate have any SII contribution? I always assumed it was effectively HOO.
Click to expand...
Thanks!

No, L-Ultimate is a dual-band Ha+OIII. The script can help you extract two images; Ha and OIII. Very useful when you want to use Ha as luminance or add Ha to your dark-nebula background or galaxies to highlight the Ha nebulas

2.jpg
 
rhusseingalindo said:
Thanks a lot Francisco :). The QE is very important to know the RGB contribution on an emission band, but it has some troubles in the OIII composition. Dark Sky Geek solved them manually for 533. Im working on solving it for any camera sensor and any object.
Click to expand...
Hi Raúl,

Do you now if the bandwidth of the dualband filters has a noticeable effect on the calculations?. I mean: Askar D1 and D2 compared to Askar C1 and C2. The light form the emission lines should be the same, but the continuum component is greater in the C1/C2 filters. I don't know if this could have an effect in the end result.

Thanks in advance.
 
fbitto said:
Hi Raúl,

Do you now if the bandwidth of the dualband filters has a noticeable effect on the calculations?. I mean: Askar D1 and D2 compared to Askar C1 and C2. The light form the emission lines should be the same, but the continuum component is greater in the C1/C2 filters. I don't know if this could have an effect in the end result.

Thanks in advance.
Click to expand...
Hi Francisco,

This script doesn't take that into account as it focuses on the exact position of the emission bands, no filter band-width. To solve it, i recommend obtaining an RGB image and using the "Subtract Continnum script". Below i show you what I did to extract the continnum from M33 Ha, extracted previusly from the L-eXtreme filter with DBXtract

Imagen1.png
 
rhusseingalindo said:
Hi Francisco,

This script doesn't take that into account as it focuses on the exact position of the emission bands, no filter band-width. To solve it, i recommend obtaining an RGB image and using the "Subtract Continnum script". Below i show you what I did to extract the continnum from M33 Ha, extracted previusly from the L-eXtreme filter with DBXtract

View attachment 23152
Click to expand...
Thanks a lot for the explanation
 
Thank you very much Raul for -currarte- working on the script and facilitating this part of the process that used to be a little doubtful for many people. Now that I'm making my first steps in narrow band with OSC camera you have been very helpful. Many thanks also for the Astrocitas channel to you and Teo. It has been of great interest to many of us.

As a detail I post here this image of M8 obtained last night and processed with DBXtract so you can see that we are taking advantage of it.
picture-1615f5ca61fce73dfa3beabb3fdbbb2b-fullhd-1300x1080.jpg


Best regards
 
monos said:
Thank you very much Raul for -currarte- working on the script and facilitating this part of the process that used to be a little doubtful for many people. Now that I'm making my first steps in narrow band with OSC camera you have been very helpful. Many thanks also for the Astrocitas channel to you and Teo. It has been of great interest to many of us.

As a detail I post here this image of M8 obtained last night and processed with DBXtract so you can see that we are taking advantage of it.
picture-1615f5ca61fce73dfa3beabb3fdbbb2b-fullhd-1300x1080.jpg


Best regards
Click to expand...
Thanks a lot, it is a pleasure for us to contribute to the astro-community. Good job with the M8 processing :)
 
Hi, thanks for this script.

A question - the final OSC image is debayered, which means some algorithm has modified the RGB values in each pixel based on its individual approach. So the R pixel has R plus some bleed for G/B pixels (as accounted in the equation) *plus* modification from the debayer algorithm.

Can this last bit be accounted for? Or is it not relevant?
 
Zerolatitude said:
Hi, thanks for this script.

A question - the final OSC image is debayered, which means some algorithm has modified the RGB values in each pixel based on its individual approach. So the R pixel has R plus some bleed for G/B pixels (as accounted in the equation) *plus* modification from the debayer algorithm.

Can this last bit be accounted for? Or is it not relevant?
Click to expand...
Thanks!

If I understood you correctly, it is just what the script does but taking into account that the matrix is RGGB. That is, it obtains what is necessary from RGGB to compose the emission band
 
Zerolatitude said:
Hi, thanks for this script.

A question - the final OSC image is debayered, which means some algorithm has modified the RGB values in each pixel based on its individual approach. So the R pixel has R plus some bleed for G/B pixels (as accounted in the equation) *plus* modification from the debayer algorithm.

Can this last bit be accounted for? Or is it not relevant?
Click to expand...
Since to stack OSC images it is a must to have debayerized them before stacking and since the script already works with stacked images I don't see much sense to the question.

In fact it is when the image is debayerized that we get the color information. And it is after having the color information and applying the chip characteristics that the script can do its work using pixelmath.

The only time we work with OSC images without debayerization is when we are calibrating them before stacking but just before stacking we must debayerize them or else we would get a false color information.
 
rhusseingalindo said:
Dual Band eXtract
The script allows you to extract the SII, HA and OIII signal from dual band filters to compose a SHO image
Click to expand...

Hello
Very nice script !
In future development, it could be nice to allow a special menu where we can enter manually b1, b2, b3, g1, g2, g3, r1, r2, r3 for people (like me) who use other sensor (I use DSLR, with IMX410 or also Canon DSLR). Maybe we could save these 9 values and recall them from a file...
 
CCD1024 said:
Hello
Very nice script !
In future development, it could be nice to allow a special menu where we can enter manually b1, b2, b3, g1, g2, g3, r1, r2, r3 for people (like me) who use other sensor (I use DSLR, with IMX410 or also Canon DSLR). Maybe we could save these 9 values and recall them from a file...
Click to expand...
Yes, im working in this new feature. Maybe the next week is finished. Thanks for feedback
 
You must log in or register to reply here.
Back
Top