DE102005007244A1 - Absolute calibration free three dimensional mirror surface measurement sensor has lamp unit giving high depth of field laterally structured light pattern from microlens array - Google Patents

Abstract

An absolute calibration free three dimensional mirror surface measurement sensor has lamp unit giving a large depth of focus laterally structured light pattern from a point source, microlens array and macro lens and an observation unit with array detector and imaging optics. Independent claims are included for the illumination and observation units.

Description

described In the following, a method for simultaneous, calibration-free, non-contact, absolute and imaging optical determination of the 3D coordinates of related reflective surfaces, wherein a deep-sharp light pattern is used for optical scanning becomes.

The surfaces can either fully mirrored, such as vapor-deposited or polished surfaces, e.g. Mirrors, and sheet metal parts or even partially mirrored or partially be reflective, such as polished or sufficiently smooth surfaces more transparent Media such as glass surfaces, Liquid surfaces or transitions of Media with different optical refractive index. The surfaces must be for Surveying should continue to be sufficiently smooth, i. the roughness or graininess the surface must be like that be that the Surface at Scanning with a diameter d light beam within each surface part with diameter d approximately free of curvature is.

With the described method can all necessary steps of data acquisition are carried out simultaneously, therefore you can called surfaces also be observed and evaluated in dynamic processes. This is important, e.g. in the vibration or deformation analysis of surfaces.

The measured object, the one surface defined above has, hereinafter referred to as reflective measurement object. The described Apparatus, hereinafter referred to as "sensor" exists at least one imaging lens and an optical surface detector (e.g., an electronic camera with CCD or CMOS element, but also e.g. chemical film for a downstream evaluation) and at least one lighting device, which creates a three-dimensional light pattern.

was standing of the technique

• 1. mechanische (taktile) Abtastung,
• 2. projizierende Verfahren, jedoch ohne ausreichend tiefenscharfes Lichtmuster, z.B. Patent Nr. DE 19731545 C1
• 3. Wellenfront-Meßverfahren (z.B. Shack-Hartmann),
• 4. konfokale Oberflächenabtastung; z.B. Patent Nr. DE 19749974 C2
• 5. Interferometrische Verfahren,
• 6. Ray-tracing mit feinem, abtastendem Laserstrahl, z.B. in Appl.Opt. Vol. 27, p. 5160ff(1988) beschrieben

The measuring tasks described above are achieved according to the prior art with the following methods:

• 1. mechanical (tactile) scanning,
• 2. projecting methods, but without sufficient deep-sharp light pattern, eg patent no. DE 19731545 C1
• 3. wavefront measuring method (eg Shack-Hartmann),
• 4. confocal surface scanning; eg patent no. DE 19749974 C2
• 5. interferometric methods,
• 6. Ray-tracing with fine, scanning laser beam, eg in Appl.Opt. Vol. 27, p. 5160ff (1988)

• 1. verschleißanfällig und langwierig; bewegliche Teile; Gefahr der Beschädigung des Prüflings;
• 2. ungenau, da hier die Lichtausbreitung zwischen Meßmuster und Prüflingsoberfläche unberücksichtigt bleibt;
• 3. nur für kleine Objekte (beschränkt von der Größe von z.B. Abbildungsoptiken und/oder Mikrolinsenarrays); weiterhin Einschränkung hinsichtlich meßbarer Höhendifferenz und meßbarem Steigungswert der Oberfläche;
• 4. zeitaufwendig, da z.B. eine sequentielle Abtastung in z-Richtung erforderlich ist; bewegliche Teile;
• 5. hoher apparativer Aufwand; insbesondere für ausgedehnte Flächen. Phaseshift-Verfahren benötigen mindestens 3 Aufnahmen mit jeweils verschobener Referenzwelle. Wie für 4 ist das Verfahren nur für geringe maximale Höhendifferenzen einsetzbar; weiterhin sehr anfällig gegenüber Umgebungseinflüssen, wie z.B. Fremdlicht, Vibrationen);
• 6. Lediglich sequentielle Abtastung möglich (daher langsam); weiterhin sind laterale Position und Strahlwinkel aufgrund der Scan-Bewegung nicht hinreichend genau bekannt; beinhaltet bewegliche Teile; Für die Verfahren aus 2, 3 und 6 ist in der Regel außerdem eine Kalibrierung notwendig.

