DE102021118429B4 - Process and device for 3D coordinate measurement using the autofocus process - Google Patents

Abstract

Method for 3D coordinate measurement using the autofocus method with the following steps: a) providing a camera (34) containing imaging optics (38) and an image sensor (36), the imaging optics (38) having an optical axis (52) and an object plane (56) which is imaged on the image sensor (36);b) illuminating a workpiece (26) to be measured with measuring light;c) generating a relative movement between the workpiece (26) and the object plane (56) along the optical axis ( 52) the imaging optics (38);d) recording images of the surface (40) of the workpiece (26) on the image sensor (36) at different times during the relative movement;e) storing the images recorded by the image sensor (36) during the relative movement ;f) calculating 3D coordinates of the surface (40) of the workpiece (26) by evaluating at least some of the images stored in step e); characterized in that during step d) a filter device (4; 44; 44', 44", 44''') attenuates at least part of first measuring light beams (48a) which emanate from the workpiece (26) at first angles with respect to a direction (z) parallel to the optical axis (52) which are smaller than a predetermined limit angle, and that the filter device (4; 44; 44', 44", 44''') does not attenuate at least part of the second measuring light beams (48b) or at least less than the first measuring light beams (48a), the second measuring light beams (48b) emanate from the workpiece (26) at second angles with respect to the direction (z) parallel to the optical axis (52), which angles are greater than the predetermined critical angle.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The invention relates to a method and a device for optical 3D coordinate measurement using the autofocus method. Such methods and devices take pictures of the surface of a workpiece from different distances. A 3D profile of the workpiece can be obtained from the sharpness information of the images.

2. Description of the Prior Art

In the prior art, coordinate measuring devices are used to measure the geometry of workpieces. Such measurements take place, for example, as part of quality assurance or so-called "reverse engineering". The sometimes complex measurement tasks are usually reduced to measuring the spatial coordinates of a number of individual points.

The coordinate measuring machines contain a sensor whose position relative to the workpiece to be measured can be changed with the aid of drives. In the case of smaller coordinate measuring machines, in particular, the workpiece is located on a cross table that can be moved with high accuracy along two horizontal coordinate axes x, y. The sensor is usually attached to a quill, which can be moved vertically (i.e. in the z-direction) with a similarly high level of accuracy. If particularly large or heavy workpieces are to be measured, coordinate measuring machines with a portal design are used, in which the workpiece is stationary and only the sensor is moved.

When it comes to sensors for coordinate measuring machines, a distinction is made between optical and tactile sensors. While with tactile sensors the information about the position of a measuring point is generated by touching the measuring point with a probing element, with optical sensors the information about the position of the measuring point is transmitted by light.

In one type of optical sensor of this type, the z-coordinates of the workpiece surface are determined using the autofocus method. For this purpose, the sensor has a camera with imaging optics and an image sensor. If the quill with the sensor attached to it is moved vertically along the optical axis of the imaging optics, a sharp image of those areas of the surface of the workpiece that are at the same height is only generated in a z-position of the camera. Outside this position, the image sensor is out of focus, so that these surface areas are blurred in the recorded images.

The contrast is usually used as a parameter for the sharpness of the images created on the image sensor. The contrast for a specific point on the surface of the workpiece is at its maximum when the point is located exactly in the object plane of the imaging optics of the camera. Since the position of the object plane in a reference system of the coordinate measuring machine is known with a high level of accuracy, the distance to this point and thus its z-coordinate can be measured very precisely in this way. The faster the contrast falls as the point moves out of the object plane, the higher the accuracy of this measurement.

In order to carry out a 3D coordinate measurement for a specific workpiece using the autofocus method, a relative movement between the workpiece and the object plane along the optical axis of the imaging optics must be generated. For this purpose, either the workpiece, the camera or both can be moved in the z-direction. Alternatively or in addition to this, the position of the object plane can be changed by moving individual lenses of the imaging optics or by using an optical element with variable refractive power (e.g. a liquid lens). The images of the workpiece surface recorded at different points in time during the relative movement are stored in order to then calculate the 3D coordinates of the workpiece surface by determining the contrast. Edges and other contours can also be determined from the images, in which certain areas of the workpiece surface are reproduced sharply, with the aid of image processing algorithms and correlated with the measured z-coordinates.

Coordinate measuring machines that calculate the z-coordinates according to the principle of the autofocus method explained above are sold by the applicant under the brand name ZEISS O-INSPECT and are, for example, in DE 10 2016 202 928 A1 described.

As already mentioned, the sensitivity of the autofocus method and thus the measurement accuracy in the z-direction depend above all on how quickly the contrast falls when the point moves out of the object plane. The smaller the axial range in which points on the workpiece surface are sharp, the greater the measurement accuracy. This one also as Schär The area designated as depth depends directly on the numerical aperture of the imaging optics. The larger the numerical aperture of the imaging optics, the smaller the depth of field and the more accurate the z-coordinates measured using the autofocus method.

However, if you want to increase the numerical aperture for a given size of the measuring field (ie the object field that can be imaged on the image sensor in the object plane of the imaging optics), this requires larger imaging optics. However, optics with larger diameters are significantly more expensive. In addition, larger optics make the sensor heavier overall. This, in turn, places higher demands on the mechanical components of the coordinate measuring machine, which are needed for the precise guidance of the sensor.

If the numerical aperture is to be increased while the size of the imaging optics remains the same, this is only possible if the size of the measuring field is reduced. However, this is also disadvantageous since small measuring fields lead to longer measuring times.

From the DE 10 2014 223 387 A1 a method for 3D coordinate measurement is known, in which a pupil filter is used, which has a higher transparency at the edge than in the middle of the pupil. During the recording of a preliminary image, the focal plane performs comparatively high-frequency modulations around a resting position. As a result, blurred image contributions that occur during image integration are removed and thus an extension of the depth of field is achieved.

