Thanks for all the nice comments. I have been busy going slowly more mad/insane than ever, trying to master PixInsight. The only way to learn is by trial and error. Be prepared for more errors than you can imagine. The results in the end when you finally win, are worth all the pain.
The image in the original post was stitched together with RegiStar. All previous processing done with PI. Even 3nm NB filters suffer from nasty gradients at the 3.5x3.5 degree field of the RH200. This means it is almost impossible to make seamless mosaics.
I have worked out a protocol with DBE to almost totally eliminate gradients.
This image was made with Gradient Merge Mosaic in PixInsight 14MB
http://d1355990.i49.quadrahosting.co..._4P_NOO_PI.jpg
The lack of seams is obvious.
If you think the above image is stunning I am now working on a Hubble Palette version of the same data plus SII of course that is even better.
Here is panel one at full stacked resolution ie x1.5 sensor size. 15MB
http://d1355990.i49.quadrahosting.co...VSNR_P1_HP.jpg
Hubble Palette is of course SII to red. NII or HA to green and OIII to blue. It is then adjusted using this method
http://bf-astro.com/hubblep.htm
The very dim detail in red and blue is vastly enhanced because human vision is far more sensitive (x20) in the green part of the visible spectrum compared to red or blue.
The simplest way to think of it is that the 'red' data modulates green to give a yellow signal and the blue data modulates green to give a cyan signal.
Bert
Exposure details. Each filter had about 20x16min or 5 hours exposure for each panel. So 4 panels X three filters x five hours = 60 hours of data at F3. To collect equivalent data at the same resultant signal to noise at F5 would take more than 160 hours.