Disadvantages of previous methods (numbering refers to prior art):

• 1. wear-prone and tedious; movable parts; Risk of damage to the test object;
• 2. inaccurate, since here the light propagation between the measuring pattern and the specimen surface is disregarded;
• 3. only for small objects (limited to the size of eg imaging optics and / or microlens arrays); furthermore limitation with regard to measurable height difference and measurable slope value of the surface;
• 4. time-consuming, since, for example, a sequential scan in the z-direction is required; movable parts;
• 5. high equipment costs; especially for large areas. Phaseshift methods require at least 3 shots, each with a shifted reference wave. As for Figure 4, the method is applicable only to small maximum height differences; still very vulnerable to environmental influences, such as extraneous light, vibrations);
• 6. Only sequential sampling possible (hence slow); Furthermore, lateral position and beam angle due to the scan movement are not known with sufficient accuracy; includes moving parts; For the methods of FIGS. 2, 3 and 6, a calibration is also usually necessary.

Objectives of the invention

• • non-contact and simultaneous detection of the surface topology of fully or partially mirrored surfaces, in particular also aspheric surfaces high measurement resolution without a priori knowledge of the surface;
• • easier, calibration-free measuring setup (Elimination of the determination of the outer orientation the lighting, object and sensor system);
• • omission movable components.

Advantages of the invention

With the sensor according to the invention can be a reflective surface, as defined above, measured absolutely and with high accuracy become. A priori knowledge about the geometry of the surface is not required.

By using the lighting device according to the invention of the sensor is a virtually unlimited depth of field of dreidi dimensional light pattern available. Thus, the measuring body can be placed in the measuring field without further presetting. Equally uncritical is the location of the recording unit (outer orientation).

at Knowledge of the spatial Course of the three-dimensional light pattern, such as by pre-calibration when mounting the projector unit (inner orientation) is no Calibration e.g. the orientation of the coordinate systems of lighting, object and receiving unit (outer orientation) more required.

The Measurement needed only at least two, usually three measuring operations at different spatial positions for the absolute and accurate scanning of the 3D topology of the specimen.

By Ausspieglung the measuring light in different positions these measurements are carried out simultaneously and without moving parts.

The specular object is illuminated with a special three-dimensional pattern of light whose geometry and structure are well known in all three space dimensions. The reflected portion of the light is picked up by a photosensitive optical surface detector. The detector records both the lateral position of the reflected light beams and the beam direction of the light pattern after reflection on the specimen surface. From both information, the topographical data S (x _i , y _i , z _i0 ) of the location of the reflection on the specimen _surface (see _Figure 5) and its local inclination can be determined. With knowledge of the sensor geometry (inner orientation), the surface topology can be determined absolutely, with high accuracy.

lighting

The Three-dimensional light patterns can e.g. by using a sufficient punctiform light source (not necessarily coherent) and at least one array of microlenses. Corresponding Drawing 1a, 2a, 3a, the microlenses cause a collimation of from the light source outgoing light beam, which in the opening of the respective microlens falls. In this way, depending on the choice of lens parameters, a pattern with high depth of field generated from collimated partial beams ("light needles") Divergence of these light needles is due to the arrangement of the microlenses and the light source completely set (Figure 1b and c, 2b and c, 3b and c) and thus exactly known. In particular, a parallel spatial course of light needles be generated (drawing 2a, b, c and 3a, b, c).

evaluation

Of the to be measured specimen is illuminated with the light needles described above. The lateral Measuring resolution corresponds doing the grid of the light pattern. The reflected light pattern is visualized on a diffusely diffusing screen and with a lens imaged on an optical surface detector. The Evaluation of the image data is done by a computer (drawing 4).

On the screen, the reflected light needles produce points of light. For example, by determining the centroid of the light spots, the intersection S (x _i , y _i , z) of each light needle can be determined with the z plane of the screen. The beam directions of the respective light needles can be determined by a second measurement, wherein the sensor unit or observation unit (drawing 4) is shifted by a known value in z (Figures 6 and 7). By sufficient displacement, a very high angular resolution can be achieved. In order to avoid moving components, this path difference in z can also be achieved by reflecting out of the beam path in light paths of different lengths to the observation unit (drawing 7).

Detailed description a typical measurement setup and the implementation a measurement:

in the The following is an embodiment of the invention with variants described.

1st projection unit:

The Projection unit consists of at least one almost punctiform light source. The punctiformity is then given sufficiently, if the spatial extent of the Light source is smaller than the diffraction spot of the projection used optical system. The light source does not have to be mandatory temporally coherent be. It can e.g. as a laser diode, an LED or too a sufficiently spatial coherent executed thermal light source be such as a halogen lamp.

