MATERIAL TESTING WITH ANGLE VARIABLE LIGHTING

Description

TECHNICAL AREA

[0001] The invention relates to an optical system with an illumination module that is configured to illuminate a sample object with several angularly variable illumination geometries.

[0002] The invention relates to techniques for determining a result image based on several transfer functions and several images of the sample object. HINTERGRUND

[0003] In materials testing, it is often desirable to generate a height profile of a sample object. Based on this height profile, anomalies can then be detected, for example.

[0004] One technique for determining an elevation profile is called "shape from shading" (SFS). See, for example, Prados, Emmanuel, and Olivier Faugeras. "Shape from shading." Handbook of mathematical models in computer vision (2006): 375-388. SFS has certain disadvantages and limitations. For example, it can often be complicated and time-consuming to obtain a sufficiently large number of images to reconstruct the elevation profile. Furthermore, such techniques can be time-consuming and prone to errors.

[0005] DE 102015110 339 A describes a determination of a height profile of a test object based on complementary illumination and a difference image calculation. BRIEF DESCRIPTION OF THE INVENTION

[0006] Therefore, there is a need for improved materials testing techniques. In particular, there is a need for techniques that address or mitigate at least some of the limitations and disadvantages mentioned above.

[0007] This problem is solved by the features of independent claims 1 and 13. The features of dependent claims define embodiments.

[0008] Based on the result image according to claim 1, anomalies in the result image can, for example, be made particularly visible. It would therefore be possible for the control system to be further configured to detect anomalies in the result image. In general, it may be possible to perform a precise material examination of the sample object using the result image. According to the claimed invention, a height profile of the sample object is visualized.

[0009] A computer program product, which, however, does not fall under the claimed invention, comprises program code that can be executed by at least one processor. The execution of the program code causes the at least one processor to execute a method. The method comprises illuminating a sample object with at least one angularly variable illumination geometry. The method further comprises imaging the sample object illuminated with the at least one angularly variable illumination geometry onto a detector. The method further comprises, based on the image, capturing at least one image of the sample object. The method further comprises determining a result image based on a transfer function and the at least one image.

[0010] A computer program, which is not covered by the claimed invention, comprises program code that can be executed by at least one processor. Execution of the program code causes the at least one processor to execute a method. The method comprises illuminating a sample object with at least one angularly variable illumination geometry. The method further comprises imaging the sample object illuminated with the at least one angularly variable illumination geometry onto a detector. The method further comprises, based on the image, capturing at least one image of the sample object. The method further comprises determining a result image based on a transfer function and the at least one image.

[0011] Using such techniques, it may be possible to reliably detect irregularities or impurities in the sample based on the resulting image, for example, by analyzing the height profile or by identifying anomalies. This is because the resulting image can depict a height profile of the sample.

[0012] The features set out above and those described below can be used not only in the corresponding explicitly set out combinations, but also in further combinations or in isolation, without leaving the scope of protection of the present invention as defined in the attached patent claims. BRIEF DESCRIPTION OF THE FIGURES

[0013] FIG. 1 Figure 1 schematically illustrates an optical system according to various examples, wherein the optical system includes an illumination module set up to illuminate a sample object with an angularly variable illumination geometry. FIG. 2 schematically illustrates the lighting module with a variety of lighting elements in greater detail. FIG. 3 schematically illustrates an exemplary lighting geometry that can be used to illuminate the sample object using the lighting module. FIG. 4 schematically illustrates an exemplary lighting geometry that can be used to illuminate the sample object using the lighting module. FIG. 5 schematically illustrates an exemplary lighting geometry that can be used to illuminate the sample object using the lighting module. FIG. 6 schematically illustrates a transfer function that can be used to determine a result image according to various examples. FIG. 7 schematically illustrates a transfer function that can be used to determine a result image according to various examples, where the transfer function is defined according to FIG. 7 compared to the transfer function according to FIG. 8 It is scaled. FIG. 8 schematically illustrates a transfer function that can be used to determine a result image according to various examples. FIG. 9 schematically illustrates transfer functions that can be used to determine a result image according to various examples. FIG. 10 This is a flowchart of an exemplary procedure. FIG. 11 schematically illustrates an illumination module and a detector in reflected light geometry according to various examples. FIG. 12 schematically illustrates an illumination module and a detector in reflected light geometry according to various examples. FIG. 13 schematically illustrates an illumination module and a detector in reflected light geometry according to various examples. FIG. 14 schematically illustrates an illumination module and a detector in reflected light geometry according to various examples. FIG. 15 Illustrates various images and result images with elevation profiles according to different examples as well as reference implementations. FIG. 16 This is a flowchart of an exemplary procedure. DETAILED DESCRIPTION OF EXECUTION FORMS

[0014] The properties, features and advantages of this invention described above, as well as the manner in which they are achieved, will become clearer and more easily understood in connection with the following description of the exemplary embodiments, which are explained in more detail in conjunction with the drawings.