The DE 10 2007 022 218 A1 discloses the use of different apertures in a measuring camera.

SUMMARY OF THE INVENTION

The object of the invention is therefore to specify a method and a device for 3D coordinate measurement using the autofocus method, in which the measurement accuracy in the z-direction is increased without this having a negative effect on the size of the imaging optics or the measurement field.

1. a) Bereitstellen einer Kamera, die eine Abbildungsoptik und einen Bildsensor enthält, wobei die Abbildungsoptik eine optische Achse und eine Objektebene hat, die auf den Bildsensor abgebildet wird;
2. b) Beleuchten eines zu vermessenden Werkstücks mit Messlicht;
3. c) Erzeugen einer Relativbewegung zwischen dem Werkstück und der Objektebene entlang der optischen Achse der Abbildungsoptik;
4. d) Aufnehmen von Bildern der Oberfläche des Werkstücks auf dem Bildsensor zu unterschiedlichen Zeitpunkten während der Relativbewegung;
5. e) Speichern der vom Bildsensor während der Relativbewegung aufgenommenen Bilder;
6. f) Berechnen von 3D-Koordinaten der Oberfläche des Werkstücks durch Auswertung zumindest eines Teils der in Schritt e) gespeicherten Bilder;

wobei
während des Schritts d) eine im Strahlengang des Messlichts angeordnete Filtereinrichtung zumindest einen Teil von ersten Messlichtstrahlen abschwächt, die von dem Werkstück unter ersten Winkeln bezüglich einer zur optischen Achse parallelen Richtung ausgehen, die kleiner sind als ein vorgegebener Grenzwinkel, und wobei
die Filtereinrichtung zumindest einen Teil von zweiten Messlichtstrahlen nicht oder zumindest weniger abschwächt als die ersten Messlichtstrahlen, wobei die zweiten Messlichtstrahlen von dem Werkstück unter zweiten Winkeln bezüglich der zur optischen Achse parallelen Richtung ausgehen, die größer sind als der vorgegebene Grenzwinkel.With regard to the method, this task is solved by a method with the following steps:

1. a) providing a camera containing imaging optics and an image sensor, the imaging optics having an optical axis and an object plane which is imaged onto the image sensor;
2. b) illuminating a workpiece to be measured with measuring light;
3. c) generating a relative movement between the workpiece and the object plane along the optical axis of the imaging optics;
4. d) recording images of the surface of the workpiece on the image sensor at different points in time during the relative movement;
5. e) storing the images recorded by the image sensor during the relative movement;
6. f) calculating 3D coordinates of the surface of the workpiece by evaluating at least some of the images stored in step e);

whereby
during step d) a filter device arranged in the beam path of the measuring light attenuates at least part of the first measuring light beams, which emanate from the workpiece at first angles with respect to a direction parallel to the optical axis, which are smaller than a predetermined critical angle, and wherein
the filter device does not attenuate at least part of the second measuring light beams or at least less than the first measuring light beams, wherein the second measuring light beams emanate from the workpiece at second angles with respect to the direction parallel to the optical axis, which angles are greater than the specified critical angle.

Expressed in simple terms, according to the invention, measuring light beams entering the imaging optics at small angles are weakened compared to the measuring light beams entering at an angle. The result of this is that measuring light beams that correspond to a smaller numerical aperture and for which the depth of field is therefore the greatest do not contribute to the imaging, or only to a lesser extent. The weakening of this proportion of the measuring light beams thus increases the contribution of the measuring light beams to the image, for which the depth of field is smaller, in relative terms.

The inventor has recognized that the depth of field is already reduced by this shift in the weighting without the numerical aperture having to be increased in comparison to conventional imaging optics. When measuring the z-coordinates, this is noticeable in that the contrast during the relative axial movement between the workpiece and the object plane changes more per length unit than if the measuring light beams entering the imaging optics at small angles are used, as shown in to can contribute to the image for a long time without weakening.

The greater the attenuation, the stronger the effect described above on the depth of field and thus the favorable influence on the accuracy of the measurement of the z-coordinate. It is therefore preferred if the filter device maximally weakens all first measuring light beams--which is tantamount to complete dimming--and does not weaken all second measuring light beams at all. However, filter devices with a graduated filter that produces an attenuation that increases continuously from large to small angles can also be considered.

In one exemplary embodiment, the filter device comprises a filter element that acts in transmission or reflection, which is arranged in a pupil plane of the imaging optics or at least in the vicinity of the pupil plane and has a transmittance or reflectance that depends on the location. The pupil plane is also often referred to as the aperture plane and is characterized by the fact that the principal rays intersect the optical axis there.

In the case of telecentric imaging optics, the pupil plane is located in an intermediate space between two partial objectives. The filter element can then also be arranged in this intermediate space. Non-telecentric imaging optics can consist of just one lens or a group of lenses, in front of or behind which the pupil plane is located.

Since all measuring light beams that emanate from the workpiece at a certain angle pass through the pupil plane at the same place, the pupil plane represents the ideal place for achieving the desired attenuation of the first measuring light beams with a simple filter element with a location-dependent degree of transmission or reflection. Since the above-mentioned relationship between angles in the object plane and locations in the pupil plane also applies to a good approximation in the vicinity of the pupil plane, the filter element does not have to be arranged exactly in the pupil plane of the imaging optics, but can be arranged in the axial direction relative to the pupil plane by one be offset by a certain amount.

