DE102017009334B3 - Method for testing an optical system - Google Patents

Abstract

The invention relates to a method for testing an optical system, in which a test pattern, in particular after imaging with imaging optics, is imaged onto an electronic light-sensitive detector via a beam path containing the optical system to be tested (specimen) and that of this detector generated signals are supplied to a data processing system for determining a modulation transformation function (MTF), wherein the test pattern and / or the imaging optics and / or the detector is optionally moved in the direction of the beam path, so that the detector thereby detects a two- or three-dimensional array of intensity values, wherein the data processing system then calculates a three- or two-dimensional data field of MTF values from the detected intensity values by means of a Fourier transformation, wherein the data processing system calculates the three- or two-dimensional data field of MTF values. using a two-dimensional array of MTF values and a three-dimensional model field of MTF values, and using a best-fit method determines the parameters of a coordinate transformation, which results in a computationally optimal correlation of the data field with the model field, and from the parameters determined Position of the optical axis of the specimen is determined.

Description

The invention relates to a method for testing an optical system in which a test pattern is imaged onto an electronic light-sensitive detector after imaging with imaging optics, via a beam path containing the optical system to be tested (which is also referred to below briefly as the test object) the signals generated by this detector are fed to a data processing system for determining a modulation transformation function MTF.

From the German patent DE 198 23 844 C1 For example, a lens tester is known which is provided for testing the imaging properties of lenses by measuring the modulation transformation function. In this case, a CCD camera with a downstream electronic image evaluation device is proposed in particular as the detector. To improve the measurement accuracy, a construction is described with a holder on which the lens to be tested is mounted pivotably.

Also the German publication DE 38 42 144 A1 describes a method for testing an optical system. For image acquisition, a two-dimensional CCD area array is used here. For the testing of high-quality lenses, the use of a three-dimensional modulation transformation function is proposed in which a contrast evaluation takes place in several directions. For this purpose, the use of a radial lattice and a measurement at different angles are proposed to determine the directions of decentration and astigmatism. The test apparatus proposed in this document requires a precise alignment of the optical axis of the specimen to be controlled in an unspecified manner by an image processing system and visually inspected by an eyepiece.

As the major cause of failure in measuring the modulation transformation function (MTF) of optical systems, such as camera lenses, has been the poor repeatability in the positioning of the testing optical system with respect to its inclination and its azimuth angle has been found. Even the slightest deviation leads to relatively large errors in the MTF results. The usually achievable repeat accuracies lead to result variations of +/- 3.5% and more.

One common technique for determining MTF values uses a crosshair as a test pattern. The optical system under test (test object) is positioned passively before the measurement. Subsequently, the test pattern is imaged through the device under test, optionally using an intermediate optical system, onto a camera chip and the captured camera image is evaluated by means of a computer-aided data processing system.

The points of a test pattern, when imaged by a real optical system, are not ideally dot-shaped, but are subject to broadening determined by the imaging characteristics of the particular optical system. These broadening can be spatially described by a so-called point spread function (also known as point spread function). The course of the point spreading function can be represented after Fourier by a sum of periodic functions whose frequency spectrum is called the modulation transfer function and whose frequency components are called MTF values.

The acquisition of a test pattern by means of a two-dimensional camera chip is generally evaluated by determining the modulation transfer function as the Fourier transform of the acquired imaging pattern. The determined MTF values represent a measure of the quality of the optical image and thus also of the quality of the optical imaging properties of the optical system to be tested.

The problem is that the results of such measurements or calculations depend very sensitively on the inclination of the specimen in the beam path. An ideal measurement requires that the optical axis of the specimen is aligned exactly in the direction of the beam path. However, it has been found that this alone can not be achieved with mechanical positioning means with the desired accuracy.

A particular disadvantage here is that the resulting error also has a large spread, because it is very difficult to ensure a sufficiently well reproducible tilt and azimuth positioning.

In addition, it is problematic that the error can be reproduced in different ways depending on the test object. The reason for this is that the error sensitivity with respect to the axis inclination and the azimuth angle is different for lower-quality parts than for better quality parts.

