AL1_SiplaceX-en.pdf - 第295页
1 - 5 S tudent Guide SIPLACE X Edition 02/2005 8 Siplace Vision 5 8.1.4 Shape of fiducials for PCB position and pla cement position recognition Following 7 (9) shapes of fiducial could be t aught as a synthetic fiducial.…
1 - 4
Student Guide SIPLACE X
8 Siplace Vision Edition 02/2005
4
Teaching fiducials (and bad mark) a complete programmed PCB is scheduled to the station - in-
clusive all fiducial coordinates -! With teach function is the 1st PCB moved in placement area for
teaching fiducials and bad marks. Teach all individual fiducials and with 'Test all fiducials' test them
on all camera's of the station. Then with the next board start production.
8.1.2.1 Fiducial arrangement on the PCB
If the position recognition fiducials on a board arranged in a row the PCB-position PCB-angle and
the length yield could be determined. Therefore 2 fid. are enough so SIPLACE Pro transmits only
2 fiducial positions.
8.1.3 Import fiducials
All fiducials, for ICOS have to be re-taught for SIPLACE Vision.
There is no mutual influence of the two vision systems because the Fiducial teach data are sensor
specific saved!! (5 SST, 24 SST, ..).
Specialties: The SIPLACE Vision accept dual cross as a kind of synthetic fiducial; ICOS-Systems
define this kind of fiducial as a ‚2D-pattern' a Import of this kind of fiducial is excluded.
Please Note:
Fiducials for SIPLACE Vision have to be teached new!
To the Station transferred PCB-
recognition fiducial positions.
Here is calculated the: PCB-
position; PCB-angle; and PCB
length and width yield.
To the Station transferred PCB-
recognition fiducial positions.
Here is calculated the: PCB-
position; PCB-angle; and PCB
length yield.
1 - 5
Student Guide SIPLACE X
Edition 02/2005 8 Siplace Vision
5
8.1.4 Shape of fiducials for PCB position and placement position recognition
Following 7 (9) shapes of fiducial could be taught as a synthetic fiducial.
Following measuring results of the fiducial issued:
Kind of fiducial (naming in fiducial
list) What to teach / comment
Cross; Double cross; Shape, dimension, stroke thickness and bright or dark picture.
Circle Shape, dimension and bright or dark picture..
Circle ring Shape, dimension, stroke thickness and bright or dark picture.
Rectangle or square (square a special
form of a rectangle)
Shape, dimension and bright or dark picture.The square is a well-known
special case of the rectangle where all sides are adjusted to identical
values.
Rectangular ring or squared ring(squared
ring a special form of a rectangle ring)
Shape, dimension, stroke thickness and bright or dark picture.The
squared ring is a well-known special case of the rectangular ring where
all sides are adjusted to identical values.
'Diamond' - a 45° turned square Shape, dimension and bright or dark picture.
'Pattern' other fiducial shapes 2D (trian-
gle, circle segment, conductive pattern or
similar structures)
Pattern recognition algorithm, the fiducial is as a - '2 D-fiducial' - in the
fiducial list named.
At fiducial teaching: At fiducial centering:
=> X-position of the structure to camera center => X-position of the structure to camera center
=> Y-position of the structure to camera center => Y-position of the structure to camera center
=> Angle deviation of the structure to PCB-camera
coordinate system
1 - 6
Student Guide SIPLACE X
8 Siplace Vision Edition 02/2005
6
8.1.5 Synthetic fiducials
In rough search step is with reduced resolution the gray value difference between both points of
a point pair determined. The shape position of the point pair calculates the algorithm from the geo-
metric model.
From the gray value difference the real position is determined.
In fine search step is along the scanning lines the largest gray value difference determined. From
this positions is the regression line calculated their position define the fiducial center position. The
positions of the scanning lines are defined from the point pair positions of the rough search. Par-
allel to the position measurement is a quality value determined. This quality value show how ac-
curate fit the theoretical model at the determined fiducial position.
This principal could be used for bad mark recognition too. The negative bad mark is recognized if
the quality level is lower than a defined limit.
Camera picture of the
fiducial
Rough search with reduced
resolution:
Position of the
scanning lines (yellow) at rec-
ognized pairs.The blue point
has a higher gray value
(brighter) than the yellow one.
Fine search: Position of the
scanning lines (yellow) at
recognized pairs.
Green crosses: found posi-
tions of the largest gray gra-
dient along the Scanning line
Green straight line: regres-
sion line