Note: Descriptions are shown in the official language in which they were submitted.
2a7~3~
A Process and Arrangement for ~he Optoelectronic Measurin~ of
Objects
The invention relates to a process for the optoelectronic measuring of the
shape, in particular the cross sectional shape, of objects, e.g. workpieces,
5 or for calibrating optoelectronic measuring systems, wherein at least one
light strip is projected from at least one light source, preferably a laser light
source, onto ~he object to be measured and/or the calibration body (light
section process), and the light strips are recorded by preferably the same
number of video cameras, preferably CCD solid-state cameras, as there are
10 light strips, and are imaged on the sensor elements of the cameras, and the
camera signals are fed to an evaluation unit comprising a computer for the
purpose of evaluating the image and calculating the dimensions of the
object, or for determining the basic data and parameters needed for
calibration. In addition, the invention relates to an arrangement for
15 implementing this process.
The image processing systems used in such processes and arrangements
have certain disadvantages specifically as regards their limited resolution,
which limits the maximum achievable accuracy. Customary systems can
resolve an image into, for example, 512 x 512 image points. If objects of
20 various size are measured by means of a measuring arrangement using the
iight section process, small objects will naturally produce a small image on
the avallable sensor element, i.e. the available measuring field will not be
fully utilized, and the accuracy of the 0valuation will suffer; small objects
use up only part of the measuring field and are therefore resolved into fewer
25 than 512 x 512 image points, and as a result the relative measuring
accuracy is reduced.
The purpose of the invention is to achieve maximum image processing
accuracy using measuring processes and arrangements which operate
according to the light section method.
~ V I IJ (~
-2-
According to the invention, a process of the type mentioned at the
besinning is designed in such a way ~hat in order to approximate or match
the size of the image or of the respective light strip(s) to the size of the
respective sensor element, the imaging scale is modified or varied or the size
5 of the measuring field of the respective camera is varied as a function of thesize of the light strips generated on the object and/or calibration body or is
adapted to the size of these strips.
An arrangement for optoelectronically measuring the shape, in particular the
cross sectional shape of objects, e.g. workpieces, or for calibrating
10 optoelectronic measuring devices, wherein at least one light strip is
projected from at least one light source, preferably a laser light source, onto
the object to be measured and/or the calibration body (light section
process), and wherein the light strips are recorded by preferably the same
number of video cameras, preferably CCD solid-state cameras, as there are
15 light strips, and are imaged on the sensor elements of the cameras, and
wherein the cameras are connected to an evaluation unit comprising a
computer for the purpose of evaluating the image and calculating the
dimensions of the object, or for determining the basic data and parameters
needed for calibration, is characterized in the manner according to the
20 invention in that devices are provided for modifying or varying the imaging
scaie of the light strip(s) and by means of these devices the image siza of
the llght strlp(s) can be varied and in particular can be adapted to the size ofthe respective sensor element, or devices are provided for modifying the
measuring field size of the cameras and for adapting the size of the
25 measuring field to the light strips which are to be measured on the objects
or on areas of the objects and/or on the calibration bodies.
In the procass and arrangemsnt according to the invention, while the sizs of
the sensor element in the camera remains unchanged, the evaluation of
small objects or small light strips can be considerably improved by
30 generating larger images of the light strips on the sensor element of the
camera, however in such a way that ail the light strips needed for the
purpose of evaluation are imaged jointly or simultaneously. Whether one or
~n~n~;j li
-3-
more light strips from the calibration body and/or from the object are imaged
on the sensor element depends on the type of evaluation process selected.
If the images of the light strips are larger than the sensor element, it is of
course also possible to select a smaller imaging scale and thus to adapt the
5 imaga size of the light strips to the size of the sensor element. The same
also applies to the light strips generated on the calibration body; these stripsare needed so that the evaluation unit can judge the dimensions of the
objects undergoing measurement; for this purpose, calibration bodies are
measured and the measurement data thus obtained are used to evaluate the
10 measured objects. It should also be noted that it is possible to project light
strips simultaneously onto the calibration bodies and onto the objects to be
measured and then to evaluate them. The process or arrangement according
to the invention thus permits optimal use to be made of the resolving power
of the image processing system which is used in the process or arrangement
15 according to the invention.
In a preferred embodiment of the invention, the imaging scale of the light
strip(s) imaged on the sensor element or the measuring field size can be
modifled by fitting the video cameras with ZOOM devicas which can be
adjusted, possibly by means of a motor drive, and/or by fitting the video
20 cameras with lenses of different focal lengths and/or by varying the
distance, possibly by means of a motor drive, between the video cameras
and the object to be measured or the calibration body. These are simple
ways of modifying the size of the images of the light strips; these devices
can be operated manually or automatically via the evaluation unit.