Another systematic error often arises from the fact that the alignment of the specimen in general can not be done directly in relation to the optical axis, since their position is not directly visible externally. Instead, it will open the outer shape of the specimen reference, for example, based on its housing surfaces. As a result, manufacturing tolerances of the housing go into the results as a further source of error.

It was the object to provide a method for testing optical systems, which allows the measurement of MFT values with a particularly high accuracy and which positioning error of the DUT as well as possible eliminated.

This object is achieved by the characterizing features of claim 1 or alternatively by the characterizing features of claim 2.

For an optical system (test object) to be tested, a three-dimensional model field of MTF values is first determined. From an assumed ideal geometry of the optical components and their positions to each other, a theoretical MTF model field is calculated, which is related to the optical axis and the position of the specimen. Alternatively, the MTF model field may be constructed from statistical measurement data of MTF measurements on many similar DUTs or formed from the combination of theoretical calculations and real measurements.

According to the invention, the real test piece becomes a part of the measuring system itself. By metrological recording and subsequent evaluation of a two- or three-dimensional data field of intensity values, a two-dimensional or three-dimensional data field of MTF values is calculated by Fourier transformation. For this data field, an optimal adaptation to the MTF model field is mathematically searched by a best-fit method, wherein the adaptation by a coordinate transformation, ie a computational spatial displacement and rotation of the measured three-dimensional MTF data field against the MTF assumed to be ideal. Model field is found.

As a result, the real position and the orientation of the specimen is computationally determined relative to an ideal orientation in the beam path, whereby it is possible to virtually completely eliminate orientation errors of the specimen for the further determination of specific MTF values.

In the simplified method variant of claim 2, only a two-dimensional measurement data field is detected, which can be regarded as a sectional plane and thus as a subset of a three-dimensional measurement data field. As a result, if a slightly lower accuracy is accepted, a scanning process can be dispensed with and the measurement setup can be simplified. It is particularly advantageous that the thus simplified test method can be carried out particularly quickly.

The optical system to be tested may be a single optical lens or may comprise a lens system comprising a plurality of optical lenses, such as a lens and in particular a camera lens.

The optical system to be tested can in particular also be a camera system, which consists of at least one pre-assembled arrangement of a camera lens and an optical detector. In this case, advantageously, the optical detector of the camera system can also be used as a detector of the test method. As a result, the test method takes into account not only the optical properties of the camera lens but the properties of the entire camera system, including the detector and any other intermediate elements of the optical path, as well as their arrangement and orientation.

The arithmetical evaluation takes place in all cases using a compensation calculation, which optimally adapts the two- or three-dimensional MTF data field determined from recorded intensity values to a previously determined three-dimensional MTF model field. This step is also referred to as a "best-fit" adjustment.

From the computationally determined best-fit parameters advantageously further information can be determined, in particular the actual orientation of the specimen relative to the measuring arrangement, and thus positioning error of the specimen, positioning and orientation error of the detector (camera chip) behind the Prüflingsoptik, and the exact location the optical axis of the specimen.

To capture a measured data field of the test object, a scanning method is preferably used in which either the distance between the detector and the test object is changed, or a defocusing / displacement of the imaging optics used of the test pattern or even a displacement of the test pattern itself is performed. It can be achieved by defocusing / shifting the imaging optics of the test pattern, a significantly higher scanning accuracy than with a shift of the detector behind the DUT.

It is advantageous to interpolate from the MTF values, which were obtained from measurement data of pixels captured by the detector, additional MTF values to unrecorded pixels. For such an interpolation can a mathematical model that takes into account the theoretical spatial dependence of the MTF values from the three-dimensional MTF model field of the respective prototype prototype.

A scan for acquiring a measurement data field can be effected, in particular, by a change in the distance between the test pattern and the test object or by a change in the distance between the test pattern and the additional imaging optics. Both are particularly advantageous for testing camera systems with a fixed-focus lens, in which a movement of the detector is not possible. Since a shift of the detector has to be dispensed with in this case, only the subset of the three-dimensional data field of MTF values resulting from the given fixed detector position can be used here for the best-fit evaluation.

