Versions Compared

Old Version 38

changes.mady.by.user Mikhail Dubrovin

Saved on

compared with

New Version 39

changes.mady.by.user Mikhail Dubrovin

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  • acq-signal-amo01509-r0125-ave-wfs.txt- contains signal averaged arrays, which can be presented by full scale and zoomed plots:

Example for module ImgAlgos::NDArrAverage

  • See Module ImgAlgos::NDArrAverage

  • The NDArrAverage module can be used for evaluation of averaged pedestals or background using dedicated runs.
    Typical configuration file may looks like this:

    No Format
    # Run this script:
# psana -c psana-meca1113-r0045-cspad-cspad-dark-hotpix.cfg

[psana]
files = exp=meca1113:run=45
events = 400
#skip-events = 0

modules = CSPadPixCoords.CSPadNDArrProducer \
          ImgAlgos.NDArrAverage \
          ImgAlgos.Tahometer


[CSPadPixCoords.CSPadNDArrProducer]
source       = MecTargetChamber.0:Cspad.0
inkey        = 
outkey       = cspad_ndarr
outtype      = int16
is_fullsize  = yes
print_bits   = 3


[ImgAlgos.NDArrAverage]
source       = MecTargetChamber.0:Cspad.0
key          = cspad_ndarr
avefile      = cspad.0-ave
rmsfile      = cspad.0-rms
maskfile     = cspad.0-msk
hotpixfile   = cspad.0-hot
thr_rms_ADU  = 10
thr_min_ADU  = 4
thr_max_ADU  = 10000
print_bits   = 29


[ImgAlgos.Tahometer]
dn         = 100
print_bits = 7
  • Evaluation of average intensity in 2 or 3 stages using gate-based algorithms excludes out-layers in intensity spectra and makes average more stable. However, the gate width is not a universal parameter. In order to set this parameter one has to look at spectrum of averaged intensities for particular device.  The same is valid for evaluation of hot/bad pixel masks. One has to plot spectra of averaged intensity and rms values. Averaged intensity and rms spectra for exp=meca1113:run=45 are shown on plots:

...

  • thr_rms_ADU = 10  – to discard very noisy pixels,
  • thr_min_ADU = 4  – to discard presumably dead pixels with 0-intensity,
  • thr_max_ADU = 10000 – to discard pixels with intensity significantly exceeding average value.

 

Example for module ImgAlgos::ImgAverage

  • See Module ImgAlgos::ImgAverage
  • The ImgAverage module can be used for evaluation of averaged pedestals or background using dedicated runs. Typical configuration file may looks like this:
Code Block
bgColor#F7F7ED
[psana]                                                               
files   = /reg/d/psdm/<instrument>/<experiment>/xtc/<file-name>.xtc
modules = ImgAlgos.CameraImageProducer \
          ImgAlgos.ImgAverage
events  = 1000

[ImgAlgos.CameraImageProducer] 
source          = DetInfo(:Opal1000)
key_in          = 
key_out         = img
subtract_offset = true
print_bits      = 1

[ImgAlgos.ImgAverage] 
source          = DetInfo(:Opal1000)
key             = img
avefile         = img-ave.dat
rmsfile         = img-rms.dat
print_bits      = 31
evts_stage1     = 100
evts_stage2     = 100
gate_width1     = 200
gate_width2     = 20

 

Example for module ImgAlgos::ImgMaskEvaluation

...

Note

In this example parameters were chosen in order to get a small number of "noisy" pixel just due to statistics.

  • Typical image:

            Image Modified

  • Plots for fraction of noisy pixels:

Plots for fraction of saturated pixels:

Image ModifiedImage Modified

Masks: noisy, saturated, and combined:

Image ModifiedImage ModifiedImage Modified

Example for module ImgAlgos::ImgCalib

...

  • Example of the mask file and resulting image:

Example for module ImgAlgos::ImgRadialCorrection

...

  • Note: the option of the background subtraction (ImgAlgos.CSPadBkgdSubtract) is commented out in this configuration file . In order to evoke this option, the comment sign (#) should be removed from the list of modules and the key=bkgd_subtracted_arr should be used in CSPadPixCoords.CSPadImageProducer.
  • Calibrated image and spectrum:
  • Calibrated and radial-corrected image, spectrum, and subtracted r-phi65 distribution for n_phi_bins=65:

Calibrated and radial-corrected image, spectrum, and subtracted r-phi12 distribution for n_phi_bins=12:
Image ModifiedImage ModifiedImage Modified

Example for module ImgAlgos::ImgPeakFinder

...

  • This algorithm consumes ~15 ms/event on psana0101 for full Opal1000 (1024x1024) camera image.
  • Smearing algorithm use a "safety margin" which is currently set to 10 pixels (offset from each boarder of the full image size).
  • Image on different stages of this algorithm:
  • raw image,
  • image in the window with amplitudes above the threshold_low
    • few peaks at the edges were discarded by the window limits,
    • image still contains many 1-photon pixels, which need to be eliminated,
  • smeared image,
  • raw image with found peaks (marked by the red circles)
  • zoom of the previous plot.

