Usage of reconstruction software

README for observer_tools software

The package observer_tools is intended for analysis and reconstruction of probe data. The package is pretty disorganized in its current versions as it was heavily modified to suit particular goals that came about during reconstruction.

The package consists of three modules:

  • observer module. This module contains routines for processing of probe data measurements. Since some routines share conventions (like shape of the data), reconstruction uses some of the files directly. Historically, observer_tools and observer were a shared resource, do the division is not really solid.
  • parseCommandLine module. This is a service for parsing configuration files and settings therein. A dedicated routines were used for parsing command line options only, but it evolved to a general parsing program for settings files. Used heavily to manage reconstruction constants, geometry settings and the like both in observer as well as observer_tools.
  • observer_tools module. It has a flood, parse_si and coinc directory. The flood directory allows calibration and characterization of BGO part, parse_si is for silicon part (including gain calibration and energy resolution studies). The coinc directory is for coincidence data matching, reconstruction, system matrix construction, SNR evaluation and whatnot. A total of hundred files with as much as 1000 rows per some of the busiest implementation files, it is a mixture of core routines and evaluation scripts. Often, transitory pieces of code get a script_ prefix which, if useful, migrate to non-script_ prepended core files. The tools are intended to be used as root libraries and are loaded on start, but they can be used in standalone programs as well.

INSTALLation

Start in a fresh folder. Whatever you call it, I will call it SFHOME (for software home).

Download neccesary files:

> cd SFHOME
> git clone http://www-f9.ijs.si/~studen/git/parseCommandLine.git
> git clone http://www-f9.ijs.si/~studen/git/observer.git
> git clone http://www-f9.ijs.si/~studen/git/observer_tools.git
> git clone git://git.code.sf.net/p/daqpp/vmedaq

Parse command line

> cd SFHOME/parseCommandLine
> make
#there is no make install

Observer

It is not scrictly necessary to make the binaries, but potentially useful.

> cd SFHOME/observer
> cp env.sh.template env.sh

Edit env.sh so it points to correct location of required:

  • GNU scientific library (GSL),
  • parseCommandLine (PCL=SFHOME/parseCommandLine)
  • root installations (ROOTCORE).
  • Downloaded observer tools (OBSERVERTOOLSBASE=SFHOME/observer_tools)
  • VMEDAQ (VMEDAQBASE=SFHOME/vmedaq)

A sample env.sh.template looks like this:

PCL=/afs/f9.ijs.si/home/studen/software/build/parseCommandLine
ROOTCORE=/data0/studen/software/install/root
#a good guess is #>basename `which root`
GSL=/usr
#for make_links.sh
OBSERVERTOOLSBASE=/afs/f9.ijs.si/home/studen/software/build/observer_tools
VMEDAQBASE=/afs/f9.ijs.si/home/studen/software/build/vmedaq

To install, set the variables and compile

>bash
>. env.sh
>make

Report any failure.

Observer tools

>cd observer_tools
>cd coinc/root_scripts
>cp env.sh.template env.sh

Correct env.sh so that:

  • PCL points to parseCommandLine installation (PCL=SFHOME/parseCommandLine)
  • OBSERVERBASE points to observer installation (OBSERVER=SFHOME/observer)
  • VMEDAQBASE points to VMEDAQ installation (VMEDAQBASE=SFHOME/vmedaq)

A sample env.sh.template points to a sample environment file.

Continue with the setup. A couple of links need to be set up to include files from other projects

>bash
>. env.sh
>./rm_links.sh
>./make_links.sh

Now you can start root. When you do, the routines will get compiled to be ready to use. First start takes a while.

>root

USE of software

Make a configuration/study directory. Call if whatever, I'll call it IMGREC. Make two additional directories, IMGREC/layout and IMGREC/studies.

Copy layouts (setup0.zip,setup2.zip) to layout directory, and unzip them.

Copy studies (setup_um.zip,setup2_um.zip) to studies directory, and unzip them.

Copy phantoms (phantoms.zip) to IMGREC directory, and unzip them.

There are paths, which are incorrect in some of the files. You should edit IMGREC/layout/setup0/setup1000.in. The variables setupPath and studyPath should be adjusted for new location:

setupPath_2=IMGREC/layout/setup0
studyPath_2=IMGREC/studies/setup_um

Similarly in IMGREC/layout/setup2/setup2000.in. For processing (running routines in observer), you should also update the calibration files, more on that in calibration entry.

Also, in setup_um/setup_um_study.in, you should probably correct all paths. A valid file would look like this:

setup=IMGREC/layout/setup0/setup1000.in
phantom=IMGREC/phantoms/phantom_0_rw_0p3_bkg_0p01.out
nCoinc=3
coincFlag_0=0xf
sinogram_0=IMGREC/studies/setup_um/sinogram_0_f.out
sensitivity_0=IMGREC/studies/setup_um/sensitivity_0_f.out
coincFlag_1=0xf0
sinogram_1=IMGREC/studies/setup_um/sinogram_0_f0.out
sensitivity_1=IMGREC/studies/setup_um/sensitivity_0_f0.out
coincFlag_2=0x100
dataImport_2=IMGREC/studies/setup_um/dataImport_0x100.in
sinogram_2=IMGREC/studies/setup_um/sinogram_0_100.out
sensitivity_2=IMGREC/studies/setup_um/sensitivity_0_100.out
sinoMatrix_2=IMGREC/studies/setup_um/sinoMatrix_0x100.in
forwardMatrix_2=IMGREC/studies/setup_um/forwardMatrix_0x100.in

You should probably be able to do the same for the setup2_um_study.in from the setup2_um directory.

Now you should already have some files to play with. For example, you could reconstruct the BGO-BGO data. To do that, do:

>cd SFHOME/observer_tools/coinc/root_scripts
>root
root[] DataSinoStorage ds
root[] ds.ImportStudy("IMGREC/studies/setup_um/setup_um_study.in");
root[] sinogramData s;
root[] ds.makeSinogram(&s,0x100);#0x100 is for BGO-BGO
root[] s.getSinogram()->Draw("COLZ") #should give you a funny sinogram
with very few lines filled
root[] ds.SetSensMatrix(&s);
root[] Reconstructor rec;
root[] rec.addSinogram(&s);
root[] rec.FBP();
root[] TH2D h2;
root[] rec.getSData()->getImage()->Copy(h2);
root[] h2.Draw("COLZ");#should give you filter-back-projection
reconstruction

links

social