Jean Favier LAPP OPERA meeting Hamburg 5/06/04 GenIma: program to simulate emulsions.

Jean Favier LAPP OPERA meeting Hamburg 5/06/04

GenIma: program to simulate emulsions


Reconstruction optimization and parameters understanding are difficult for emulsions:

• Lack of data for various conditions ( particle type, angles, energies, fog densities, ambiant radioactivity integration times, etc) and poor statistics.

• Large feed-back time between beam tests and scanning

• ECC efficiency evaluation will be needed


GEN: Particles generation in GEANT 3.2

• Anything you like as particles, energies, dE/dX, vertices, neutrinos events, etc

• Keep the coordinates of energy deposites in a volume =200 200 288 3 21 dead-zone

• Density of 30 grains/100 -rays generated• Track density controlled ( ex: 10/field of view)


ccd plane

Focused

plane

Grain plane

d=3.254 mm d=200 mm

r=6mm

The systeme objective-ccd (in red) can move along the optical axis: we keep 16

tomographic planes ; is defined by objective characteristics

The virtual microscope

objective


IMA : makes pictures fom GEN output• For each “grain” x,y,z , intercept the ccd surface by a

conical prolongation surface S of pixels to “darken”. cone aperture : N.A (numerical aperture):

= arcsin(N.A/nrefr )• Fill them with grey in each of the 32 picture planes;

grey density varies with 1 / S and e-d/ ( is an effective emulsion attenuation lenghth)

• Lateral size of grains: appropriate smearing matrix • z- grain size: 5 z values displaced by (0.6 /5) (luxuous)• Linear or parabolic distorsions included (ex:flat from 0

to 40mrad)• Objective geometrical distorsions, non-uniform lighting• Pictures are written in Fortran (PPM format), then

converted in JPEG or BMP with XView.


“My objective”

• Dry ( oil) objective N.A = 0.8 , G= “ 50 “ • Focal length : f=3.23 mm• Distance objective-ccd plane: 200 mm G=62.5• ccd: 1024 1024 pixels 12 12 2

• thin lense approximation, no explicite diffraction ( included in matrix)

sin = N.A/nrefr

(in a camera: objective luminosity = f/Dobj=1/2.tg () ) = 32o in emulsion


Format PPM

http//cedric.cnam.fr/~gouet/Var/ppm.html

#Example: toto.ppmP34 4 255# red green blue0 0 0 16 16 16 32 32 32 48 48 4864 64 64 80 80 80 96 96 96 112 112 112128 128 128 143 143 143 159 159 159 175 175 175191 191 191 207 207 207 223 23 223 239 239 239


Fog and background

• Fog is randomly generated as single grain with a density of 6 grains/ (10 )3

• time consuming, but

• Albums of fogs can be generated only once, and pictures added to the tracks ones

• Album of real emulsions data can be used also (fog+radioactivity fossiles tracks after refreshing)

• An ultimate uniform grey layer is added at the end (light scattering, long distance shadows)


Oil and air (without fog) NA=0.8 electrons 0.2-1 MeV


Oil and air (without fog) NA=0.8 electrons 0.2-1 MeV


Emulsion simulation

• 16 images


Emulsion simulation

• 16 images


Emulsion simulation

• 16 images


Emulsion simulation

• 16 images


Emulsion simulation

• 16 images


Emulsion simulation

• 16 images



Depth of field

• From geometry, depth of field can be shorter for oil since is larger.

• But diffraction is imposing a minimum spot radius (Airy distribution)

• NIKON handbook :

dof=.n/NA2+n.esep/NA


80 MeV 2 GeV

2

Oil objective Dry objective


An application:base track efficiency

(extremely preliminary)• Distorsion (0-40 mrad)• Clustering: just threshold and saturation• Start from top microtrack, open a 50 mrad cone• Find n “clusters” in bottom road : n6?• Efficiency for base track association: 0.93 (0 mrad)• Increase with angle 0.95 (300 mrad)• Sysal filter or micro track finder responsible for this measured

fact? Inject Genima pictures (or code ) in Sysal• What is this dependency (if any) with the UTS,SUTS tools? :simulate pulse height


Next steps…• Release a Fortran version with

documentation (Sysal)• Pictures are in FEDRA (Dominique)• “Albums” of various background• Refreshing action parametrization• Filtering in GenIma to reduce heavy

pictures?

and…neutrino events with several emulsions and lead ?

Jean Favier LAPP OPERA meeting Hamburg 5/06/04 GenIma: program to simulate emulsions.

Documents

mmdistance objectiveccd

ccd surface

fog na

objective characteristics

objectivedry oil objective

mmthe systeme objectiveccd

fog densities

objective luminosity