In a measurement data acquisition by a scanning method can be dispensed with the use of imaging optics in many cases even, since their function can be achieved in principle with a moving test pattern. Defocusing is achieved either by moving the test pattern out of the focus range of the test piece or by changing the distance of the detector behind the optics formed by the test piece or by using both options.

Advantageously, it can be provided that during scanning along the (here arbitrarily chosen) z-direction of a coordinate system whose position should correspond as well as possible to the optical axis of the specimen, the test pattern is additionally moved within the perpendicular xy plane. This allows MTF values for a higher number of x - y -, and z Positions are obtained by a direct measurement instead of by interpolation.

Advantageously, it can also be provided that when scanning a test pattern or its image, for each set test pattern position or imaging position in the z Direction, a complete three-dimensional intensity values data field behind the DUT is detected by movement of the detector within the xy plane.

It can also be advantageously provided that only the detector is moved while the test pattern and its image are held at a fixed position. As a result, intensity values or MTF values are not determined at all points behind the test object, but only that subset belonging to this fixed position of the test pattern is detected.

Due to the well-known phenomenon of field curvature of an optical image, the points sharply imaged by the specimen optics are generally not within a flat but within a curved plane. A scan that only moves the detector completely captures the points within that curved plane. However, only a subset of the complete three-dimensional model data field of the MTF values can be adapted to these. By using the best-fit method, however, the position of the test object can be determined almost as well as when compared with a fully acquired three-dimensional MTF data field.

Equally advantageously, it can alternatively be provided that the detector remains at a fixed position, and relative thereto only the test pattern and / or the imaging optics is moved. As a result, the MTF values behind the test object can not be determined at all points, but only the subset which can be determined with the aid of the intensity detection at this one position of the detector. However, the shift of the imaging optics replaces the movement of the detector and makes it possible, despite the field curvature, to capture an approximately equal number of pixels. From this a similarly large subset of the complete three-dimensional data field of the MTF values can be determined as with a movement of the detector. Again, by applying the best fit method, the position of the device under test can be determined almost as well as when compared with a complete three dimensional data field of the MTF values.

Likewise, alternatively, it can advantageously be provided that both the detector, the test pattern and the imaging optics remain at a fixed position. From this, only a two-dimensional subset of the complete three-dimensional data field of the MTF values can be determined. Again, by applying the best fit method, the position of the device under test can be determined almost as well as when compared with a complete three dimensional data field of the MTF values.

The principle of the method according to the invention and advantageous variants will be described below with reference to the schematic representations of 1 to 4 be explained further.

The 4 schematically represents a measurement setup for the proposed method. Sketched are the required components, as well as their Relative arrangement to each other, being dispensed with a scale true representation.

Provided is a light source 16 which is a target 11 illuminated or illuminated. On the target 11 is a test pattern 5 applied, which may be performed, for example, as a crosshair or as a linear or surface applied dot pattern. With the help of an imaging optics 6 is from the test pattern 5 generates an intermediate image which is from the optical system to be tested 1 when on a two-dimensionally registering detector 3 is shown. The detector 3 is preferably designed as a camera chip designed as a CCD camera chip and with a data processing system 17 coupled.

The data processing system 17 is through a calculator 14 formed, which executes an image processing software. The computer 14 may for example be a standard personal computer. In addition, at least one is through the computer 14 controlled drive device 12 . 13 . 18 present the target 11 and / or the detector 3 and / or the imaging optics 6 in the direction of the optical beam path 15 which is arbitrarily set as z-direction here. Optionally, the drive devices 12 . 13 . 18 the target 11 , the detector 3 and / or the imaging optics 6 also within the xy plane perpendicular to the direction of the beam path 15 move.

The optical system to be tested 1 ("DUT") may for example consist of an optical lens or a lens system, such as a camera lens. Alternatively, the examinee 1 also be a preassembled camera unit, which consists of at least a lens system and an image sensor. In this case, the image sensor of the camera unit can advantageously also be the detector 3 form the test system, and for this purpose with the data processing system 17 get connected.

