[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [APML] What's the Best Way to "Combine" Two ImagesofDifferingTimes?
Hi Wade,
Carlos pretty much answered the question. Let me just repeat it in a slightly different way.
You can measure the standard deviation in Photoshop. Depending on the PS version you
are using, this information would appear somewhere around the histogram window.
In each image, you select similar background area and read the standard deviation value
for each of them.
Suppose you have two images. Their stddev values are 50 and 25 (2:1). Then the inverse
square of the stddev is 1:4. Therefore in Registar, you put the weights 1 and 4 for these
images and average them. It will give you the best results in terms of shadow quality.
If you don't use Registar, you can also use the "normal blending" of Photoshop layers to
average the images. In the above case, suppose the first image (stddev=50) is the background
layer and the second image (stddev=25) is the upper layer. Then you give an opacity 80%
to the top layer and flatten the images. In this way, the top layer contributes 80% to the final
image and the background layer contributes the rest 20%. This is exactly 1:4 weighting.
Cheers,
Wei-Hao
On 9/20/05, Carlos Milovic F. <cmilovic@yahoo.es> wrote:
>>Then we can measure the standard deviation (rms) in the images and
use the inverse-squares
>>of the rms as the weights.
>What program can handle inverse-squares. Do you have a snapshot?
>I'm a little slow in visualizing where/what values need to be
>entered.
Hi Wade
PixInsight, throught the PixelMath process, has included the power operator, wich allows you to perform a square... Anyway, you don't need this feature. What you have to do is the following:
- Make a preview of a "pure" background area in every image, or in the object of mayor interest (a small section).
- With the Statistics tool calculare the standard deviation of the samples.
- Now copy this value in a paper and calculate the square. Inverse the result (1/x).
- In the PixelMath process, use those calculated values as the weighting factor for the propper channels/images.
- It is a good idea to check the Rescale box just in case the values go out of range. Remember that the 32bits acuraccy is distributed in the normalizated range as floating point numbers, not integers.
Good luck!
_______________________________________________
Astro-Photo mailing list
Astro-Photo@seds.org
http://seds.org/mailman/listinfo/astro-photo
--
________________________________________________________________
Wei-Hao Wang :)
Institute for Astronomy at University of Hawaii
Address:
2680 Woodlawn Drive Personal Website:
Honolulu, HI 96822 http://www.ifa.hawaii.edu/~wang
________________________________________________________________
_______________________________________________
Astro-Photo mailing list
Astro-Photo@seds.org
http://seds.org/mailman/listinfo/astro-photo