In the simplest case, the filter element is a central circular aperture which completely obscures a circular area of the pupil plane with a specific diameter. Another example of a suitable filter element is a mirror, the inner area of which is absorbing or faceted in such a way that the measuring light beams impinging there are deflected out of the beam path of the measuring light.

As an alternative to a filter element arranged close to the pupil, the filter device can comprise a filter element which acts in transmission or reflection and which is arranged in a space between the workpiece and the imaging optics or in a space between the image sensor (36) and the imaging optics (38) and at an angle has dependent transmittance or reflectance. Such filter elements can be designed, for example, as plate-shaped glass substrates that are provided with special multi-layer coatings that produce the angular dependency.

Due to the inventive weakening of the first measuring light beams, less light reaches the image sensor. This is usually not a problem for distance measurement. In some applications, however, it may be desirable for additional images of the surface of the workpiece to be recorded at least in some relative positions between the workpiece and the object plane, in which case the filter device does not attenuate the first measuring light beams, or at least not more than the second measuring light beams.

These additional images are therefore not used for the distance measurement in step f), but for other evaluation purposes. For example, these additional images can be used as a basis for image processing to detect edges or contours.

In order to be able to record images in which the first measuring light beams contribute to the image to a different extent, the filter device should be able to switch between a first mode in which it attenuates the first measuring light beams and the second measuring light beams to different extents, and a second mode in which they does not weaken the first measuring light rays or weakens them just as little as the second measuring light rays. Ideally, the filter device can be switched between the two modes very quickly. It is then possible for the switching to take place during a continuous relative movement between the object plane and the workpiece. Examples of such rapidly switchable filter devices are LCD panels or switchable micromirror arrays such as DMDs (digital mirror devices).

If switching takes a long time, as is the case, for example, with transmission filters that can be inserted into the beam path, the additional images can be measured in a separate measurement process gang in which only pictures are taken in the second mode.

1. a) einer Kamera, die eine Abbildungsoptik und einen Bildsensor enthält, wobei die Abbildungsoptik eine optische Achse und eine Objektebene hat, die auf den Bildsensor abgebildet wird,
2. b) einer Beleuchtungseinrichtung, die dazu eingerichtet ist, ein zu vermessendes Werkstück mit Messlicht zu beleuchten,
3. c) einem Antrieb oder einem verstellbaren optischen Element zum Erzeugen einer Relativbewegung zwischen dem Werkstück und der Objektebene entlang der optischen Achse der Abbildungsoptik,
4. d) einem Speicher zum Speichern von Bildern der Oberfläche des Werkstücks, die von dem Bildsensor zu unterschiedlichen Zeitpunkten der Relativbewegung aufgenommen wurden,
5. e) einer Recheneinheit, die dazu eingerichtet ist, 3D-Koordinaten der Oberfläche des Werkstücks durch Auswertung zumindest eines Teils der in Schritt e) gespeicherten Bilder zu berechnen,
6. f) einer im Strahlengang des Messlichts angeordnete Filtereinrichtung, die dazu eingerichtet ist, während der Aufnahme von Bildern der Oberfläche des Werkstücks
   • - zumindest einen Teil von ersten Messlichtstrahlen abzuschwächen, die von dem Werkstück unter ersten Winkeln bezüglich einer zur optischen Achse parallelen Richtung ausgehen, die kleiner sind als ein vorgegebener Grenzwinkel, und
   • - zumindest einen Teil von zweiten Messlichtstrahlen nicht oder zumindest weniger abschwächt als die ersten Messlichtstrahlen, wobei die zweiten Messlichtstrahlen von dem Werkstück unter zweiten Winkeln bezüglich der zur optischen Achse parallelen Richtung ausgehen, die größer sind als der vorgegebene Grenzwinkel.

With regard to the device, the above task is solved by a device for 3D coordinate measurement using the autofocus method, with:

1. a) a camera containing imaging optics and an image sensor, the imaging optics having an optical axis and an object plane which is imaged on the image sensor,
2. b) an illumination device that is set up to illuminate a workpiece to be measured with measuring light,
3. c) a drive or an adjustable optical element for generating a relative movement between the workpiece and the object plane along the optical axis of the imaging optics,
4. d) a memory for storing images of the surface of the workpiece, which were recorded by the image sensor at different times of the relative movement,
5. e) a computing unit that is set up to calculate 3D coordinates of the surface of the workpiece by evaluating at least some of the images stored in step e),
6. f) a filter device arranged in the beam path of the measuring light, which is set up to record images of the surface of the workpiece
   • - to attenuate at least part of first measuring light beams emanating from the workpiece at first angles with respect to a direction parallel to the optical axis, smaller than a predetermined critical angle, and
   • - At least some of the second measuring light beams are not weakened or at least less weakened than the first measuring light beams, the second measuring light beams emanating from the workpiece at second angles with respect to the direction parallel to the optical axis, which angles are greater than the specified critical angle.

The filter device can be set up to completely screen out all first measuring light beams and not to attenuate all second measuring light beams.

The computing unit can be set up to control the device in such a way that at least in some relative positions between the workpiece and the object plane, additional images of the surface of the workpiece are recorded, in which the filter device does not detect the first measuring light beams or at least not more than the second measuring light beams weakens.

For this purpose, the computing unit can control the filter device in such a way that the filter device changes between a first mode in which it attenuates the first measuring light beams and the second measuring light beams to different extents and a second mode in which it does not attenuate the first measuring light beams or attenuates them just as little like the second measuring light beams.

