Overview of the system, modules, parameters and functionality

1
FEDRA status and progress
Valeri Tioukov, INFN, Napoli

• Overview of the system, modules, parameters and
  functionality

2
The purpose of FEDRA is the emulsion data analysis

• To provide the library of methods and tools for
  emulsion data processing
• To simplify the possibility of finding new
  effects in emulsion data
• To be easily upgradeable with new functions and
  modules
• Possibility of the interactive analysis

3
Maximum of freedom for usersdevelopers

• Full data set must be available with the
  transparent access (root files)
• Raw data
• Final data
• Intermediate data
• Functions and libraries should be available in
  the interactive session
• Batch processing driven by users cards

4
Framework development cycle of the new
function(s)

• Interactive analysis of the data set (root
  session)
• Script programming and testing of new functions
• Class and/or functions development
• Including into libraries
• Testing and release

5
Emulsion scanningstorage structures

• View the basic structure with
• Header
• frames (images)
• clusters
• segments
• Run all data of one scanning session
• Run header
• View tree
• Fragment (SySal only) rectangular region
  scanned together with some common settings

6
Data reconstruction structures

• EdbSegP segment with some additional attributes
• EdbPattern 2-dimensional pattern of segments
• EdbVolume 3-dimensional, array of patterns
• EdbDataSet keep information for the set of data
  to be reconstructed together (brick)
• EdbLayer geometrical attributes of one emulsion
  layer (could be associated to pattern)
• EdbScanCond parameters like resolutions, etc
  which depends on the emulsionSS conditions

7
Essential operations

• Link Pattern to pattern linking I.e.
  segment-segment association
• Alignment pattern to pattern affine
  transformation calculated for the case of maximum
  number of associations found
• tracking recognition of chains of segments
  satisfied to linking criteria

8
Recset front-end for batch processing

• recset -l -a -t -f ...
  input_data_set_file
• -ccd - to remove ccd defects
  (update par/xxx.par file)
• -l - link up/down
• -ang - correct up/down angles
  offset and rotations
• -a - plate to plate
  alignment
• -f - fine alignment based on
  passed-throw tracks
• -tn - tracking
• -nu - suppress the update of
  par files

9
Data-set structure

• Dataset.def structure definition file
• Runs.lst - runs list
• Default.par common parameters setting
• rawdata/ - raw runs corresponding to runs.lst
• ./par/ - individual per peace parameters
  d-base
• ./data/ intermediate data (per peace)
• Tracks.root - output

10
Parameters overview layer

• Convention for layer ID
• 1 and 2 emulsion surfaces (patterns consists
  of microtracks)
• 0 - emulsion plate (patterns consist
  of basetracks)
• Reference system for 1 and 2 in-plate RS
• Reference system for 0 in-brick RS
• ZLAYER 1/2 0./210. 0. 0.
• SHRINK 1 .89
• AFFXY 1 1 0 0 1 0 0
• AFFTXTY 1 1 0 0 1 0 0
• ZLAYER 0 1300.
• AFFXY 0 1.000636 0.000099 0.000529
  1.000098 -43.073853 -47.092899

11
Parameters overview ScanCond

• Convention concerning 1,2,0 is the same as for
  layers
• "Prototype_OPERA_microtrack"
• SIGMAGR 1 .5 .5 3.
  sigmaX sigmaY sigmaZ for one grain
• SIGMA0 1 .5 .5 .010 .010
  sigmaX sigmaY sigmaTX sigmaTY
• DEGRAD 1 5.
  angular degradation sig(tx) sig(0)(1degtx)
• BINS 1 5 5 4 4
  binsigma acceptance
• RAMP0 1 8 12
  puls(Signal 1) puls(Signal 80) at 0 rad angle
• RAMP04 1 6 9
  puls(Signal 1) puls(Signal 80) at .4 rad angle
• CHI2MAX 1 3.5
  chi2 simple
• CHI2PMAX 1 3.5
  chi2 with puls weighting (ramp used)

12
Parameters overview ScanCond

• "Prototype_OPERA_basetrack_300micron"
• SIGMA0 0 9. 10. .006 .006
• DEGRAD 0 4.
• BINS 0 5 5 3 3
• RAMP0 0 10 10
• RAMP04 0 10 10
• CHI2MAX 0 3.5
• CHI2PMAX 0 2.5
• OFFSET 0 100. 100.
  maximal offset for alignment search (dX dY)

13
Parameters overviewSelections
Cuts for segments (view coord - same as in run
files) used for up/down linking XCUT - reject
inside the defined phase volume (accept the
eXternal part) ICUT - reject outside the
defined phase volume (accept the Internal part)
xmin xmax ymin ymax txmin
txmax tymin tymax Pmin Pmax ICUT 1 -500.
500. -500. 500. -.9 .9
-.2 .2 0. 17. ICUT 2 -500. 500.
-500. 500. -.9 .9
-.2 .2 0. 17. Cuts for basetrack
(coord - same as in couples tree) used for
alingment search ICUT 0 CUTCP can be used
before alignment and linking Max for N1
N1tot N2 N2tot Chi2 Chi2P CUTCP 0
1 10 1 10 3.5 3.5

14
Parameters overview other

Should be defined once/dataset AFID 0
1-use fiducial marks, 0 - stage coord (stored in
the run files) CLUST 0 1- with grains
analysis (time consuming), 0 - no grains used
for raw data selection (reading from run.root)
x0 y0
z0 tx ty VOLUME0 -62000.
58000. 0. 0. 0. VOLUMEA
5000. 5000.
15
Additional information

• FEDRA CVS repository
• http//ntslab01.na.infn.it/cgi-bin/cvsweb.cgi/fed
  ra/
• Installation instruction
  fedra/README.INSTALL
• Raw data structures definition
  fedra/src/doc/edb_data_structure.txt
• Data-set organization and processing
  fedra/src/appl/jun2003/README
• Progress information
  fedra/doc/fedra_release_notes.txt
• Raw and processed data sample
  http//ntslab01.na.infn.it/public/CMOS/

16
Useful scripts in the tarball

• Rootlogon.C - load fedra libraries
• Makepar.C - create dummy par-files for the
  following update
• Check_raw.C - process run-file and plot
  characteristic distributions
• Check_set.C - display the summary plots for
  the data-set

17
Additional modules not stable yet but already
functional

• Emulsion display developed by Igor
• Ali/Op-Root display is adopted for Fedra
• Image Processing module (VTIK)
• FIR-filtering of the images
• Clasterization with output to root tree
• Grain Analysis module
• Vertical grains sequence reconstruction (useful
  for hardware parameters estimation)

18
Example of using EdbIP

• EdbIP ip
• ip.SetFIR( new EdbFIRF("6x6A"))
• ip.SetThr(thr)
• EdbFrame frame new EdbFrame()
• frame-gtSetID(i)
• frame-gtSetZ(z)
• EdbImage image new EdbImage(im.bmp,"BMP")
• frame-gtSetImage(image)
• TTree treeip.InitTree()
• ip.Clusterize(frame,tree)

19
Prospectives

• Tracking (ideas are welcome, Imad?)
• Vertex reconstruction (first pass could be done
  in a month)
• Marriage with Op-simulation?
• Code polishing
• Documentation