An idealized measuring procedure is achieved by the 2 explained. Schematically represented are the target 11 with the test pattern 5 , the imaging optics symbolized by a single lens 6 , as well as the specimen symbolically represented by another single lens 1 , In addition, a coordinate friction is drawn, which shows the three mutually perpendicular directions in the measurement setup x . y . z referred to hereinafter.

The examinee 1 make that on the target 11 arranged and by the light source 16 illuminated test pattern 5 on the detector 3 from. Between the target 11 and the examinee 1 is the imaging optics 6 inserted. The imaging optics 6 prepares the beam in front of the test object 1 so that the respective test requirements are met. So can the imaging optics 6 For example, an intermediate image of the test pattern, not shown here 5 produce, for example, at infinity in front of the examinee 1 can lie.

To by means of the two-dimensionally registering optical detector 3 a three-dimensional array of light intensity values I ( x . y . z ), several images are taken, in which, after each shot, the relative positioning of the optical components Target 11 , Imaging optics 6 and detector 3 is changed relative to each other. This happens because at least one of these components 11 . 6 . 3 by a drive device in or against the direction of the beam path 15 is moved.

The detection takes place in the form of a scan in which, for example, the detector 3 repeated by a predetermined small distance in or against the z Direction is moved and makes an image capture of the xy plane after each shift.

Alternatively or additionally, the target can also be used 11 in the z-direction against the test object 1 be moved, which here by a triple representation of the target 11 symbolically indicated. As with a shift of the target 11 also the of the target 11 against the detector 3 shifts, so detects the two-dimensionally measuring detector 3 again a three-dimensional intensity data field I ( x . y . z ).

By successive shifts of both the target 11 as well as the detector 3 In addition, even for each one z Position of the target 11 a three-dimensional data field of a light intensity distribution I (x, y, z) are recorded, resulting in a large evaluable amount of data.

In any case, the three-dimensional array of intensity values I ( x . y . z ) (or a corresponding two-dimensional data field when the scanning method is omitted) to that in the 4 illustrated data processing system 17 transmitted and evaluated by this.

In the evaluation, a Fourier transformation of the intensity values I ( x . y . z ) in each case in a surrounding area around the intensity values I ( x . y . z ) assigned spatial points ( x . y . z ) performed around. The Fourier transformation is subsequently normalized by means of a reference Fourier transformation in order to obtain the values of the modulation transformation function MTF (FIG. x . y . z ) to these points in space ( x . y . z ) behind the examinee 1 to surrender. The reference Fourier transformation is the Fourier transformation of the test pattern 5 arising from the shape of the test pattern used 5 can be calculated.

Idealizing was until now assumed that the position of the optical axis 10 of the optical system to be tested 1 exactly in the direction of the beam path 15 and in the scanning direction of the detector 3 , That is aligned exactly parallel to the illustrated z-direction of the measuring arrangement, and that the optical system to be tested 1 at the distance of their own focal length f from the detector 3 located. These assumptions are only approximately true for real measurements, and it has been found that precisely in this way the achievable accuracy of measurements of the modulation transformation function MTF (FIG. x . y . z ) is significantly limited.

This will be described with reference to the schematic representation of 1 be explained. Outlined in turn is a measurement setup as in the 2 , referring to a representation of the detector 3 was omitted here.

The coordinate system S the measuring arrangement again has the directions x . y . z on. An ideal with the optical axis 10 parallel to the z-direction aligned specimen 1 is shown schematically as a lens contour with a dashed edge. As a picture of the target 5 through this ideally aligned test piece 1 the detector would use the dot pattern shown with bright dots with the intensity values I ( x . y . z ) to capture.