The adjustable optical element can be, for example, an axially movable lens group or individual lens, a deformable mirror, a liquid lens or a switchable Fresnel lens.

character list

• 1: eine perspektivische Darstellung eines erfindungsgemäßen Koordinatenmessgeräts;
• 2: einen meridionalen Schnitt durch einen Sensor des Koordinatenmessgeräts und ein auf einem Kreuztisch befestigtes Werkstück;
• 3a und 3b: einen Strahlengang mit kleiner numerischer Apertur zum Messen von z-Koordinaten an unterschiedlichen Orten auf dem Werkstück;
• 4a und 4b: den 3a und 3b entsprechende Darstellungen für einen Strahlengang mit großer numerischer Apertur;
• 4c: der in der 4b gezeigte Strahlengang nach Verfahren der Abbildungsoptik in z-Richtung;
• 5a: eine den 3a und 4a entsprechende Darstellung für einen Strahlengang mit mittlerer numerischer Apertur;
• 5b: eine den 3b und 4b entsprechende Darstellung für einen Strahlengang mit mittlerer numerischer Apertur, wobei erfindungsgemäß Messlichtstrahlen, die unter kleinem Winkel zur Richtung der optischen Achse verlaufen, abgeschwächt werden;
• 6: ein Diagramm, auf dem die Intensitätsverteilung eines Lichtpunkts auf dem Bildsensor in unterschiedlichen Abständen z zwischen der Objektebene und der Oberfläche des Werkstücks für den in der 5a gezeigten Fall dargestellt ist;
• 7: ein der 6 entsprechendes Diagramm für den in der 5b gezeigten Fall;
• 8: der in der 2 gezeigte Sensor, wobei sich die Filtereinrichtung in einem ersten Modus für die Abstandsmessung befindet, in dem Messlichtstrahlen abgeschwächt werden;
• 9: einen der 2 entsprechenden meridionalen Schnitt durch einen Sensor gemäß einem zweiten Ausführungsbeispiel, bei dem die Filtereinrichtung als digitales Mikrospiegelarray ausgebildet ist und sich in einem zweiten Modus befindet, in dem keine Messlichtstrahlen abgeschwächt werden;
• 10: der in der 9 gezeigte Sensor, wobei sich die Filtereinrichtung im ersten Modus befindet, in dem Messlichtstrahlen abgeschwächt werden;
• 11: einen der 2 entsprechenden meridionalen Schnitt durch einen Sensor gemäß einem dritten Ausführungsbeispiel, bei dem das Filterelement eine fest in einer Pupillenebene des Sensors angeordneten Kreisblende umfasst;
• 12: eine Draufsicht auf die in der 11 gezeigte Kreisblende;
• 13: einen der 2 entsprechenden meridionalen Schnitt durch einen Sensor gemäß einem vierten Ausführungsbeispiel, bei dem die Filtereinrichtung als winkelabhängiges Filterelement ausgebildet ist;
• 14: einen Graphen, auf dem die Abhängigkeit des Transmissionsgrades T vom Einfallswinkel α schematisch dargestellt ist.

Exemplary embodiments of the invention are explained in more detail below with reference to the drawings. In these show:

• 1 : a perspective view of a coordinate measuring machine according to the invention;
• 2 : a meridional section through a sensor of the coordinate measuring machine and a workpiece fixed on a cross table;
• 3a and 3b : a small numerical aperture beam path for measuring z-coordinates at different locations on the workpiece;
• 4a and 4b : the 3a and 3b corresponding representations for a beam path with a large numerical aperture;
• 4c : the Indian 4b Beam path shown after moving the imaging optics in the z-direction;
• 5a : a den 3a and 4a Corresponding representation for a beam path with a mean numerical aperture;
• 5b 3b and 4b: a representation corresponding to FIGS. 3b and 4b for a beam path with a mean numerical aperture, with measuring light beams running at a small angle to the direction of the optical axis being attenuated according to the invention;
• 6 : a diagram showing the intensity distribution of a point of light on the image sensor at different distances z between the object plane and the surface of the workpiece for the in the 5a case shown;
• 7 : one of 6 corresponding diagram for the one in the 5b shown case;
• 8th : the Indian 2 Sensor shown, wherein the filter device is in a first mode for the distance measurement, are attenuated in the measuring light beams;
• 9 : one of 2 corresponding meridional section through a sensor according to a second exemplary embodiment, in which the filter device is designed as a digital micromirror array and is in a second mode in which no measuring light beams are attenuated;
• 10 : the Indian 9 Sensor shown, wherein the filter device is in the first mode are attenuated in the measuring light beams;
• 11 : one of 2 corresponding meridional section through a sensor according to a third exemplary embodiment, in which the filter element comprises a circular diaphragm fixedly arranged in a pupil plane of the sensor;
• 12 : a top view of the in the 11 circular aperture shown;
• 13 : one of 2 corresponding meridional section through a sensor according to a fourth exemplary embodiment, in which the filter device is designed as an angle-dependent filter element;
• 14 : a graph on which the dependency of the transmittance T on the angle of incidence α is shown schematically.

DESCRIPTION OF PREFERRED EMBODIMENTS

1. Structure of the coordinate measuring machine

The 1 shows a coordinate measuring machine, designated overall by 10, in a schematic perspective representation.

The coordinate measuring machine 10 includes a base 12 which supports a table 14 to which a control panel 16 is attached. A stand 18 which carries a quill 20 extends upwards from the table 14 . As indicated by a double arrow 22, the quill 20 can be moved precisely in the vertical direction (z-direction) with the aid of a drive (not shown).

An optical sensor 24 is fastened to the underside of the quill 20, with which an image of a workpiece 26 can be recorded. The workpiece 26 is mounted on a compound table 28 with which the workpiece 26 can be moved precisely in the horizontal plane (x-direction and y-direction), as is shown in FIG 1 is indicated by arrows 30 and 32, respectively. In this way, it is also possible to successively measure larger workpieces 26 with the aid of the sensor 24 by gradually introducing the workpiece 26 into the measuring field of the sensor 24 with the aid of the compound table 28 .