25 It is preferable when, for calibration purposes, at least one calibration body
of a given size is measured and the data obtained from the imaging of at
least this one calibration body on the sensor element are then stored in the
evaluation unit and are used to evaluate the data obtained from measuring
an object. In the process and arrangement according to the invantion the
30 measurements of the light sections imaged on the sensor element are used
in the evaluation unit to determine the enlargement or reduction scale
chosen for imaging the respective light strip on the object and/or calibration
~07 082~
body and these data are also taken into account when evaluatin~ the
camera signals. The devices used to adjust the imaging size or the imaging
scale or the measuring field size can be controlled by the evaluation unit,
namely by comparing the data from the calibration body with the data from
5 the object to be measured or as a function of size of ~he Images of the light
strips on the sensor element. This permits the creation of an almost fully
automatic measuring process having optimum accuracy characteristics.
In a preferred embodiment of the invention, provision is made for at least
one light section projected onto a calibration body or a calibration mark to
10 be imaged on the sensor element; and the size of the image of at least the
one light section is adapted to the size of the ~ensor element; and the light
strips generated on the caiibration body or at the calibration marks are
measured and the imaging scale or the measuring field size are determined;
and the data and parameters relating to the size of the light strips and the
15 imaging scale are stored in the evaluation unit; and when measuring at least
one light strip generated on an object, the size of the image of at least the
one light strip is adapted to the size of the sensor element; and by using the
same imaging scale or an imaging scale deviating therefrom in a known or
given manner, the object is measured using this imaging scale; and with the
20 aid of the imaglng parameters determined during the measuring and
calibration steps, the dimensions of the object are calculated.
In addltlon, in the process and arrangement according to the invention the
evaluatlon unit comprises a comparator for comparing the data and
parameters obtained from at least one measured calibration body with the
25 data and parameters supplied to the evaluation unit in the course of
measuring an object; and the evaluation unit is connected to the devices for
modifying the image scale or the measuring field size and feeds to them a
control signal, based on the results of the data comparison, for the purpose
of adjusting the imaging scale or the measuring field size.
30 The invention will be described in more detail in the following on the basis
of drawings. Fig. 1 shows in diagrammatic form the structure of an
i: ` 207a8~
-5-
arran~ement according to the invention. Fig. 2 depicts the principle behind
the measurement of an object or a calibration body. Figs. 3a, 3b, 3c depict
various arrangements for modifying the imaging scale and Figs. 4a, b, c and
d depict the measurement of calibration bodies.
5 Fig. 1 illustrates the principle of the measuring process. The object 4 to be
examined is illuminated by a number of light sources 5, in the present case
by four laser light sources 5, which may emit light of different wavelength.
The light beam from the laser 5 is spread into a plane, by means of an
optical system which is not depicted here, so that a bright outline or a
10 bright strip 6 is projected onto the object being measured. Each laser 5
projects a bright strip 6 and these usually partially overlap on the object.
Each light plane created by a laser 5 is advantageously oriented at right
angles to the longitudinal axis of the object ~angle i3); the light planes of the
individual lasers are oriented in such a way that they lie as far as possible ini 5 one and the same plane, so that the strips projected by the individual lasers
are as far as possible superimposed one on the other, or lie in a defined
plane intersecting the object, in order right from the start to avoid evaluationerrors based on positional inaccuracies.
The llght sections 6 projected by the lasers onto the object 4 are recorded
20 by solid-state cameras 7 arranged at an angle a relative to the optical axis or
light plane 21 of the laser 5 allocated to the respective camera 7. Each
camera 7 may be fitted with a filter 8 which allows only light having the
wavelength of the light emitted by the associated laser 5 to pass through,
so that each camera 7 can receive light only from the laser light source 5
25 with which it is paired. This prevents each camera 7 from being influenced
by light emitted by other light sources 5. At the same time, a very accurate
evaluation can be made of the outline of the strips 6 or the basic data of a
calibration object can be accurately evaluated. This also increases the
subsequent measuring accuracy.
30 Appropriate control wires for the light sources 5 or the cameras 7 are
indicated by the number 1 1; the control unit 10, which controls the on/off
., ~ .
: . ~
` 2Q70~2~
-6-
switching of the illumination and the cameras, can be coupled with the
evaluation unit 9 for the camera signals or operates in conjunction with that
unit.
