definite and precise relation between sensor ele-
ments and frame storage pixels.
Hence electronic insufficiencies in conventional
CCIR standard video interfaces destroy the advanta-
ge of geometric stable CCD sensor elements. A ty-
pical value for jitter is 20 ns, which will cause
statistical deviations of about 0.2 pixel in profi-
le positions from line to line. This value corre-
sponds very well to the experimentally gained valu-
es for the standard deviation of profile positions
around the regression line.
The solution of this problem follows a simple prin-
ciple: pixel clock signals from the CCD sensor must
be transmitted to the frame storage. Fortunately
more and more manufacturers of cameras and frame
storage boards get rid of the conventional CCIR vi-
deo standard and offer pixel-synchronous image ac-
quisition systems. Applications of pixel-synchro-
nous image acquisition systems report subpixel re-
solutions of 1% to 3% of a pixel period [12],
(13]. This resolution is about 10 times higher
than the reported resolutions for CCIR standard
image acquisition systems [5], [14], [15].
CONCLUSIONS
The industrial application of an optical 3-D sen-
sing method requires careful considerations about
system design. Especially the condition of a scan-
ning illumination system influences the criteria
for the choice of a CCD camera for the detection
system.
Evaluation of 3-D data from the camera image de-
mands for the application of photogrammetric me-
thods for calibration of the detection system. The
basic equations for a simple camera model of per-
spective projection with pinhole imaging are deri-
ved. The results gained with the experimental setup
with high symmetry show promising results even for
this simple model.
The analysis of the observed errors shows again the
importance of pixel synchronous image acquisition.
In spite of this avoidable errors, the accuracy of
the optical 3-D sensor is comparable with usual in-
dustrial measuring methods, which can not be used
for an in-process measurement. Thus the presented
non-contact, simple and fast optical 3-D sensor ba-
sed on the principle of light sectioning meets all
requirements for the in-process measurement of ben-
ding angles.
ACKNOWLEDGEMENTS
The investigations presented in this paper are car-
ried out at the University of Erlangen-Nürnberg,
Chair of Manufacturing Technology, and supported by
the 'Deutsche Forschungsgemeinschaft' (Ref.: Ge
530/6-1, 'Biegewinkelerfassung').
The author would like to thank his professor in or-
dinary, Univ.-Prof. Dr.-Ing. Manfred Geiger, for
the opportunity to work on this project and for his
scientific advise and confidence.
REFERENCES
[1] Häusler, G.; Heckel, W.: Light sectioning with
large depth and high resolution. Applied Op-
tics, Vol. 27 (1988), No. 24, p. 5165.
[2] McLeod, J.H.: Axicon: A new type of optical
element. J. Opt. Soc. Am., Vol. 44 (1954), p.
592.
[3] McLeod, J.H.: Axicons and their uses.
J. Opt.
Soc. Am., Vol. 50 (1960), p. 166.
[4] Slevogt,
Gruyter,
H.: Technische
1974, p. 56.
Optik. Berlin: de
[5] Lenz, R.: Zur Genauigkeit der Videometrie mit
CCD-Sensoren. In: Proc. DAGM = Informatik
Fachberichte 180, Vol. 10 (1988), p. 179.
[6] Alexander, B.F.; Ng, K.C.: Elimination of sy-
stematic error in subpixel accuracy centroid
estimation. Optical Engineering, Vol. 30
(1991) 9, p. 1320.
416
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
Seitz, P.: Optical superresolution using so-
lid-state cameras and digital signal proces-
sing. Optical Engineering, Vol. 27 (1988) 7,
p. 535.
Heckel, W.: Lichtschnittverfahren mit erwei-
tertem Meßbereich und hoher Auflösung. Diploma
Thesis, University of Erlangen, Physics Insti-
tute, 1988.
Tsai, R.Y.: A versatile camera calibration
technique for high accuracy 3D machine vision
metrology using off-the-shelf TV cameras and
lenses. IBM Research Report RC 11414 (#51342),
1985.
Lenz, R.K.; Tsai, R.Y.: Techniques for cali-
bration of the scale factor and image center
for high accuracy 3D machine vision metrology.
IBM Research Report RC 12266 (#54867), 1986.
Dähler, J.: Problems in digital image acquisi-
tion with CCD cameras. Proc. of ISPRS Inter-
commission Conf. on "Fast processing of photo-
grammetric data", Interlaken, June 2-4, 1987,
p. 48.
Raynor, J.M.; Seitz, P.: The technology and
practical problems of pixel-synchronous CCD
data acquisition for optical metrology appli-
cations. Proc. SPIE, Vol. 1395 (1990), p. 96.
Gale, M.T.; Seitz, P.: High resolution optical
metrology and edge detection using a PC-con-
trolled smart CCD Camera. Proc. SPIE, Vol. 701
(1986), p. 254.
Hachicha, A.; Simon, S.: Subpixel edge detec-
tion for precise measurements by a vision sy-
stem. Proc. SPIE, Vol. 1010 (1988), p. 148.
Kleinemeier, B.: Measurement of CCD interpola-
tion functions in the subpixel precision ran-
ge. Proc. SPIE, Vol. 1010 (1988), p. 158.