If larger or particularly heavy workpieces 26 are to be measured, the coordinate measuring machine 10 can also have a different mechanical design and, for example, instead of the cross table 28, have a movable portal to which the quill 20 is attached. In this way, the quill 20 can be precisely moved not only along the z-direction, but also along the x-direction and y-direction, as is known per se in the prior art. The workpiece 26 then does not have to be moved during the measurement.

2. Structure of the sensor

The 2 shows the structure of the sensor 24 according to a first exemplary embodiment in a simplified meridional section. The sensor 24 includes a camera 34 with a two-dimensional image sensor 36, which can be a conventional CCD sensor or CMOS sensor. The camera 34 also includes imaging optics 38 in the 2 is indicated for the sake of clarity by only two lenses L1 and L2. The imaging optics 38 can be microscope optics that generate an enlarged image of the surface 40 of the workpiece 26 to be measured on the image sensor 36 . In the illustrated embodiment, the increase is β = -1.5; typically, however, the magnification of the imaging optics 38 is significantly greater in terms of absolute value. Instead of dioptric imaging optics, catadioptric imaging optics (ie containing refracting and reflecting surfaces) can also be used.

The sensor 24 also includes a filter device with a filter element which is arranged in a pupil plane 42 of the imaging optics 38 and is designed as an LCD panel 44 in the illustrated exemplary embodiment. For this purpose, the LCD panel 44 comprises a multiplicity of liquid crystal cells 46 arranged in a matrix-like manner, the transmittance of which can be changed individually between a minimum value close to 0% and a maximum value close to 100%. In the mode shown, all liquid crystal cells 46 are maximally translucent, so that all measuring light beams 48a and 48b, which emanate from the surface 40 of the workpiece 26 and enter the imaging optics 38, reach the image sensor 36 without any significant attenuation.

Below the camera 34 is an illumination device 50 which, in the exemplary embodiment shown, comprises LEDs arranged in a ring around the optical axis 52 . The lighting device 50 can be designed, for example, in such a way that it can direct light with different colors and/or from adjustable directions onto the surface 40 of the workpiece 26 . An example of a suitable lighting device 50 is in FIG WO 2013/167168 A1 described by the applicant. In addition to the incident light illumination, a further illumination device can be provided, with which the measuring light is coupled into the beam path of the imaging optics 38 via a coupling mirror in order to be able to illuminate the measuring field directly from above and approximately axially parallel. Illumination in transmitted light is also possible with certain designs.

At 53 is in the 2 a drive is indicated, with which the quill 20 with the sensor 24 attached to it can be moved vertically along the z-direction, as indicated by the double arrow 22 . Because the quill 20 in the 2 not shown, the drive 53 is shown as acting directly on the sensor 24 . In reality, however, drive 53 acts on sleeve 20.

The image sensor 36, the LCD panel 44, the lighting device 50, the drive 53 and the cross table 28 are connected to a processor 54 containing a memory 55, which controls the measuring sequence in a manner to be explained in more detail below and processes the images recorded by the image sensor 36 .

3. function

To explain the function of the sensor 24 is first on the 3a and 3b referenced.

The 3a shows the surface 40 of the workpiece 26, which is illuminated by the illumination device 50 with measuring light. The surface 40 of the workpiece 26 reflects the incident measuring light diffusely in all spatial directions. However, only a small part of the reflected measuring light contributes to the imaging of the surface 40 on the image sensor 36 . This is because only those reflected measuring light beams 48 that emanate from the workpiece 26 at an angle with respect to a direction parallel to the optical axis 52 (in this case the vertical) that is smaller than the object-side opening angle specified by the numerical aperture NA of the imaging optics 38 can of the imaging optics 38 and reach the image sensor 36 . In the 3a it is assumed that the numerical aperture NA and therefore also the object-side aperture angle is small. Therefore, only those measuring light beams 48 that emanate from the surface 40 of the workpiece 26 at small angles with respect to the vertical contribute to the imaging.

At 56 is in the 3a the object plane of the imaging optics 38 is indicated. A point P on the surface 40 is considered, which has the coordinates x, y in the measuring field of the sensor 24 and from which the two measuring light beams 48 emanate. In the constellation shown, the point P is located exactly in the object plane 56. The point P is therefore imaged sharply and with maximum contrast on the image sensor 36.

If the workpiece 26 is now moved along the x-direction by a certain distance, as is shown in FIG 3a is indicated by an arrow 58, another point P' on the surface 40 arrives at the location x, y of the measuring field. It is assumed that the point P' is not at the level of the object plane 56 but, for example, slightly above it, as is shown in FIG 3b is shown. Consequently, this point P′ is no longer imaged on the image sensor 36 with maximum sharpness.

However, the loss of sharpness and the corresponding reduction in contrast is very small in this case. This is due to the fact that the measuring light beams 48 enclose only a small angle to the vertical. How to by comparing the 3a and 3b can recognize, the point P' on the surface 40 is still approximately at the point of intersection of the two measuring light beams 48, so that the loss of sharpness on the image sensor 36 is hardly perceptible. The measuring light beams 48 were, in order to make the position of their point of intersection more recognizable, in the 3a and 3b and also in the following ones 4 and 5 extended downwards through the workpiece 26, although of course no measuring light usually propagates there.

The imaging with a small numerical aperture thus causes a large depth of field. The image on the image sensor 36 therefore remains comparatively sharp when the surface 40 is no longer positioned exactly in the object plane. However, a large depth of field is undesirable for the distance measurement using the sensor 24, since the contrast on the image sensor 36 changes only slightly when the distance between the surface 40 and the object plane 56 varies.