The evaluation unit 9 comprises a computer, to which are fed the digitalized
5 video signals from the cameras 7 and which stores these signals. The
computer selects from the image matrix those image points which represent
the light section. When a measurement is performed, the positional data of
the object and the position and orientation of the camera relative to the light
plane and also the focal length of the lens are all known, so that the image
10 points which are found can be geometrically rectified and calculated back
into the actual coordinates of the object.
'
It should be noted that in favourable cases, two cameras alone are sufficient
to take the measurements; for round cross sections at least three cameras
are needed; and when four cameras are used, as shown in Fig. 3, almost all
15 customary structural sections with convex and concave cross sectional
shapes can be completely measured, as long as no undercuts exist.
The light sources 5 used may be white light sources, with appropriate
colour filters, lasers and laser diodes whose wavelengths or frequencies can
be adlusted, or similar. Appropriate optical systems are known for forming
20 the very narrow light sections on the object.
The cameras used may be video cameras, solid-state cameras, especially
CCD cameras, and cameras with sensor elements which are specifically
colour-sensitive or respond to certain colours, i.e. so-called colour cameras.
In particular, imaging is carried out using CCD cameras comprising a solid
25 state sensor element which is built up from about 500 x 500 photo diodes
and supplies substantially distortion-free images.
In order to evaluate the camera video signals stored in the evaluation unit 9,
the image is normally scanned for contour starting points by checking the
,
: ,
~; ~
2~7~824
-7-
brishtness contrasts of the image points and joining up the individual line
segments to form traverses. Once the appropriate traverses have been
determined, separately for each individual video camera signal, each
traverse is transformed into the coordinates of the object and then the
5 traverses obtained from the individual cameras are combined together to
give the overall contour from which the desired dimensions are calculated.
The imaging or rectification parameters are determined in the course of the
calibration process, for which purpose a calibration body is placed in the
measuring field of the cameras and the exact dimensions of the calibration
10 body are stored in the computer. When the light section of the calibration
body is recorded in the described manner, the rectification parameters can
be calculated by comparing the stored dimensions with the measured light
sectlons.
In Fig. 1, tha reference number 12 denotes the unit used to process the
15 signals from the individual cameras into the traverse sections recorded by
each carnera. Reference number 13 denotes the unit for combining the
individual traverses to form the outline or cross section of the object.
Reference number 14 is a monitor on which the measured object is
displayed, and 15 is a printer for printing out the measurements of the
20 object or other measurement data, and 16 is a plotter for graphically
reproducing the measured object.
-8- 2~7082~
Figure 2 depicts a spatial arrangement in which the object 4 is illuminated
by four lasers 5 and the tight sections 6 are recorded by CCD-cameras 7
fitted with filters 8. It can be seen that the optical axes 21 of the lasers 5
and the optical axes 22 of the cameras 7 enclose an angle a of 45, and
5 the planes sf laser light formed by each laser 5 lie in one common overall
plane .
.
Figs. 3a, 3b and 3c show various ways of modifying the imaging scale or
of adjusting the measuring field size of the cameras relative to the object to
be measured or the calibration object. In Fig. 3a a camera 7 is fitted with a
10 ZOOM lens 18 and it is shown that objects 23 of various size or light
sections or light strips 6 of various size, can be imaged in such a way, as
shown on the left and right-hand sides of Fig. 3a, that they fill the
measuring field 30 or 31 of the camera 7 to the maximum extent possible,
or that the size of the measuring field 30 or 31 is adapted to the objects to
15 be mcasured or to the light strips 6.
Fig. 3b shows an arrangement similar to that in Fig. 3a, wherein a lens-
changing device 18', e.g. a rotating lens turret system, is provided to adapt
or vary the imaging scale or the measuring fieid size of the video camera 7.
In Fig. 3c a camera-displacing device 18" is provided to adapt the measuring
20 field size by varying the distance between the camera 7 and the object 23,
although alternatively the object could be moved reiative to the camera 7 or
9 2070824
both the camera 7 and the object or calibration body 2~ could be moved
relative to each other.
