**IRIS TUTORIAL
Add a sequence of images**

The simplest method for combine images is the arithmetic addition. For example for adding the images i1, i2, i3, run the command

>ADD2 I 3

If you wish to normalize the result if the sum exceeds value 32767, run the command

>ADD_NORM I 3

The **ADD_MEAN** command compute
the mean of a set of images. For example:

>ADD_MEAN I 3

The **COPY_ADD** evaluate
a subset of images by adding group of images of the input sequence. Suppose
for example the images:

I1, I2, I3, I4, I5, I6

If the desired result is a new sequence J like:

J1=I1+I2, J2=I3+I4, J3=I5+I6

run the console command

>COPY_ADD I J 6 2

SI the desired output sequence is

J1=I1+I2+I3, J2=I4+I5+I6

run

>COPY_ADD I J 6 3

Note: the **COPY_MED** is similar,
but compute the median sum for each subset, see below.

You can also calculate the median of the sequence

>SMEDIAN K 15

The result is a single image in which each pixel is the median intensities for the same pixel in the images k1..k15.

A more global panel for combine
images is the **Add
a sequence** dialog box of the **Processing **menu. For example:

The **Add a sequence** command
is used for sum, median combine, statistical combine a set of images.

The *Arithmetic mode *provides
a simple addition (sufficient is most condition).

The *Median mode* is useful
when some pixels are deviants (cosmic ray, hot pixels). But the final signal
to noise ratio is lower comparatively to the arithmetic sum.

The *Min-Max mode* reject before
addition the maximal and the minimal values find for the sequence of images
and a given pixel.

The *Adaptative weighting mode*
is powerful for reject bad pixel. The method based on an analysis of local
noise of the image to stack and the process iterative. Normaly select no more
than 3 - 5 iterations:

The *Sigma clipping* is a rejection
method of non representive pixels values. When using this method, the pixel
value that deviates the most from the mean by more than S*igma
x Sigma coef* will be discarded from the
sum. *Sigma*
is the local standard deviation, and *Sigma
coef* is a parameter entered by the user.
Typicaly, *Sigma coef = 1 to 3*.
Decreasing this coefficient will reject more pixels.

The process can be iterative.

The *Sigma median method* is
similar to the *Sigma clipping*, but the rejected image pixel is replaced
by the median of the stacked image for the given pixel

**Tip:** Somewhat, you may notice that the background level is not
the same for all the image of the sequence. Before combine with tools like median,
sigma clipping, ... it is strongly
suggested to normalize the background to a same value for all images. The
bakground level of the in-memory image is returned by the command** BG**:

>BG

To normalize the background of a set of images, open the dialog box Offset normalisation of a sequence of Processing menu

The background of the output sequence j1, ...,
j8 is normalized to 100. Use the **BG** command for control.

You can also run the console command

>NOFFSET2 I J 100 8