Example for module ImgAlgos::ImgPeakFilter

Code Block
bgColor#F7F7ED
[ImgAlgos.ImgPeakFilter] 
source         = DetInfo(:Opal1000)
key            = peaks
threshold_peak = 5
threshold_total= 0
n_peaks_min    = 10
print_bits     = 11
fname          = img
selection_mode = SELECTION_ON

Example for module ImgAlgos::ImgPeakFinderAB

Code Block
bgColor#F7F7ED
[psana]                                                               
files   = /reg/d/psdm/<instrument>/<experiment>/xtc/<file-name>.xtc
modules = ImgAlgos.CameraImageProducer \
          ImgAlgos.ImgPeakFinderAB
events  = 10

[ImgAlgos.CameraImageProducer] 
source          = DetInfo(:Opal1000)
key_in          = 
key_out         = img
subtract_offset = true
print_bits      = 1

[ImgAlgos.ImgPeakFinderAB]
source                 = DetInfo(:Opal1000)
key                    = img
key_peaks_out          = peaks
#key_signal_out        = signal-arr
#hot_pix_mask_inp_file = ana-misc-exp/mask.dat
#hot_pix_mask_out_file = noise-mask-out.dat
#frac_noisy_evts_file  = noise-frac-out.dat
evt_file_out           = tmp/img-

rmin                =    10
dr                  =     1
SoNThr_noise        =     3
SoNThr_signal       =     3
frac_noisy_imgs     =   0.9
peak_npix_min       =     3
peak_npix_max       =   100
peak_amp_tot_thr    =     0.
peak_SoN_thr        =     4.
event_npeak_min     =     5
event_npeak_max     =  1000
event_amp_tot_thr   =     0.
nevents_mask_update =     0
nevents_mask_accum  =    50
selection_mode      = SELECTION_ON
out_file_bits       =    15
print_bits          =   513

Results:
Image ModifiedImage ModifiedImage Modified

Example for module ImgAlgos::ImgHitFinder

  • See Module ImgAlgos::ImgHitFinder
  • ImgHitFinder in regular mode needs in file with pedestals (offset) to correct the image and file with threshold.
  • In amo74213 run 93 these files can be obtained directly from data, discarding signal hits as outliers using
  • ImgAlgos.ImgAverage module as follows with configuration file:

...

Code Block
bgColor#F7F7ED
psana -c psana-amo74213-r0093-opal-img-hit-finder.cfg

Which creates the file: img-sum-result-r0093.dat for thre_mode = 2 and thre_mode = 3, respectively:
Image Modified Image Modified

Example for module ImgAlgos::ImgSpectra

...

  • This script can be used in order to produce text files with image and spectral array:
  • or graphics for several images:

Example for module ImgAlgos::ImgSpectraProc

...

Code Block
bgColor#F7F7ED
[info:ImgAlgos.ImgSpectraProc] Spectral array shape =3, 1024
[info:ImgAlgos.ImgSpectraProc] Image spectra for run=0060 Evt=000100
Column:       0     100     200     300     400     500     600     700     800     900    1000
Signal:    1211    4062   11150   17070   16406   12949    7991    5168    3968    3542    3811
Refer.:     933    3485   10425   17128   17791   13522    8315    5000    3390    2967    3193
Diff. :   0.259   0.153   0.067  -0.003  -0.081  -0.043  -0.040   0.033   0.157   0.177   0.176
[info:ImgAlgos.ImgSpectraProc] Run=0060 Evt=000100 Time=20120507-125420.982421325 done...

Example for module ImgAlgos::ImgSaveInFile

Code Block
bgColor#F7F7ED
modules = ... ImgAlgos.ImgSaveInFile:1 ...

[ImgAlgos.ImgSaveInFile:1]
source         = DetInfo(:Opal1000)   # or CxiDs1.0:Cspad.0
key            = img
fname          = my-img
#ftype          = txt
#ftype          = bin
#ftype          = png
ftype          = tiff
#eventSave     = 5
saveAll        = true

Example for module ImgPeakFinder and ImgPeakFilter for CSPad

  • Module ImgAlgos::ImgPeakFinder works on image. In order to apply this algorithm to CSPad the image should be produced. In next example the image is produced using consequtive modules cspad_mod.CsPadCalib, ImgAlgos.CSPadMaskApply, and CSPadPixCoords.CSPadImageProducer:
Code Block
bgColor#F7F7ED
[psana]
files   = /reg/d/psdm/<instrument>/<experiment>/xtc/<file-name-1>.xtc \
          ...
          /reg/d/psdm/<instrument>/<experiment>/xtc/<file-name-N>.xtc

events      =  1000
experiment  = cxii0212
calib-dir   = ana-cxii0212/calib

modules = cspad_mod.CsPadCalib \
          ImgAlgos.CSPadMaskApply \
          CSPadPixCoords.CSPadImageProducer \
          ImgAlgos.ImgPeakFinder \
          ImgAlgos.ImgPeakFilter \
          ImgAlgos.ImgSaveInFile:1