In a real measurement is the optical axis of a DUT 1' but never exactly parallel to z Direction of the measuring arrangement aligned. The here as Linsenkontur represented by a solid line real sample 1' can have its own coordinate system S ' be assigned, whose axes x ' y ' . z ' all (exaggerated here) not parallel to the axes x . y . z of the coordinate system S the measuring arrangement can be accepted. In particular, the as z ' -Axis represented direction of the optical axis 10 ' of the test piece 1' In general, tilted against both the x and the y direction of the coordinate system S of the measuring arrangement by a respective unknown angle and shifted by an unknown amount.

By deviating from the ideal position results in the optical detection of the real DUT 1' in the coordinate system S of the measuring system a clearly deviating pattern of intensity values I '( x . y . z ), which is represented here symbolically with dark dots, their positions and image sharpnesses clearly from those of the bright dot pattern of an ideally aligned test object 1 differ.

Although in this 1 The detector, not shown, naturally detects intensity values at each point of its sensor surface, but is used to determine the imaging quality of a real test object 1' less the absolute intensities I '( x . y . z ) of mapped points are interesting, as their areal or spatial expansions. The diameters of two- or three-dimensionally detected imaging points represent a measure of the image sharpness, and are represented by the values MTF '(FIG. x . y . z ) of the modulation transfer function MTF.

In the presentation of the 1 represent larger points values MTF ( x . y . z ) or MTF '( x . y . z ) with greater blur and smaller points MTF values with a higher image sharpness. In the example shown here, the MTF values with the least fuzziness are directly on the optical axis 10 . 10 ' both the ideal and the real sample 1 respectively. 1' ,

By tilting and / or shifting the real test object 1' opposite to its ideal position in the measuring system S The position and the sharpness of the dots of the dark-colored dot pattern also deviate significantly from the corresponding values of the bright dot pattern of an ideally oriented sample 1 from.

It is essential that a tilting and / or displacement of the real test object 1' compared to its ideal position in the measuring system S, the image of the test pattern 5 on the detector is not twisted or displaced to the same extent. This means that the intensity distribution I ( x . y . z ) during a displacement or tilting of a test object 1 his ideal position does not follow this shift or tilt. This means that from the intensity pattern I '( x . y . z ) of a real test object 1' even if the intensity distribution I ( x . y . z ) of an ideally positioned specimen 1 would be known exactly.

To clarify this fact, one can use a single test pattern point 19 consider that in the case of the ideal examinee 1 directly on its optical axis 10 lies and its sharply imaged picture point 20 correspondingly also in the image field on the optical axis 10 the ideal test object 1 lies. In the slightly twisted position of the real test object 1' becomes this test pattern point 19 no longer on its optical axis 10 ' lie; there is now another test pattern point 19 ' that is on the optical axis 10 ' of the real test object 1' is located, as well as an associated image point 20 ' which has an optimum MTF value. Thus one can the position of the optical axis 10 ' of the real test object 1' not determined from specific test pattern points or imaging points, but on the basis of the location of the sharpest points. For this, it is necessary that there are enough imaging points for a statistical evaluation.
In contrast to the intensity distribution I ( x . y . z ) or I ' (x . y . z ) are the sharpness distribution and thus the values of the modulation transfer function MTF ( x . y . z ) or MTF '( x . y . z ) spatially fixed with the location of the specimen 1 . 1' connected. Therefore, to determine the location and position of the real test object 1' relative to the measuring system S, not the positions of the points of the intensity distribution I '( x . y . z ), but their sharpness distribution. This is represented by a three-dimensional data field MTF '( x . y . z ) given by a Fourier transformation from the determined intensity distribution I '( x . y . z ) is calculated.

The orientation of the real test object 1' relative to the measuring system S with the coordinate directions x . y . z is determined as follows.

As already mentioned, one would lead to the measuring system S ideally aligned test piece 1 a detectable by the detector intensity distribution I ( x, y . z ). Fourier transformation could result in this intensity distribution I ( x . y . z ) a three-dimensional array of MTF values MTF ( x . y . z ) be calculated. Individual MTF values at defined points of this data field MTF ( x . y . z ) could be used to describe the quality of the candidate and the task would be fulfilled.

