First light with my new urban setup: the Rosette Nebula in SHO from a city rooftop

I’m very happy to share my first image captured with my new setup, which was designed from the ground up with a very specific goal in mind: imaging large, extended nebulae from a highly constrained urban environment, in this case the sloped roof of my building, right in the middle of the city.

The philosophy behind this setup is straightforward:

maximize signal on faint, extended targets despite heavy light pollution
reduce operational constraints as much as possible (meridian flip, cable management, accessibility)
fully exploit narrowband imaging (SII, Hα, OIII), while still keeping the option to experiment with some LRGB when conditions allow

The setup

📷 IMG_7180.JPG

For this configuration, I went with:

Optics: SQA85
Camera: ToupTek SkyEye62AM (IMX455 sensor)
Mount: Avalon M-Zero

The Avalon mount is a real game changer in this context. No meridian flip.

When imaging from a rooftop, often under mediocre or unstable conditions, removing one major source of complexity makes a huge difference. One less thing to manage is always a win.

The target: the Rosette Nebula

For this first light, I deliberately chose a very well-known nebula, making it easier to evaluate the performance of the setup and the quality of the data.

I simply let the system run as much as possible. There was a fair amount of discarded data due to frequent cloud passages, but enough usable signal to push the processing quite far.

Final integration

Hα 2.5 nm: 35 × 600 s
OIII 2.5 nm: 50 × 600 s
SII 2.5 nm: 64 × 600 s

That adds up to 25 hours total integration time, my personal record so far.

First impressions on the data

As expected:

Hα is extremely strong.
OIII shows a lot of signal in the central regions.
The real surprise came from SII.

There is a remarkable amount of faint SII extension, much richer than I initially anticipated. This quickly became the main challenge of the processing: bringing out those SII extensions without destroying the global balance or overprocessing the image.

I believe the final result is well balanced, but I’ll let you be the judge.

Processing approach for the Rosette SHO

A hybrid workflow inspired by multiple techniques

I published a detailed blog post on Deep Space Astronomy including:

a breakdown of the processing steps
and a download link for the calibrated SHO data, so anyone interested can try their own processing approach

Below is my complete workflow.

1. Preprocessing

WBPP on the SHO data

2. Gradient correction

Automatic DBE from Seti Astro
Gradients were corrected individually on each channel before any nonlinear transformation.

3. Linear noise and detail treatment

Still fully in the linear stage:

BlurXTerminator
NoiseXTerminator
StarXTerminator

One thing worth testing in the future is doing a light MAS stretch before StarX, as MAS sometimes seems to perform better when stars are still present.

4. Building the HOO image

4.1 Histogram stretch

MAS on Hα and OIII
- Background around 0.2
- Aggressiveness around 0.7

4.2 Hα / OIII combination

Perfect Palette Picker

4.3 Narrowband normalization

Lightness set to Hα
Synthetic green blend amount set to 0, resulting in a theoretically pure HOO image

4.4 Structure enhancement

HDR Multiscale
NoiseXTerminator

5. Dedicated SII processing

This is where most of the work happened.

BlurXTerminator, NoiseXTerminator, StarXTerminator
MAS, with the background pushed as dark as possible while preserving faint structures
StarXTerminator to keep the SII stars separate
NB Colour Mapper
- Hue: 47
- Saturation: 0.75
ImageBlend to progressively add the SII signal into the HOO image

6. Color work in Photoshop

Fine color balancing, local contrast adjustments and saturation control were done in Photoshop to refine the overall look while preserving the faint extensions.

7. Back to PixInsight: star recombination

NB to RGB Star Combination
Formula used:

~(~$T * ~$T_stars)

Tilt issues

There is some tilt present, clearly visible in the data.

The SkyEye camera includes a tilt adjustment ring, but manually dialing this in is far from trivial.

A potential solution would be an Electronic Tilt Adjuster from Wanderer Astro, which I’m seriously considering. This will likely be the subject of a future article.

Conclusion

This first image fully validates the concept behind the setup:

narrowband imaging from the city works
operational stability matters more than ever in constrained locations
and SII on extended nebulae is often underestimated

If you’re interested in processing, I strongly encourage you to download the calibrated data and try your own interpretation. Comparing approaches is one of the best ways to improve.

Clear skies.

Carlos Burkhalter