Tests of hsi_overres_eval.pro using flare testdata

;+
FUNCTION HSI_OVERRES_EVAL,CBE,NX,DMAX,ftmap,NODISPLAY=nodisplay,$
                          PIXEL_SIZE=pixel_size,VERBOSE=verbose,  $
                          STATS=stats,TITLE=title
; PURPOSE:
;   Tries to evaluate a HESSI calibrated event list to determine which
;   subcollimators over resolve the source.
;
; METHOD:
;   Starts with detector dmax, makes a Fourier transform map
;   (hsi_phasor_sum), selects the half-power contour of the source,
;   then goes to the finer subcollimators, making FT maps and 
;   determining the ratio of the maximum inside the contour 
;   to the maximum outside the contour.
;   Usually, if this ratio is less than one, the subcollimator
;   overresolves the source, but noise, erronious pointing, or other
;   factors can throw it off.  Therefore the default is to show the
;   single-collimator maps.

Examples for testdata flares 1-9

FLARE 1 Single-Collimator FT Maps

Output of hsi_overres_eval:
FLARE 1 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1      1.33643
       2      1.97347
       3      2.37062
       4      2.05639
       5      2.11248
       6      1.90198
COMMENT:
This flare is a point source, and the program correctly finds a ratio > 1
for all single-collimator maps.

Compare the above with the results of Back projection:

   INSIDE/OUTSIDE RATIO OF MAXIMA
       1      2.31373
       2      3.23918
       3      3.34304
       4      2.47523
       5      2.89087
       6      1.90433

FLARE 2 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 2 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1      1.67966
       2      2.28242
       3      2.36724
       4      2.36474
       5      2.35560
       6      1.91350
COMMENT:
This flare is also a point source, and the program correctly finds a ratio > 1
for all single-collimator maps.

FLARE 3 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 3 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1      1.08606
       2      1.26030
       3      1.30649
       4      1.38147
       5      2.07780
       6      1.24848

COMMENT:
This flare is a weaker point source, and the program correctly finds a
ratio > 1 for all single-collimator maps.
 

FLARE 4 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 4 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1     0.936529
       2      2.56635
       3      2.13025
       4      2.01656
       5      1.40083
       6      1.88308   

COMMENT:
Also a point source, but for some reason the coarsest map shows an offset
of the source from that of the finer-collimator maps.  (The cause is being
investigated.)

FLARE 5 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 5 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1     0.278255
       2      1.02931
       3     0.710092
       4     0.736607
       5     0.362777
       6      1.95015
COMMENTS:
This flare has only a 0.5-second duration, so all the maps are poor.  Nothing
should be expected from hsi_overres_eval in this case.

 

FLARE 6 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 6 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1     0.816310
       2     0.887275
       3      1.06506
       4      1.46256
       5      1.14944
       6      1.88405
COMMENTS:

This is a double source where the components have sizes of 5 and 10
arcsec, so collimators 1 and 2 should over resolve both
components.  The outputs succesfully indicate this.

 

FLARE 7 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 7 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1      1.06818
       2     0.936754
       3     0.477178
       4     0.723433
       5     0.939746
       6      1.91462

COMMENTS:
This flare has two 4 x 20 arcsec components, and should be over resolved
by detectors 1,2,3 and 4, but the output shows ratios of 1.06 and 0.94
for detectors 1 and 2.  These values are clearly wrong.  Looking at
the maps for collimators 1 and 2, there doesn't seem to be anything of
significance in the contours.

This is an example of how the human eye can make a better determination
than a programmed single-number statistic.

Does anyone have a suggestion for how a program could make a better
assessment?

FLARE 8 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 8 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1     0.934892
       2      1.06527
       3     0.824776
       4      1.10071
       5      1.41920
       6      1.90997

COMMENTS:
This flare is two 5-arcsec Gaussians.  The output rightfully finds
collimators 4, 5 and 6 not over resolving, but says that collimator 2
is also not overresolving (ratio=1.06527).

Again, the single-ratio statistic fails, and a better statistic is needed.

FLARE 9 Single-Collimator FT Maps

Output of hsi_overres_eval:

FLARE 9 PHASOR SUM MAPS, dets 1-6 INSIDE/OUTSIDE RATIO OF MAXIMA
       1      1.06612
       2      1.98707
       3      2.89170
       4      2.24095
       5      2.26897
       6      1.88859

COMMENTS:
This is a pair of 1-arcsec Gaussians, and, as desired, the program finds them
not over resolved by all the collimators.

The program and included functions:

Some calibrated event lists (idlsave files):