For the distance measurement is cheaper in the 4a until 4c illustrated case, in which the imaging optics 38 has a large numerical aperture NA and a correspondingly large object-side opening angle. The measuring light beams 48 can now enter the imaging optics 38 at a large angle to the vertical. Like from one comparison of 4a and 4b with the 3a and 3b becomes clear, the loss of contrast of the image on the image sensor 36 is significantly greater in this case when the surface 40 leaves the object plane 56. Because measuring light beams 48, which converge exactly on the image sensor 36, no longer emanate approximately from the point P' on the surface 40 of the workpiece 26, but - in the example shown - from an imaginary point below the surface 40, as in the enlarged section of the 4b is recognizable. The large numerical aperture NA of the imaging optics 38 thus leads to a small depth of field, as is desired for the distance measurement.

Due to the low contrast on the image sensor 36 in FIG 4b In the constellation shown, the sensor 24 can detect that the surface 40 is outside of the object plane 56 . If the sleeve 20 is now moved upwards in the z-direction with the aid of the drive 53, the object plane 56 also moves upwards by the same amount. Once the object plane 56 is exactly level with point P', as shown in FIG 4c is shown, a maximum sharp image with the greatest possible contrast is achieved again. Since the position of the object plane 56 in a reference system of the coordinate measuring device 10 is precisely known, the z-coordinate of the point P′ on the surface 40 can be determined exactly in this way.

During the measurement, a relative movement is thus generated continuously or intermittently between the workpiece 26 and the object plane 56 along the optical axis 52 of the imaging optics 68 by the sleeve 20 being moved along the z-direction with the aid of the drive 53 . During this relative movement, several images of the surface 40 of the workpiece 28 are recorded at different times with the aid of the camera 34 and stored in the memory 55 of the computing unit 54 .

All areas of the surface 40 that are within the lateral measuring field and the axial measuring range and have the same z-coordinate are imaged sharply on the image sensor 36 at a specific point in time during the relative movement between the quill 20 and the workpiece 26. A three-dimensional profile of the surface 40 of the workpiece 26 can be obtained in this way, since the x and y coordinates of the measuring points can be derived directly from the position on the recorded images.

In order to increase the numerical aperture NA for a given size of the measuring field and thereby reduce the depth of focus, the imaging optics 38 would have to be significantly larger, heavier and more expensive. If the size of the imaging optics 38 remains the same, the numerical aperture NA can otherwise only be increased at the expense of the size of the measuring field. Smaller measuring fields are disadvantageous, however, because this significantly increases the measuring time for larger workpieces. In practice, one usually chooses smaller numerical apertures NA between 0.1 and 0.2, which already ensure sufficient lateral resolution of the images.

In the 5a is in a to the 3a and 4a Ajar representation shown again the surface 40 of the workpiece 26. First measuring light beams 48a emanate from a point P on the workpiece 26 at first angles with respect to a direction parallel to the optical axis (here again the vertical), which are smaller than a predetermined critical angle. For these first measuring light beams 48a, the depth of field is large, as is explained above with reference to FIG 3a and 3b was explained.

Also shown are second measuring light beams 48b, which emanate from point P on surface 40 at second angles with respect to the vertical, which are greater than the specified critical angle. For these second measuring light beams 48b, which run more obliquely than the first measuring light beams 48a, the depth of field is smaller than for the first measuring light beams 48a.

According to the invention, the first measuring light beams 48 are attenuated by a filter device, as is shown in FIG 5b is indicated. The attenuation is maximum here and is indicated by an aperture 4, which completely absorbs the first measuring light beams 48a. The critical angle is indicated by broken lines 60 . As a result of the absorption of the first measuring light beams 48a, only the more inclined second measuring light beams 48b, for which the depth of field is smaller, contribute to the imaging of the point P on the surface 40 . In this way, an image is obtained with a depth of field that is less than if the first measuring light beams 48a would also contribute to the image.

This difference is in the 6 and 7 illustrated. Each curve represents locations of the same intensity in an image generated on the image sensor 36 of a point on the surface 40 of the workpiece 26. The radial distance r from the center of the image point is on the abscissa and the distance of the location from the object plane is on the ordinate 56 plotted in the z-direction. The curves thus reflect the three-dimensional course of the intensity of a rotationally symmetrical aerial photograph.

The 6 shows this aerial image without attenuation of the first measuring light beams 48a. As expected, the lateral dimension r of the pixel is minimal when the surface is located exactly in the object plane 56 (z=0). The intensity distribution widens as the distance from the object plane 56 increases. However, the intensity in the center (r=0) falls only very slowly with increasing distance z from the object plane 56 . This can be seen from the fact that the lines of equal intensity shown in the graph are relatively far apart.

The 7 shows the situation when the first measuring light beams 48a are attenuated in the manner according to the invention. Assuming full attenuation here, as shown in the 5b is indicated. In comparison to 6 is in the 7 It can be seen that the initially disk-shaped intensity distribution is converted into an annular intensity distribution as the distance z from the object plane 56 increases. This is accompanied by the fact that in the center of the image point (r=0) the intensity decreases significantly more as the distance from the object plane 56 increases (ie z increases in absolute value) than in the case of the one in FIG 6 shown case. As can be seen, the lines of constant intensity are much closer together at the position indicated by a dashed vertical line in the center of the image point (r = 0) than in the 6 . Due to this greater intensity gradient, the measurement accuracy when measuring the z-coordinate is significantly higher than in the prior art, as is the case with 5a and 6 is illustrated.

