×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Close-range imaging, long-range vision


CALIBRATION OF AN SEM, USING A NANO POSITIONING TILTING TABLE
AND A MICROSCOPIC CALIBRATION PYRAMID
Olaf Sinram', Martin Ritter”, S. Kleindiek', A. Schertel’, H. Hohenberg" and J. Albertz"
? TU Berlin, Photogrammetry and Cartography, D-10623 Berlin, Germany, sinram@fpk.tu-berlin.de
* HPI Hamburg, Electron Microscopy Group, D-20251 Hamburg, Germany, ritter@hpi.uni-hamburg.de
* Kleindiek Nanotechnik, D-72770 Reutlingen, Germany
4 FE] GmbH, D-85622 Feldkirchen-Munich, Germany
Commission V, WG 2
KEY WORDS: Adjustment, Calibration, Photogrammetry, Orientation, Bundle, Close Range
ABSTRACT:
In this paper we present the calibration of a scanning electron microscope, using a high precision tilting sample stage and a new
microscopic calibration pyramid. Difficulties when using extrem
ely high magnifications will be stated and means of solutions are
presented and evaluated. Since the scanning electron microscope cannot be moved around the object, the object has to be tilted
instead. By this movement, the object can be seen in several virtual points of view, which is a necessity for any three dimensional
reconstruction. As for generating the virtual views, the first difficulty encountered is to position the sample into the eucentric axis.
Only when positioned in the rotation axis, the sample remains within the field of view, instead of being moved outside. Therefore, a
special tilting table was required, which provided maximum precision and accuracy. Furthermore, an object had to be found, which
met the requirements of a calibration object. Here, a new microscopic cascade pyramid was developed, supplied with 38 spatially
distributed control points. Finally, for each tilting series, all desired images were oriented using a bundle block adjustment following
the rules of parallel projection. In this paper, the mathematics of parallel projection will be an important chapter, pointing out the
differences to central projection. As a result, the accuracy of the tilting table and the imaging properties of the microscope are
presented. These properties are used as the basic essential parameters for further evaluation of microscopic images.
KURZFASSUNG:
In dieser Arbeit stellen wir die Kalibrierung eines Rasterelektronenmikroskops vor, bei der wir einen kippgenauen Probentisch und
eine neuartige mikroskopische Stufenpyramide als Kalibrierkórper verwenden. Die Probleme, die sich stellen, wenn man bei extrem
hohen Vergróferungen arbeitet, sollen aufgezählt und Lösungswege vorgestellt und beurteilt werden. Da das Mikroskop nicht um das
zu untersuchende Objekt herum bewegt werden kann, muss statt dessen das Objekt gekippt werden um virtuelle Kamerastandpunkte,
die für jede räumliche Auswertung eine Grundvoraussetzung sind, zu erhalten. Bei der Generierung der virtuellen
Aufnahmeanordnung durch die Kippung, wird die Schwierigkeit auftreten, die Probe in die euzentrische Achse zu bewegen. Nur
direkt auf der Rotationsachse liegend, verbleibt die Probe bei de
r Drehung im extrem kleinen Sichtfeld, anstatt dieses zu verlassen.
Um diese Anforderungen zu erfüllen, wurde ein hochpräziser und zuverlässig genauer Kipptisch benötigt. Zudem musste ein Objekt
gefunden werden, welches den Ansprüchen für ein Kalibrierobjekt genügt. Es wurde daher eine Stufenpyramide entwickelt, die mit
38 räumlich verteilten Passpunkten ausgestattet ist. Schließlich wurden alle ausgewählten Bilder einer Kippserie durch eine
Bündelblockausgleichung orientiert, die in unserem Falle den Gesetzen der Parallelprojektion folgt. Die Mathematik der
Parallelprojektion stellt in dieser Arbeit ein wichtiges Kapitel dar, in welchem des weiteren die Unterschiede zur Zentralprojektion
erklärt werden sollen. Als Ergebnis stellen wir die Genauigkeit des Kipptisches und die Abbildungseigenschaften des Mikroskops
vor, welche die Ausgangsparameter für jede folgende räumliche Auswertung bilden.
1. INTRODUCTION
The surface reconstruction of biological objects is the
motivation for exploiting the possibilities modern microscopic
techniques are offering. At the Technical University of Berlin in
cooperation with the Heinrich-Pette-Institut in Hamburg a
research project under the title “photogrammetric reconstruction
of biological surfaces on the basis of SEM Data” is currently
sponsored by the Deutsche Forschungsgemeinschaft.
The reconstruction of real world objects can be achieved in
various ways. Photogrammetry offers means for non-destructive
reconstruction since the object, being depicted is not touched.
The derivation of spatial coordinates is done indirectly through
the images. This becomes especially interesting, if the sample of
interest is very sensitive to tactile treatment, or may even suffer
geometric distortions.
Cellular biological samples are very small. Only few options
remain to capture the sample properties. Among others, optical
microscopy, electron microscopy and AFM are to be mentioned.
For a brief overview see (Hemmleb, 2001). In this project we
will explore the properties of SEM images for further
evaluation. The SEM has always been an interesting approach
for photogrammetry, since it provides a large depth of focus and
images can be captured over a wide range of magnification,
very much in contrast to optical microscopes. Also, the good
signal to noise ratio makes them a first choice for evaluation.
Since the data are generated digitally (Koenig et. al., 1987) it is
a good basis for automation. A brief overview about how SEM
images are generated will be given in the next chapter (Maune,
1976).
—210—
Ther
The
imag
imag
are |
requ
diffe
to m
that
impl
matc
intro
first
stan
Whe
para
Ther
para
com
21°
An
acce
elec
micı
sam
calle
sam
of S
posi
emit
give
a ca
topo
maii
the ]
The
allo:
the |
to-n
phot
appl
elec
be i
100!
will
2.2
of v
posi
a Wi
not .
to ti
fron
an |
unw
may
not
axis
mo