A real exam 1' is against the measuring system S tilted and moved, and therefore can by its own coordinate system S ' with the coordinate directions x ' . y ' . z ' the direction z 'with the direction of the real optical axis 10 ' matches.

In the measuring system S generates the real test object 1' an intensity distribution I '( x . y . z ), which differ from the ideal intensity distribution I ( x . y . z ), but the deviation of the real test object 1' from the position of the ideal test object 1 not directly reproduces.

From the intensity distribution I 'measured by the detector ( x . y . z ) of the real test object 1' Now, by means of a Fourier transformation, a three-dimensional data field MTF '( x . y . z ) calculated on the measuring system S is related.

Because of the coupling of the sharpness distribution to the positioning of the test object 1' corresponds to the searched data field MTF ( x . y . z ) the data field MTF '( x ' . y ' , z ') on the coordinate system S ' of the real test object 1' is related.

The empirically determined data field MTF '( x . y . z ) is thus still in the coordinate system S ' of the real test object 1' to convert the data field MTF '( x ' . y ' . z ' ) and correspondingly at the same time the data field MTF ( x . y . z ) to obtain.

Thus, the displacement and tilt of the coordinate systems remain to be determined S and S ' up to today. These properties can be completely described by a coordinate transformation, for example by a transformation matrix T that the coordinate system S in the coordinate system S ' transferred.

This transformation T also converts the MTF values MTF '( x . y . z ) into the data field MTF '( x ' . y ' . z ' ) and thus at the same time into the data field of an ideally positioned test object 1 MTF ( x . y . z ) through the operation: $$\text{MTF}\text{'}\left(\text{x}.\text{y}.\text{z}\right)*\text{T}=\text{MTF}\text{'}\left(\text{x}\text{'}.\text{y}\text{'}.\text{z}\text{'}\right)\equiv \text{MTF}\left(\text{x}.\text{y}.\text{z}\right),$$

The problem here is that in addition to T also MTF ( x . y . z ) is not yet known. MTF ( x . y . z ) therefore becomes a model data field MTF_Model (x . y . z ), which either from the known geometry of an ideal test specimen 1 by theoretical calculations, or by empirical acquisition of measurement data from prototypes of real candidates 1' and averaging is obtained.

The relation to the determination of the coordinate transformation matrix T is thus now: MTF '(x, y, z) * T ≈ MTF_model (x, y, z).

The determination of the coordinate transformation matrix T thus takes place by a statistical adaptation of the measured intensity values I '( x . y . z ) calculated data field MTF '( x . y . z ) and a predetermined model data field MTF _model ( x . y . z ). The adaptation takes place according to a best-fit method which determines the coordinate transformation matrix T so that the deviations between the data fields MTF '(x, y, z) * T and MTF_model ( x . y . z) total will be minimal. By statistically evaluating the largest possible number of measuring points, the transformation matrix can also be used to determine the position of the real test object 1' relative to the measuring system S be determined with a high accuracy. The positioning error of the DUT 1' relative to the measuring system S thus play no role anymore because the positioning of the specimen 1' from the MTF measurement itself is determined with.

The exact knowledge of the positioning of the measuring system S relative to the DUT system S ' allows the MTF data field MTF 'determined by measurement ( x . y . z ) into an MTF data field MTF ( x . y . z ), which is an ideally positioned candidate 1 and from the quality descriptive MTF values of the real test object 1' are calculable.

In some cases, the appearance of a specimen 1 . 1' designed so that the sharpest parts of the not at all on the optical axis 10 . 10 ' lie. Nevertheless, the theoretically expected sharpness distribution of an ideal candidate 1 with respect to the optical axis 10 and the main plane due to its geometry always known. Therefore, it is always possible the spatial location of the real test object 1' retroactively calculated from the measured sharpness distribution.

In the 1 It is assumed that the image field of the a plane is and the sharpest point is directly on the optical axis 10 . 10 ' located. In addition, it is assumed that the optical design is optimized for infinity. This means: if the test pattern 5 from the focus of imaging optics 6 is moved, can the Although by a shift of the detector, not shown here focus again, but the sharpness values will be worse overall than they would be in the case that the test pattern 5 in the nominal position, ie at a distance of the focal length f from the imaging optics 6 , would be.