[cspad_mod.CsPadCalib]
inputKey      = 
outputKey     = calibrated
doPedestals   = yes
doPixelStatus = no
doCommonMode  = yes

[ImgAlgos.CSPadMaskApply]
source            = DetInfo(CxiDs1.0:Cspad.0)
inkey             = calibrated
outkey            = masked_arr
mask_fname        = <your-local-directory>/<mask-file-name>.dat
masked_amp        =  0
print_bits        =  5
mask_control_bits = 15

[CSPadPixCoords.CSPadImageProducer]
calibDir       = /reg/d/psdm/<instrument>/<experiment>/calib
typeGroupName  = CsPad::CalibV1
source         = CxiDs1.0:Cspad.0
key            = masked_arr
imgkey         = img
print_bits     =  0
#tiltIsApplied  = true

[ImgAlgos.ImgPeakFinder] 
source         = DetInfo(CxiDs1.0:Cspad.0)
key            = img
peaksKey       = peaks
threshold_low  = 2
threshold_high = 5
sigma          = 1.5
smear_radius   = 5
peak_radius    = 7
xmin           = 20
xmax           = 1700
ymin           = 20
ymax           = 1700
#testEvent      = 5
print_bits     = 3
#finderIsOn     = true

[ImgAlgos.ImgPeakFilter] 
source         = DetInfo(CxiDs1.0:Cspad.0)
key            = peaks
threshold_peak = 5
threshold_total= 0
n_peaks_min    = 10
print_bits     = 11
fname          = cspad-img
selection_mode = SELECTION_ON

[ImgAlgos.ImgSaveInFile:1]
source         = CxiDs1.0:Cspad.0
key            = img
fname          = cspad-img
#eventSave     = 1
saveAll        = true

Example for module ImgAlgos::CSPadArrAverage

...

  • Images of the CSPad arrays for averaged and rms values, respectively, in one of the CXI runs:

Example for module ImgAlgos::CSPadBkgdSubtract

...

  • The file with the background array, bkgd_fname, was obtained by averaging 1000 events using module CSPadArrAverage. Subtraction is done with normalization for norm_sector=0.
  • Event image and pixel amplitude spectrum before and after the background subtraction are shown in plots:

  • Other event with better subtracted background:

Example for Module ImgAlgos::CSPadMaskApply

  • See Module ImgAlgos::CSPadMaskApply
  • The array for mask contains zeros and ones for masked and passed pixels, respectively, and has a shape of full-size CSPad array 4*8*185388.
  • For example, it can be generated by the command

...

  • for the averaged background amplitude array <input-background-cspad-arr-file-name> abtained as a result of ImgAlgos::CSPadArrAverage module.
  • Plots show the averaged background, and the mask arrays generated from this background for three thresholds 10, 20, and 30 EDU:

    The best results in filtering can be achieved in combination of modiles:
Code Block
bgColor#F7F7ED
modules = cspad_mod.CsPadCalib \
          ImgAlgos.CSPadBkgdSubtract \
          ImgAlgos.CSPadMaskApply \
          ...
[cspad_mod.CsPadCalib]
inputKey      = 
outputKey     = calibrated_arr
doPedestals   = yes
doPixelStatus = no
doCommonMode  = no

[ImgAlgos.CSPadBkgdSubtract]
source        = DetInfo(CxiDs1.0:Cspad.0)
inputKey      = calibrated_arr
outputKey     = bkgd_subtracted_arr
bkgd_fname    = ana-cxi49012/cspad-cxi49012-r0025-background-ave.dat
norm_sector   = 0
print_bits    = 0

[ImgAlgos.CSPadMaskApply]
source            = DetInfo(CxiDs1.0:Cspad.0)
inkey             = bkgd_subtracted_arr
outkey            = masked_arr
mask_fname        = ana-cxi49012/cspad-cxi49012-r0025-mask-40.dat
masked_amp        =  0
print_bits        =  3
mask_control_bits =  1

where

  • cspad_mod.CsPadCalib - subtracts the pedestals from raw CSPad data,
  • ImgAlgos.CSPadBkgdSubtract - subtracts the background,
  • ImgAlgos.CSPadMaskApply - apply the mask.
  • In the test with images for background represented by the water and solvent rings this filter provides the background suppression factor about 100.
  • The background images that still pass this filter have significantly larger intensity with respect to averaged background:

    Input parameter mask_control_bits allows to control masking regions of 2x1. For example, if all edges need to be masked, then use mask_control_bits = 15, which gives image array like:

    where red regions/lines of pixels of amplitude=8 are masked.

Example for module ImgAlgos::CSPadArrNoise

...