The one in the 5b The arrangement of the aperture 4 shown in the beam path between the workpiece 26 and the sensor 24 would of course only be possible if only a single measuring point is to be measured at a given point in time or the image field is very small compared to the working distance of the optics. As a rule, however, it is desirable to determine the z-coordinate for all points on the surface 40 of the workpiece 26 that are in the measuring field. In that case, of course, no screens can be arranged between the workpiece 26 and the sensor 24 because no light would then reach the imaging optics 38 at all.

The weakening of the first measuring light beams 48a is, however, easily possible in the pupil plane 42 of the imaging optics 38 . Since all measuring light beams 48a, 48b, which emanate from the object plane 56 at a specific angle, pass through the pupil plane 42 at one location, an angle-dependent weakening of the measuring light beams 48 can be carried out there in a simple manner.

The 8th shows the in the 2 sensor shown, but with activated LCD panel 44. As can be seen from the drawn beam path, the obliquely running second measuring light beams 48b only penetrate an outer annular area of the pupil plane 42. Axis-parallel first measuring light beams 48a cross the optical axis 52 in the pupil plane 42. The critical angle 60 corresponds in the pupil plane 42 to a radius r _G ..

In order to maximally weaken the first measuring light rays 48a, the computing unit 54 controls the liquid crystal cells 46 of the LCD panel 44 in such a way that all liquid crystal cells 46 within the radius r _G reduce their transmittance to a maximum (T ≈ 0), which is 8th is indicated by a blackening. The remaining liquid crystal cells 46 are controlled by the arithmetic unit 54 in such a way that they are maximally transparent to the measuring light passing through (T≈100%). In this way, only the second measuring light beams 48b, the angle of which is greater than the critical angle, contribute to the imaging of the surface 40. As a result, in the manner explained above, with an unchanged numerical aperture NA and an unchanged size of the measuring field, a reduction in the depth of field and, associated therewith, a greater measuring accuracy in the z-direction is achieved.

Due to the activation of the LCD panel 44 described above, part of the measurement light reflected by the workpiece 26 is obscured and thus does not contribute to the imaging. This also reduces the brightness of the images captured by the image sensor 36 . Above all in the case of more strongly absorbent surfaces 40 of the workpiece 26, this can lead to errors occurring more frequently in an image processing that is additionally carried out, for example for the purpose of detecting edges or profiles.

In order to avoid this, additional images can be recorded of the surface 40 of the workpiece 26 in which the LCD panel 44 does not attenuate the first measuring light beams 48a or at least not more than the second measuring light beams 48b. For this purpose, the computing unit 54 controls the LCD panel 44 in such a way that all liquid crystal cells 46 allow the measurement light to pass through with maximum transmittance, as is shown in FIG 2 is shown.

Since the switching of the LCD panel 44 between a first mode in which the first measuring light beams 48a and the second measuring light beams 48b are attenuated to different degrees, and a second mode in which the first measuring light beams 48a are not attenuated or are attenuated just as little as the second measuring light beams 48b, in which the LCD panel 44 can be done very quickly, can be in the different modes captured images in quick succession. As a result, images with a large and a small depth of field can be recorded alternately or quasi-simultaneously in a single displacement movement of the quill 20 .

4. Second embodiment

The 9 shows in a the 2 and 8th based on a second exemplary embodiment of a sensor 24 according to the invention.

While at the in the 2 and 8th shown first embodiment, the filter element is designed as an LCD panel 44 with a location-dependent transmittance is located at the in the 9 shown second embodiment in the pupil plane 42 a digital mirror array 44 ', which contains a plurality of matrix-like arranged and individually controllable micro-mirrors 64. The mirror array 44' can be in the form of a DMD (digital mirror device), such as is used in projectors, for example. A plane mirror 63, which folds the beam path by 90°, is arranged in the beam path between the pupil plane 42 and the second lens L2. There in the in the 9 shown neutral position, in which the surfaces of all micromirrors 64 are aligned parallel to one another, and the mirror array 44' also folds the beam path by 90°, the optical axis 52 - apart from the section between the mirror array 44' and the plane mirror 63 - remains vertically aligned.

In the in the 9 shown neutral position of the mirror array 44 'there is no weakening of the first measuring light beams 48a. If the micromirrors 64 arranged in the center of the pupil are folded down at the instigation of the arithmetic unit 54, as is the case 10 illustrated, the first measuring light beams 48a striking it no longer reach the lens L1 and thus do not contribute to the imaging. Thus, in this exemplary embodiment as well, a quick change between two modes can be carried out, in which the first measuring light beams 48a contribute to the imaging or not.

5. Third embodiment

The 11 shows in a the 2 and 8th based representation a third embodiment with a particularly simple structure filter device.

The filter device does not have a switchable filter element arranged in the pupil plane 42 here. Instead, the filter element there is formed by a simple circular aperture 44" with a circular circumference, which is arranged centered on the optical axis 52. Like the top view of the circular aperture 44" according to FIG 12 shows, this is held by four narrow webs 70 in the housing of the camera 34, the meridional section of the 11 however, are not shown. The portion of the second measuring light beams 48b that are absorbed by the webs 70 is negligibly small.

The circular diaphragm 44" preferably remains permanently in the beam path, so that no images can be recorded in which the first measuring light beams 48a also contribute to the imaging. Alternatively, provision can be made for the circular diaphragm 44" to be automatically removed from the beam path with the aid of an exchange mechanism (not shown). remove it and reinsert it into the beam path if necessary. However, such an exchange requires a comparatively long time and therefore cannot be carried out during a continuous movement of the quill 20 .

6. Fourth embodiment

The 13 shows in a the 2 based on a fourth exemplary embodiment of a sensor 24 according to the invention.