In other words, the MTF values worsen, first, the farther away to the side of the center to the edge of the (from the optical axis of the DUT 1 . 1' to the side) and secondly the farther up or down the position of the DUT 1 . 1' deviates from the focal length f. Thus, the best MTF values lie in a plane which is at a distance of the focal length f of the test object 1' from the main plane of the specimen 1' is removed and perpendicular to the optical axis of the specimen 1' is oriented. The absolute best MTF value lies in this plane directly on the optical axis 10 . 10 ' , This is a theoretical behavior of the MTF values with respect to the coordinate system S 'of the real test object 1' described qualitatively.

To the location of the test object 1' with high accuracy, it is advantageous to scan a data field I '( x . y . z ) of as many and close measuring points in all three dimensions x . y , and z capture. However, these results are due to a discrete test pattern 5 in the also a discrete pattern of measurement points with a limited resolution. By interpolation, further data points can be added between the acquired measuring points. This makes it possible to take into account even very fine positional differences.

Instead of a camera lens as an individual part, there is often also the task of measuring the optical properties of a preassembled camera unit consisting of at least one camera lens and an image sensor in a fixed spatial arrangement. Since it is not possible to irradiate the entire unit, it is expedient to simultaneously use the image sensor of the camera unit as a detector 3 provided.

An advantageous embodiment of the measurement setup is in the 3 sketched in which the test pattern neither a uniform area, nor with an optic completely to infinite needs to be displayed. The test pattern here consists of several individual patterns 5a . 5b . 5c , These each have their own imaging optics 6a . 6b . 6c each provided a multipoint pattern 2a . 2 B . 2c maps, which in each case in a certain area of the detector 3 is detected.

As a result, instead of an otherwise relatively large and expensive imaging optics, a plurality of generally less expensive smaller imaging optics 6a . 6b . 6c be used and it is also a particularly compact test setup for carrying out the method of the invention possible. There may well be gaps in the picture, here the multipoint pattern 2a . 2 B . 2c forms and thus also when the intensity values are acquired. An arithmetical adaptation to the calculated MTF values of an MTF model field and a subsequent orientation of the test specimen 1' and its optical axis 10 ' are nevertheless possible with high accuracy according to the adaptation method proposed here.

Reference and formula symbols

1

Optical system (ideal test object)

1'

real examinee

2

Illustration

2a, 2b, 2c

More dot pattern

3

Detector (camera chip, CCD matrix)

5

test pattern

6

imaging optics

5a, 5b, 5c

individual samples

6a, 6b, 6c

Imaging optics (multi-part)

7a, 7b, 7c

subregions

9

Image field (sharpness plane, focal plane, curved image plane)

10, 10 '

Optical axis

11

target

12

Drive device (target drive)

13

(Detector) drive device

14

Computer (control)

15

beam path

16

light source

17

Data processing system

18

Drive device (imaging optics drive)

19, 19 '

Test pattern point

20, 20 '

imaging point

f

focal length

I (x, y, z)

Intensity data field (ideally positioned candidate)

I '(x, y, z)

Intensity data field (real positioned candidate)

MTF (x, y, z)

MTF data field (ideally positioned specimen)

MTF '(x, y, z)

MTF data field (real positioned test sample)

MTF '(x', y ', z')

Data field of MTF values (in the coordinate system S 'of the test object)

MTF_model (x, y, z)

MTF model field

S

Coordinate system of the measuring system

S '

Coordinate system of the test object

T

Transformation matrix (coordinate transformation)

Claims (12)

