IRIS

Procedure for co-aligning SJI images

This describes my attempts at co-aligning IRIS SJI images. A key point to note is that when you read a SJI file, the xcen and ycen values of the images do not vary smoothly: they contain jitter. So, before attempting to cross-correlate image, it is best to first align the images by interpolating them using the xcen and ycen values. This is done with:

IDL> outmap=iris_sji_align(filename [, imagen=, xrange=, yrange=])

where the output is an array of IDL map structures. Making a movie from this output (with movie_map) will show that there is still jitter in the sequence, and so it's necessary to cross-correlate between neighboring frames.

Study of SJI sequences - case study

For the 3-Oct-2014 22:00 data-set (4 hours) I took the 1400 SJI images (1152 frames) and created a map using:

IDL> read_iris_l2, filename, index, data
IDL> index2map, index, data, map    ; this takes a long time!
IDL> mapx=sdo_align_map(map, xrange=[130,190],yrange=[-160,-100])

This last routine (which I wrote for SDO) takes a map and interpolates the individual frames using the xcen, ycen coordinates. So it does not perform any co-alignment. The movie produced in this way is fairly smooth, but there's clearly drift and jitter.

The data-set also includes 2832, and I repeated the same steps

1. Use read_iris_l2 to read the data sequence. Use imagen=[i0,i1] to only read a subset of the images.
2. Extract a sub-array for region of interest, using interpolation to correct jitter in pointing values (equivalent of sdo_align_map), and using dustbuster to remove dust.
3. Perform coalignment between pairs of frames using cross_corr.pro.
   

Data analysis notes

Wavelength calibration