The light from the light source is partially collimated using microlenses (microlens array) arranged laterally in a plane with the aid of at least one anay. Each individual lens causes the formation of a parallel sub-beam (drawing 1a), which propagates in the direction of the main beam of the respective microlens in space, which is called in the following light needle. According to the laws of optical diffraction, the beam diameter of the individual light needles changes only slightly within a large z-range. In this way, a light pattern can be generated, the over meh centimeter up to a few meters unchanged (depth sharp) is.

By using at least one macrolens and a microlens array (FIG. 2a), or at least one macrolens and two microlens arrays (FIG. 3a), a deep-focused light pattern can be generated which consists of light needles running parallel in space. For such a light pattern running in each distance z _p to the projection unit both the lateral position x, y and the propagation angle of each individual light needle is known (drawing 1b and c, 2b and c, 3b and c).

The Variants of the structure described in the drawings 1a, 2a, and 3a do not include any moving parts. The user must therefore no adjustments, e.g. Adjusting the sharpness or Calibration on the projection unit.

Farther can be a DUT be placed at any z-distance to the projection unit, since all geometric properties of the light pattern are known. This provides a further advantage in the handling of the measuring device described here It only has to to be sure that the whole length of the unfolded beam path to the observation unit within of the depth of field range the light needles lies.

For the rest Course of the description we go without limitation of generality from the generation of n × m parallel to the optical axis running, Cartesian in n columns and m lines arranged light needles. The light needles can be from each other except can not be distinguished from each other by their lateral position.

The Sizes for n and m are in practice in the order of magnitude from a hundred to a few thousand. This number is fundamental not limited. For A sufficient detectability of the light needles may be a certain relationship n between the surface the scanning field of a single light needle and illuminated by her area not exceeded become. The upper limit is η <1.

2. Data Acquisition and Detection Unit:

The reflective measuring object is under consideration the depth of field of the light needles placed in the light pattern described in 1..

The reflected light pattern drops to one, at a distance z to the measuring surface, diffusely diffusing screen. The distribution of the luminous points is by means of a lens is imaged on a photosensitive area detector and evaluated with the help of a computer (drawing 4).

For the following described measurement of the surface data by evaluating the points of light it makes sense, though not imperative and therefore not limiting, that the reflective surface so oriented that will be the main part of the reflected light pattern reflected in the direction of the optical axis becomes. Furthermore, the observation unit, without violation of generality positioned on the optical axis of the overall system (drawing 7). In this case, the reflected light pattern with the help of a beam splitter reflected from the beam path (drawing 7).

The position of the points of intersection of the light needles S (x _i , y _i , z) with the screen are determined by the respective mean local inclination of the reflection location with the size of the illuminating surface of the light needle on the specular surface and the distance of the screen from the reflection S (x _i , y _i , z _i0 ) of the respective light needle.

by virtue of the different local inclination of the measuring surface, it is typical that the arrangement the points of light on the screen no longer the original ones Arrangement of n × m corresponds to Cartesian points of light. Rather, it typically has a blending of the locations of the points of light yield the points In general, they will no longer be in Cartesian Arrangement available. Furthermore, it will be typical that by areas of the test object, have very strong inclination, individual light needles the screen will not meet again. Therefore, the number of, on the screen detected light points i ≤ (n · m).

The clear determination of the surface topology the sample surface sets then from the steps listed under a) to d) together, where e.g. through reflections at different z-positions all operations the data recording can be performed simultaneously.

The Steps of evaluation can also simultaneously, ie directly downstream of the data acquisition respectively. It is also possible (for example, to increase the measuring speed), without restriction the simultaneity of all physical measuring steps to buffer the data and only in a subsequent process step in a computer to evaluate:

Individual measuring steps:

a. Determination of the beam directions of the light needles after reflection on the specimen

To determine the beam direction of each re The lateral positions of the light spots on the screen are measured in at least two different z-distances (Figure 6). The positions S (x _i , y _i , z ₁ ) of each luminance point in the screen plane at z ₁ = z and the positions S (x _i , y _i , z ₂ ) at z ₂ = z + Δz are eg by determining the center of gravity of the intensity distribution determined. Conveniently, Δz will be chosen to be small so that the points of the two measurements identified i and j can be clearly assigned to the same light needle and no further mixing due to different beam directions can result.