The same improvement in accuracy when measuring objects can also be
achieved when measuring calibration bodies by matching the size of the
5 measuring field to the calibration body or to individual areas thereof. As
already msntioned, the measurement system is calibrated using a calibration
body of precisely defined shape. By comparing this defined shape with the
image data of the measured calibration body stored in the evaluation unit, it
is possible to obtain the necessary rectification parameters to rectify images
10 of objects which are to be measured. As is the case when measuring
objects, there are errors involved in determining the rectification parameters,
depending on the resolving power of the image processing system. As when
measuring objects, small errors or calibration errors can be reduced or
avoided when measuring calibration bodies, or when carrying out
15 callbratlon, by ensuring that the light strips generated on the calibration
body extend as far as possibie over the entire measuring field, or the image
or desired section of the image of the calibration body extends over the
~ntir~ sensor element.
If variable measuring ranges are used, as described above, then it is
20 advantageous, for measuring ranges with different imaging scaies, to use
dlfferent calibration bodies or calibration bodies with specific calibration
marks for different imaging scales. According to the invention, calibration
bodies for differently sized measuring fields or for different imaging scales
.
.. . .
,:
L- 2 0 7 0 8 2 4
10-
may possess different sections with marks indicating predetermined
locations and/or predetermined dimensions, and these sections are placed in
the measuring field or imaged on the sensor element and their known
dimensions are evaluated.
5 Fig. 4a shows in diagrarnmatic form the measuring of a calibration body 23'
bearing the calibration marks 26 and 26'. Fig. 4b shows a top view of the
calibration body 23' on which the peripheral calibration marks 26 and the
centrally located calibration marks 26' can be seen. The aimin~ directions or
optical axes of the four video cameras are denoted by the number 22. The
10 calibration body shown in Figs. 4a and 4b is designed for the simultaneous
calibration of four cameras. Calibration is carried out by forming appropriate
light sections 6 on the calibration marks 26 or 26', or the calibration marks
26 and 26' are intersected by the corresponding light planes 33. If the
image transmission system is on the "large scale" setting, i.e. if the ZOOM
15 lens is in the ~telephoto" position, then as shown in Fig. 4d, only the four
central calibration marks 26' are located in the measuring fieid and only the
light sections 24' of the central calibration marks 26' are imaged on the
sensor alement 34' of the video cameras 7. When a smaller imaging scale is
selected, e.g. when the ZOOM lens is in the wide-angie setting, all the
20 calibration marks 26 and 26' can be imaged, or the light sections 24 and
24' are imaged, on the sensor eiement 34', as shown in Fig. 4c. By
comparing the measuring data from the calibration body with stored data on
the dimensions of the caiibration marks, the evaluation system can
determine the imaging scale or the image fieid size; in addition, by
.
207 0824
measuring the light strips, the evaluation system can determine the size of
the object and measure the latter with the two imaging scales. In this way,
the measurement error and the calibration measurement error can be
reduced in absolute terms, and the measuring accuracy can be increased.
5 Advantageously, the imaging scale selected should be as large as possible.
The imaging scale used for the measurements should be known. Measuring
is carried out either at the same imaging scale used for calibration, or the
imaging scale used for the measurements is adjusted to a known value
either automatically by the evaluation unit or manually.
10 A further improvement in the measurement or calibration procegs is
achieved when a light strip is projected on an elongate object, e.g. a ruler or
measuring rod, and measurements are taken across the planes of the
measuring field. The object is then shifted to a parallel position and again
moasured; this process is repeated until light strips have be~n measured at
15 regular Intervals across the measuring field. The object (ruler) is then rotated
throu~h 90 and ths same process is repeated. This measurement grid, the
spacing of which can be varied depending on the measuring accuracy
desired, is used for calibration and is related to the image of the object or is
compared with the light strips projected onto the object. If certain areas of
20 the object have to be measured more accurately than others, a narrower
grid is formed in those particular areas. The formation of the grid or of the
calibràtion light strips can thus be locally varied over the m0asuring field or
image field.
,, ~ . , ,
; ` 2~7082~
-12-
Advantageously, according to the invention, the light emitted by the light
strip formed on the object is fed directly to the video cameras or is fed
directly or without any deflection of the beam path to the sensor ~10ment by
the imaging devices 18, 18', 18" provided. According to the invention, the
5 size of the light section(s) is directly matched to the size of the sensor
element, or vice versa, so that light losses are prevented and the evaluation
can be improved and made more accurate by an imaging scale, which can
be varied during the measurement process itself. It is thus possible, by
quickly modifying the imaging scale, to measure one and the same light
10 section in various sizes or in more or less image-filling forrn, o- to sxamine
desired detailed areas using a particular desired imaging scale. Since each
vldeo camera is equipped with such variable imaging units, it is also possible
to image and evaluate the light sections allocated to each video camera,
uslng different Imaging scales.