M104 A mosaic using images of the HLA

[Eddy Timmermans]

Hi,

I finally succeeded in putting together a mosaic of M104 using 18 images from the HLA.
I forst combined them to 6 RGB images using Pixelmath.
At first the orientation af the images didn't leave enough overlap to merge the images.
Also the last two images always seemed to have a different background. DBE didn't help to correct that.

So I cropped seperate parts of the images so there was enough overlap and merged them together.

This is the result : http://dl.dropbox.com/u/22839703/M104_Mosaic_DBE.jpg

The image is rather dark and I am still searching a proper way to lighten everythin up without brightening the background to much or vblowing up the core

Anybody have any ideas ?

Eddy

[Nocturnal]

Hi Eddy,

I think the image looks great, proper background/foreground proportions. Of course this is a matter of taste. You can use curves and the histogram tool to brighten the image. With the curves tool you can selectively only brighten bright areas. You could also first extract a Lightness image from the image, modify it as needed for your needs and then use that to shield the dark areas as you stretch.

Nice work on the mosaic!

[Eddy Timmermans]

Thank you.

The tutorials and processing examples here on this forus were of great help.

Eddy

[naavis]

Is the image supposed to be very blue and monochromatic? I think the color balance is very off. The mosaic is still very well done, and great data too!

[Eddy Timmermans]

I don't think it should be that blue.
Perhaps something went wrong during the merging of the Hubble images.

Now that I have the mosaic, I can focus on the processing.

Eddy

[pavel]

Hi Eddy!I ask you if it is time to describe in detail the process completely, is in the details ....In my very well!

[Eddy Timmermans]

Pavel,

I first downloaded all 18 images. I then combined the 3 channels of each image into 1 RGB image using PixelMath.
Because the images were tilted, I couldn't use StarAlignment directly. I had to crop each image to get rid of the blank parts. But by cropping them, I lost quite a bit of overlap between the image, so Star Alignment didn't find enough stars. So I cropped the sides of the images independently. This caused quite a bit of background abberation, but I was able to crop that away.
I then used DynamicAlignment and Gradient MergeMosaic to put together everything into 1 large picture.

I then cropped it again to eliminate the mosaic artifacts and was left with an image approximately 9500x3500 pixels.
I used DBE to try to get rid of some gradients, but it didn't make muche difference.

I then used HistogramTransformation and MMT. But the image stayed rather dark and blue.
I enhanced the contrast a bit so the equatorial dust band showed more detail.

Then I posted the image.

Yesterday I used Curves and Colorsaturation to change the color a bit, so it looks a bit more orange but not too much.
I will post it this evening.

I find it truly amazing that someone like me , with not much background in image processing and using PI for only 5 months, was able to use a complex program like PI and produce this image. I think it goes to show that by using this forum and the tutorials available on internet, anyone can do this. So the learning curve is not that steep at all. You just need to make a little effort of putting things together.
I would like to change my profile status from newbie to addict :-)

Eddy

[pavel]

Eddy!Thank you so much for the detailed and interesting answer! I, too, have had problems with creating mosaics from HLA,but your description of the process gives me a new positive impetus to proceed with HLA.Once again thank you for your quick and helpful reply!
P.S I think that all users of the PI will sooner or later become addicted PI, but it is a good drug!
Pavel

[Eddy Timmermans]

Pavel,

Putting together a mosaic using the Hubble data would be easier if PI would be able to crop images using a parallellogram instead of a rectangle. But most imagers produce rectangular images, so this feature would only be helpful for Hubble images.

In the HLA there is an enormous amount of data of M81. Theoretically it would be possible to create one gigantic mosaic of it, but I doubt my PC will be able to process it.

Eddy