By the relationship θ _i = atan (x _i2 -x _i1 ) / Δz and φ; = atan (y _i2 -y _i1 ) / Δz, the beam directions of all the i needles visible on the screen can be determined (Figures 5 and 6).

b. Determination of local Inclination of the reflective specimen

According to the laws of reflection, from the angles of the light needles θ _i , and φ _i, the associated local normal vector n _i , and thus the average slope of the illuminated area element n _{x; i,} and n _{y; i,} are determined very accurately. In order to achieve high accuracy, at least one measurement of the positions S (x _i , y _i , z ₃ ) of the points of light center is required, in which the greatest possible value for Δz 'is chosen for the displacement (Figure 6). According to the approximate beam directions measured under a.), The light spots found at z ₃ = z + Δz 'can be assigned to those from the measurement plane at z ₁ or z ₂ . Corresponding to the relationships θ _i = arctan Δx '/ Δz' and φ _i = arctan Δx '/ Δz', the beam directions can be determined therefrom with improved accuracy, where Δx '= x ₃ -x ₁ and Δz' = z ₃ -z ₁ or Δx '= x ₃ -x ₂ and Δz' = z ₃ -z ₂ , depending on whether the measured values from plane z ₁ or z _{2 are} referred to.

c. Assignment of the light needles for orientation of the illumination pattern

By knowing the points of intersection as well as the beam direction of the reflected light needles, the light needles can be traced back to the positions S (x _i , y _i , z _i0 ) _{n, m} in the computer, ie the spatial position immediately at the reflection location on the reflecting surface of the test object. This position can be found when in turn results in the original arrangement of the light needles, corresponding to the unreflected, known pattern.

These Finding happens iteratively in a computer program, by possible intersections of the Light needle with the undisturbed Be determined course and doing the maximum possible tilt between adjacent surface elements be taken into account on the reflective specimen surface (Drawing 8).

d. Integration of the local inclinations

After the spatial coordinates of the reflecting surface S (x _i , y _i , z _i0 ) _{n, m have been} determined, the surface can be obtained by integrating the inclination angles n _{x; i,} and n _{y; i,} which are determined very precisely by measurement in b) high accuracy can be calculated.

pictures:

• 1.a generation of light needles, wherein the beam directions the light needles are not parallel, consisting of a point light source and a microlens array
• 1.b intensity distribution in a plane at the position z = z ₁
• 1.c intensity distribution in a plane at the position z = z ₂ , where for the grid dimensions: px ₁ <px ₂ and py _l <py ₂ ; d ₁ = d ₂ (with d: diameter of the light needles in different z position)
• 2.a Generation of parallel light needles by using a punctate Light source, a microlens array and a macrolens
• 2.b intensity distribution in a plane at the position z = z ₁
• 2.c intensity distribution in a plane at the position z = z ₂ ; px ₁ = px ₂ and py ₁ = py ₂ ; d ₁ = d ₂
• 3.a generation of parallel light needles by using a punctate Light source, two microlens arrays and a macro lens
• 3.b intensity distribution in a plane at the position z = z ₁
• 3.c intensity distribution in a plane at the position z = z ₂ ; px ₁ = px ₂ and py ₁ = py ₂ ; d ₁ = d ₂
• 4. basic structure of the observation unit
• 5. Positioning of the DUT in the light pattern, orientation of the beam angle
• 6. Principle of angle determination
• 7. Positioning of the observation unit in the beam path to Measurement of the beam angles at different positions by steel division of the reflected light pattern.
• 8. Virtual repatriation of the Lightpins on the surface of the reflective specimen by consideration potential Tilts and intersections with the original light pattern.

Literature:

• Breuckmann B .: "Image processing and optical metrology in industrial practice "Franzis-Verlag ISBN 3-7723-4861-0
• Daniel Malacara, "Optical Shop Testing", ^2nd Edition, John Wiley & Son, Inc. (1992)

1

punctiform or almost punctiform light source

2

Microlens array with focal length f

3

teilkollimiertes light beam with divergent course

4

arranged in parallel light needles

5

macro lens

6

optical axis

7

Microlens array with focal length f

8th

reflective measurement object

9

observation unit

10

imaging optics

11

Umbrella, diffusely scattering

12

light sensitive area detector

13

reflected light Adel

14

semitransparent mirrors, beamsplitter

15

intersections between reflected and undisturbed light needle

16

reconstructed reflection places

Claims (25)