Method for testing an optical system (1, 1 '), in which a test pattern (5) after imaging with an imaging optics (6), via an optical system to be tested, which is referred to below as the test object (1, 1') , beam path (15) is imaged on an electronic light-sensitive detector (3), and in which the signals generated by this detector (3) are fed to a data processing system (17) for determining a modulation transformation function MTF, characterized in that a three-dimensional model field MTF values (MTF_model (x, y, z)) are computationally determined from the geometric structure of the test piece (1, 1 ') or statistically by measuring a plurality of prototypes of the test piece (1, 1') with the determination of mean values that the test pattern (5) and / or the imaging optics (6) and / or the detector (3) in the direction of the beam path (15) is moved, so that the detector (3) thereby a three-dimensional data field vo n intensity values (I '(x, y, z)) detects that thereafter the data processing system (17) from the detected intensity values (I' (x, y, z)) by a Fourier transform a three-dimensional array of MTF values (MTF ' (x, y, z)) calculates that the data processing system (17) the three-dimensional array of MTF values (MTF '(x, y, z)) and the Three-dimensional model field (MTF_Modell (x, y, z)) used by MTF values, and therefrom according to a best-fit method determines the parameters of a coordinate transformation (T), which a computationally optimal correlation of the data field (MTF '(x, y , z)) with the model field (MTF_model (x, y, z)), and that the data processing system (17) determines from the determined parameters the position of the optical axis (10) of the test object (1 '). Method for testing an optical system (1, 1 '), in which a test pattern (5) after imaging with an imaging optics (6), via an optical system to be tested, which is referred to below as the test object (1, 1') , beam path (15) is imaged on an electronic light-sensitive detector (3), and in which the signals generated by this detector (3) are fed to a data processing system (17) for determining a modulation transformation function MTF, characterized in that a three-dimensional model field MTF values (MTF_model (x, y, z)) are determined mathematically from the geometric structure of the test object (1, 1 ') or statistically by measuring a plurality of prototypes of the test object (1, 1') with the determination of average values (3) detects a two-dimensional data field of intensity values, that after that the data processing system (17) from the detected intensity values by a Fourier transform A two-dimensional data field of MTF values calculates that the data processing system (17) uses the two-dimensional data field of the MTF values and the three-dimensional model field (MTF_model (x, y, z)) of MTF values, and then for a best fit Method determines the parameters of a coordinate transformation (T) which results in a computationally optimal correlation of the data field with the model field (MTF_model (x, y, z)), and in that the data processing system (17) determines the position of the optical axis from the parameters determined ( 10) of the test piece (1 ') determined. Method according to Claim 1 or 2 , characterized in that MTF values which lie on or in a defined position relative to the determined optical axis (10 ') or to the spatial position of the test object (1') are determined as relevant MTF results. Method according to Claim 1 or 2 , characterized in that the three- or two-dimensional data field of the MTF values (MTF '(x, y, z)) for pixels not detected by the detector (3) is supplemented by interpolated intermediate values. Method according to Claim 1 or 2 , characterized in that during the detection of the two- or three-dimensional data field of intensity values I '(x, y, z) the test pattern (5) or the imaging optics (6) is moved laterally to the beam path (15). Method according to Claim 1 or 2 , characterized in that from the determined parameters in addition at least one of the following information is calculated: spatial orientation of the specimen (1 '), positioning error of the specimen (1'), positioning and orientation error of the detector (3) to the optics of the specimen (1 '). Method according to Claim 1 or 2 , characterized in that the specimen (1, 1 ') is an objective. Method according to Claim 1 or 2 , characterized in that the specimen (1, 1 ') is a camera system which has at least one pre-assembled arrangement of a lens and an image sensor. Method according to Claim 8 , characterized in that the image sensor of the camera system is also used as a detector (3) of the test method. Method according to Claim 1 , characterized in that the position of the specimen (1 ') is actively readjusted to bring the test pattern (5) with the computationally determined position of the optical axis (10') or the spatial position of the specimen (1 ') in accordance. Method according to Claim 1 , characterized in that displacements of the detector (3) on the one hand, and of the test pattern (5) or the imaging optics (6) on the other hand are alternately made in each case for each position of the test pattern (5) or its imaging position intensity data I '(x, y , z) at several positions of the detector (3). Method according to Claim 1 , characterized in that only one of the components detector (3), test pattern (5) and imaging optics (6) in the direction of the beam path (15) is moved.

Images