Lateral X-ray marks finding with ESS

1
Lateral X-ray marks finding with ESS
C. Bozza, E. Carrara, L.S.Esposito and G.Sirri

• Goal implementation of the lateral mark finding
  in the ESS software
• First Test Plate-to-plate intercalibration study
• Second Test Start a scan back without
  intercalibration zones

2
Lateral X-ray mark
X
3
Mark recognition

• Not possible to be done
• No extra camera with low magnification optics are
  available
• Only a small amount of the lines can be scanned
• The two strip are not always long enough to print
  a clear crossing point
• Not possible to scan only one FOV per mark
• The approach is
• Three segments of each strip are scanned and
  lines are obtained by a linear fit. The
  coordinates of intersection point are the mark
  coordinates.

NO
NO
Intersectionpoint
4
Mark finding procedure
This is the procedure currently under test.
4. Repeat 2,3for the 2nd line
1. Set the corner
2. Scan this edge to find a first line segment
3. Scan further two line segment here
6. Evaluate the intersection points and evaluate
the transformation
5. Repeat 2,3,4 for the second mark
5
How the transformation is evaluated?
Rdet A Rstage B
Affine transf. of the plane are composed of
scaling (reflections), rotations, and
translations. Emphasizing this order, the
components of a transformation can be encoded
with the IFS formalism r scaling (reflection)
in the x direction s scaling (reflection) in
the y direction theta - rotation of horizontal
lines phi - rotation of vertical lines e -
horizontal translation f - vertical translation
With 2 mark coordinates up to 4 independent
parameter can be evaluated? rs , thetaphi ,
e , f (same scaling for both axis, rigid
rotation, translation)
6
Mark finding reproducibility
5 measure of the mark coordinates without
removing the emulsion plate
IntersectionMark1 ( 121722.1 , 3101.5 ) (
121722.4 , 3099.6 ) ( 121722.4 , 3107.0 ) (
121722.4 , 3104.7 ) ( 121722.3 , 3102.8 )
IntersectionMark2 ( 121820.2 , 97942.4 ) (
121818.3 , 97937.0 ) ( 121820.3 , 97943.1 ) (
121822.9 , 97938.3 ) ( 121820.3 , 97936.6 )
The system is able to find the marks within 10
microns( better for the X coordinate )
7
First Test plate-to-plate intercalibration study

• A first test has been done by scanning 5 1 cm²
  zones in two adjacent emulsion plates of a brick
  exposed to cosmic rays.
• The scanning was done at LNGS Scanning station
  with a vacuum system equipped with a vacuum
  channel larger than the emulsion size.

Adhesive tape
emulsion
3
1
glass
5
4
2
Vacuum channel

• The emulsion were attached to the adhesive tape
• In this setup the mark lines images are not
  affected by the vacuum channel image
• Scanning done using the mark finding module in
  the acquisition sw

8
First Test plate-to-plate intercalibration study

• Tracks of the first plate are projected to the
  second one without extra intercalibration
• Pattern match results
• ZONE 1 DX E 100 M -14.0 RMS 8.8
• ZONE 1 DY E 100 M 33.0 RMS 8.2
• ZONE 2 DX E 51 M -7.0 RMS 11.6
• ZONE 2 DY E 51 M -3.2 RMS 9.1
• ZONE 3 DX E 105 M -3.7 RMS 8.0
• ZONE 3 DY E 105 M -7.6 RMS 10.3
• ZONE 4 DX E 86 M 10.5 RMS 5.6
• ZONE 4 DY E 86 M 13.5 RMS 5.9
• ZONE 5 DX E 91 M -6.5 RMS 7.0
• ZONE 5 DY E 91 M 9.4 RMS 6.7

9
First Test plate-to-plate intercalibration study
- zone 1
10
First Test plate-to-plate intercalibration study
- zone 2
11
First Test plate-to-plate intercalibration study
- zone 3
12
First Test plate-to-plate intercalibration study
- zone 4
13
First Test plate-to-plate intercalibration study
- zone 5
14
First Test Evaluation of the best transformation
After matching the track couples can be used to
evaluate the intercalibration between the 2
plates. The results have to be compared with the
Identity transformation which is the one supposed
to be used without intercalibration zones. 406
matches found. QMAP 6 parameters Aff2D
0.999894 -0.000181 0.000207 0.999933 18.65 -20.81
(DET0.999827) r 0.999894 s 0.999933 theta
0.000207 phi 0.000182 e 18.65 f
-20.81 QMAP - 3 parameters (EdbAffine class
implemented) Aff2D 1.000000 -0.000199 0.000199
1.000000 12.09 -23.41 (DET1.000000) r
1.000000 s 1.000000 theta 0.000199 phi
0.000199 e 12.09 f -23.41 Scaling and
Rotation are very small (lt 1/10000). The
translation error is about 20 micron.
15
Second Test start a scan back

• The scan back of brick 8199 with a predicted
  track has been started for test.
• This scanning has been done at Bologna using a
  vacuum system with a standard vacuum channel
• To avoid interference with the image of the
  vacuum channel the emulsion plate was placed at
  the edge of the channel with lateral mark outside
  the vacuum area.

Vacuum channel
16
Second Test scan back history (5 plates)

• Scan Back done without computing infrastructure
  assistance. Linking and Prediction Finding has
  done manually
• POS TOL 100 micron SLOPE TOL 0.03
  0.05SLOPE

PL PPX PPY PSX PSY FPX FPY FSX FSY FP DPX DPY 1 -2
968655 -3247649 -0.405 0.508 -2968628 -3247628 -0.
388 0.506 23 -27 -21 -2968628 -3247629 -0.374
0.519 20 -27 -20 2 -2968124 -3248286 -0.388 0.506
-2968073 -3248362 -0.381 0.499 33 -51 76 3 -29675
78 -3249011 -0.381 0.499 4 -2967082 -324965
9 -0.381 0.499 -2967042 -3249708 -0.393 0.503 23 -
41 48 5 -2966531 -3250362 -0.393 0.503 -2966546 -3
250378 -0.382 0.503 24 16 16 -2966564 -325032
7 -0.383 0.492 16 34 -34

• Plate 3 not found also for the scan back with
  intercalibration zones
• Plate 4 skipped by the scan back with
  intercalibration zones because optical marks were
  printed on the wrong emulsion side.

17
Technical considerations

• The sw implementation is working, but it cannot
  be considered in the final state.
• For the version Im testing
• Automatic search of the First Mark procedure
  not tested
• The Mark Finding can fail if used above the
  vacuum channel
• Automatic recovery strategy and manual set
  procedure are not implemented
• No autofocus
• Sometimes it recognize the (wrong) X-ray line
  placed 1 mm far from each vertical strip.

18
How to deal with

• For these tests
• Set the upper-right corner manually
• Use of large vacuum area plate and attach
  emulsion on an adhesive tape
• Use the standard but put the markers outside the
  vacuum area
• Use precise mechanical reference on the glass
  plate
• Put the exact emulsion sizes in the map string
• Failures if marks are outside the vacuum area
  set the Z with care.
• Open the log file and check manually if wrong
  strip is recognize
• The development is in progress.

19
Conclusions

• The current implementation of the lateral mark
  finding in the ESS software has been tested and
  is working correctly (with some technical
  prescriptions).
• Mark finding reproducibility is good
• Plate-to-plate intercalibration of the order of
  20 micron.
• The development to have a foolish proof version
  is in progress