Sensor for the absolute, calibration-free 3D surface measurement of specular surfaces such that he from at least one lighting unit and at least one observation unit consists. Lighting unit of claim 1 such that these creates a deep, laterally structured light pattern. Lighting unit of claim 1 such that these produces a deep-focused, laterally structured light pattern, the three - dimensional structure of the pattern over a wide range of its Propagation in the room unchanged remains. Lighting unit of claims 1, 2 and 3 such that the light pattern from claim 2 and 3 by using at least one punctiform or almost punctiform Light source and at least one microlens array consists. Lighting unit of claims 1, 2 and 3 such that the light pattern from claim 2 and 3 by using at least one punctiform or almost punctiform Light source and at least one microlens array and at least a macrolens. Lighting unit of claims 1, 2 and 3 such that the light pattern from claim 2 and 3 by using at least one punctiform or almost punctiform Light source and at least two successively arranged microlens arrays and at least one macrolens. Lighting unit of claim 1 to 6 such, that the Light sources either as laser diode, laser, LED or as thermal, polychromatic light source is executed. Observation unit of claim 1 such that these from at least one diffusely scattering screen, at least one photosensitive area detector and at least one imaging optics exists. Observation unit of claim 8 such that these Unit in the, on the DUT reflected rays, causing the reflected Light pattern of claim 2 to 7 on the screen of the observation unit intensity distributions (Light points), which claim by the receiving optics 8 on the photosensitive surface detector be imaged from claim 8. Observation unit of claim 8 and 9 such that the positions of intersections of the light rays S (x _i , y _i , z) are determined with the screen. Observation unit of claim 8 to 10 such that this unit at at least two different z-positions, the positions intersections of the light beams S (x _i , y _i , z) receives. Observation unit of claim 8 to 11 such that measurements according to claim 10 and 11 are divided by beam splitting to Δz different lengths of light paths, where by at least one receiving unit, the light points are recorded and the positions intersections of the light beams S (x _i , y _i , z) are determined. Observation unit of claim 8 to 12 such, that by the at least two measurements from claims 1 l and 12 the angles the individual beams are determined. Observation unit according to claims 8 to 13, such that an observation unit at at least one second z-position receives the positions of the intersections of the light beams S (x _i , y _i , z) according to claims 9 and 10, wherein the distance in z between the z-positions Positions of the individual measurements according to claims 11 to 12 are chosen to be large according to the desired angular resolution. Observation unit of claims 8 to 14 such that for the unambiguous assignment of the intersection points to the corresponding beams and intersections from the measurements of claim 14, the respective beam angles and positions determined from the measurements according to claims 10 to 13 intersect points of the light beams S (x _i , y _i , z) are used. Observation unit of claim 8 to 15 such, that out the measured according to claim 10 and 15 positions of the light spots the beam angles are determined. Observation unit of claim 8 to 16 such, that the Position for all measurements from the claims 11 and 14 by at least one beam splitter of the reflected Light pattern on each of different lengths of light paths to each at least an observation unit is achieved. An observation unit as claimed in any of claims 8 to 17, wherein the measurements of the positions of intersections of the light beams S (x _i , y _i , z) of claims 10 to 14 are made at measurement locations differing only in their position along the optical axis. The observation unit of claims 8 to 18, such that the positions of the intersections of the light rays S (x _i , y _i , z) are traced back to the immediate vicinity of the surface of the specimen reflecting surface. Observation unit of claim 8 to 19 such that the tracing back of claim 19 by appropriate software in a computer program using the determined from claim 10 to 16 beam directions and positions intersections of the light beams S (x _i , y _i , z) is performed. Observation unit according to claims 8 to 20, such that the beams measured with the aid of the beam directions and positions intersections of the light beams S (x _i , y _i , z) of claims 10 to 16 are the original sort of the known light pattern of claim 1 to 7 are assigned. Observation unit of claim 8 to 21 such, that the Assignment from claim 21 considering a maximum allowed surface slope between two laterally adjacent, intersected points of intersection the light needle with the original one Considered light pattern become. Observation unit of claim 8 to 22 such, that out the, according to claim 13 and 15 measured beam directions the local surface slopes be converted. The observation unit of claims 8 to 23, such that the local surface _slopes of claim 23 are assigned to the local positions of the reflective surface S (x _i , y _i , z _i0 ) by the association of claims 21 and 22. Observation unit of claim 8 to 24 such, that out the, according to claim 24 associated surface slopes of the surface profile itself is calculated.