Instead of arranging a filter element with a location-dependent degree of transmission or reflection in the pupil plane 42, a plate-shaped transmission filter 44''' is arranged there as a filter element in an intermediate space between the workpiece 26 and the second lens L2. The transmission filter 44''' has a degree of transmission T that depends on the angle of incidence α, as is shown in FIG 14 is shown schematically. In the graph of 14 it can be seen that first measuring light beams 48a, the angle of which with respect to a direction parallel to the optical axis is smaller than a critical angle α _G , are greatly attenuated, while second measuring light beams 48b with larger angles of incidence than the critical angle α _G penetrate the transmission filter 44''' almost without attenuation. can happen.

As in the other exemplary embodiments, an angle-selective weakening of the first measuring light beams 48a and thus an increase in the measuring accuracy for the measurement of the z-coordinate is thus also achieved here.

Claims (11)

Method for 3D coordinate measurement using the autofocus method with the following steps: a) providing a camera (34) containing imaging optics (38) and an image sensor (36), the imaging optics (38) having an optical axis (52) and an object plane (56) has that on the image sensor (36) is imaged; b) illuminating a workpiece (26) to be measured with measuring light; c) generating a relative movement between the workpiece (26) and the object plane (56) along the optical axis (52) of the imaging optics (38); d) recording images of the surface (40) of the workpiece (26) on the image sensor (36) at different points in time during the relative movement; e) storing the images recorded by the image sensor (36) during the relative movement; f) calculating 3D coordinates of the surface (40) of the workpiece (26) by evaluating at least some of the images stored in step e); characterized in that during step d) a filter device (4; 44; 44', 44", 44''') arranged in the beam path of the measuring light attenuates at least part of the first measuring light beams (48a) emitted by the workpiece (26) emanate at first angles with respect to a direction (z) parallel to the optical axis (52) which are smaller than a predetermined limit angle, and that the filter device (4; 44; 44', 44", 44''') covers at least a part of second measuring light beams (48b) is not weakened or at least less weakened than the first measuring light beams (48a), wherein the second measuring light beams (48b) emanate from the workpiece (26) at second angles with respect to the direction (z) parallel to the optical axis (52), which are greater than the specified limit angle. procedure after claim 1 , characterized in that during step d) the filter device (4; 44; 44', 44", 44''') completely screens out all first measuring light beams (48a) and does not attenuate all second measuring light beams (48b). Procedure according to one of Claims 1 or 2 , characterized in that the filter device comprises a filter element (44; 44';44") that acts in transmission or reflection and is arranged in a pupil plane (42) of the imaging optics (38) or at least in the vicinity of the pupil plane (42) and has a location-dependent transmittance or reflectance. Procedure according to one of Claims 1 or 2 , characterized in that the filter device comprises a filter element (44''') acting in transmission or reflection, which is located in an intermediate space between the workpiece (26) and the imaging optics (38) or in an intermediate space between the image sensor (36) and the Imaging optics (38) is arranged and has a degree of transmission or reflectance that depends on the angle. Method according to one of the preceding claims, characterized in that additional images of the surface (40) of the workpiece (26) are recorded at least in some relative positions between the workpiece (26) and the object plane (56), in which the filter device (44; 44') does not attenuate the first measuring light beams (48a) or at least not more than the second measuring light beams (48b). procedure after claim 5 , characterized in that the filter device (44; 44') changes between a first mode, in which it attenuates the first measuring light beams (48a) and the second measuring light beams (48b) to different extents, and a second mode, in which it changes the first measuring light beams (48a) does not weaken or weakens just as little as the second measuring light beams (48b). procedure after claim 6 , characterized in that the filter device has a switchable LCD panel (44) or a switchable micromirror array (44'). Device for 3D coordinate measurement using the autofocus method, having: a) a camera (34) which contains imaging optics (38) and an image sensor (36), the imaging optics (38) having an optical axis (52) and an object plane (56 ) which is imaged on the image sensor (36), b) an illumination device (50) which is set up to illuminate a workpiece (26) to be measured with measuring light, c) a drive (53) or an adjustable optical element for generating a relative movement between the workpiece (26) and the object plane (56) along the optical axis (52) of the imaging optics (38), d) a memory (55) for storing images of the surface (40) of the workpiece (26) , which were recorded by the image sensor (36) at different times of the relative movement, e) a computing unit (54) which is set up to calculate 3D coordinates of the surface of the workpiece (26) by evaluating at least some of the data stored in step e). to calculate images, characterized by a filter device (4; 44; 44', 44'', 44''') which is set up to attenuate at least part of first measuring light beams (48a) emitted by the workpiece ( 26) emanate at first angles with respect to a direction (z) parallel to the optical axis (52) which are smaller than a predetermined critical angle, and - at least some of the second measuring light beams (48b) are not attenuated or at least less than the first measuring light beams (48a), the second measuring light beams (48b) emanating from the workpiece (26) at second angles with respect to the direction (z) parallel to the optical axis (52), which are greater than the predetermined critical angle. device after claim 8 , characterized in that the filter device (4; 44; 44', 44'', 44''') is set up to completely screen out all first measuring light beams (48a) and not to attenuate all second measuring light beams (48b). device after claim 8 or 9 , characterized in that the computing unit (54) is also set up to control the device (10) in such a way that at least in some relative positions between the workpiece (26) and the object plane (56) additional images of the surface (40) of the Workpiece (26) are recorded, in which the filter device (44; 44 ') does not weaken the first measuring light beams (48) or at least not more than the second measuring light beams (48b). device after claim 10 , characterized in that the computing unit (54) controls the filter device (44; 44') in such a way that the filter device (44; 44') switches between a first mode in which it transmits the first measuring light beams (48a) and the second measuring light beams (48b ) weakens to different degrees, and changes to a second mode in which it does not weaken the first measuring light beams (48a) or just as little as the second measuring light beams (48b).

Images