[0015] In the figures, identical reference numerals denote identical or similar elements. The figures are schematic representations of various embodiments of the invention. Elements depicted in the figures are not necessarily shown to scale. Rather, the various elements depicted in the figures are represented in such a way that their function and general purpose are understandable to a person skilled in the art. Connections and couplings between functional units and elements shown in the figures can also be implemented as indirect connections or couplings. A connection or coupling can be implemented as a wired or wireless connection. Functional units can be implemented as hardware, software, or a combination of hardware and software.

[0016] Techniques for determining a result image with tailored contrast are described below. According to the claimed invention, the result image provides a height profile of the sample object. Therefore, the result image can be used for material testing of a sample object.

[0017] According to the claimed invention, the resulting image is determined by digital post-processing of several images of a sample object. For example, it would be possible that one or more of the images of the sample object are intensity images that do not themselves represent a height profile.

[0018] According to the claimed invention, the multiple images of the sample object are associated with different illumination geometries. This means that the multiple images are each captured by a detector when the sample object is simultaneously illuminated using a corresponding illumination geometry.

[0019] The different illumination geometries can be associated with different illumination directions. These different illumination geometries, or the associated different images, can be separated by time-division multiplexing or frequency-division multiplexing. Separation using different polarizations would also be possible. The illumination geometries can exhibit a directional dependency; for example, they can show a gradient of illuminance along one or more spatial directions. For instance, the illuminance could vary in steps along a spatial direction, such as between zero and a finite value, or between two different finite values.

[0020] The sample object could be opaque to the light used. For example, the sample object could comprise a bulk solid element, such as a textile or fabric, a metallic component, etc. According to the claimed invention, the illumination module and the detector are operated in reflected light geometry.

[0021] According to the claimed invention, several transfer functions are used for the digital post-processing of multiple images to obtain the resulting image. For example, the transfer function can be an object transfer function and / or an optical transfer function of the optical system. The transfer function can be suitable for predicting at least one image under a specific illumination and with a specific sample object. For example, the transfer function can have a real component and / or an imaginary component. Techniques for determining the imaginary component of the transfer function are described below, among others. For the sake of simplicity, it is not always explicitly stated that the techniques relate to the imaginary component of the transfer function. In some examples, a purely imaginary transfer function without a real component can be used.

[0022] Various techniques can be used to determine the transfer function. Depending on the transfer function used, different techniques can be used to determine the resulting image. According to the claimed invention, the technique is based on a Tikhonov regularization, i.e., an image is obtained using an inverse Fourier transform based on the transfer function. H * and furthermore based on the spatial frequency representation of a combination Ĩ DPC determined from two images of the sample object under different lighting geometries: F − 1 ∑ j H j ∗ u ⋅ I ˜ DPC , j u ∑ j H j u 2 + α α is a regularization parameter. It describes Ĩ DPC the spectral decomposition of a combination of two images I T and I B , which were captured under different lighting geometries illuminating complementary semicircles: I DPC = I T − I B I T + I B

[0023] Dies sind Examples. In general, the illumination geometry does not have to be strictly semicircular. For example, four LEDs could be used, arranged in a semicircle. Defined or discrete illumination directions could also be used, i.e., individual LEDs. Furthermore, in Eq. 2, the normalization to one could be performed instead of to I T + I B , or to another value. Instead of a settlement of I T and I B In other examples, the raw data itself could also be used, e.g. I DPC = I T or I DPC = I B . By forming a corresponding quotient in Eq. 2, otherwise interfering influences such as other material properties, color, etc., can be reduced. In particular, by calculating the difference, any absorption component due to a real-valued part of the transfer function can be reduced. Ĩ DPC is proportional to the local increase in phase shift due to the sample object. The phase shift can be caused by a change in the thickness or topography of the sample object and / or by a change in its optical properties.

[0024] For example, two pictures I DPC ,1 as well I DPC ,2 can be determined, once with a pair of semicircular illumination geometries arranged top-bottom in a lateral plane perpendicular to the beam path ( I DPC ,1 ), and once with a pair of semicircular lighting geometries arranged left-right in the lateral plane ( I DPC ,2 ). Then a combination of I DPC ,1 with I DPC ,2 are taken into account when determining the result image, i.e. by summing over the summation index j in Eq. 1.

[0025] The spatial frequency domain can be defined as the conjugate of the spatial domain. Transformations between the spatial domain and the spatial frequency domain can be performed using Fourier analysis and inverse Fourier analysis. Spatial frequencies represent the inverse of a spatial period length.

[0026] Such techniques rely on certain assumptions and simplifications, for example, in the case of the aforementioned formulation of a weak object approximation. However, other approximations and formalisms can be used in other examples. For instance, a different inversion could be used instead of Tichonov regularization, such as direct integration or a differently implemented Fourier filter. Even in such modifications, the fundamental properties of the transfer function, as described in the various examples herein, can be preserved.

[0027] FIG. 1 Figure 100 illustrates an exemplary optical system. For example, optical system 100 could be designed according to the example of the FIG. 1 Implement a light microscope in reflected light geometry. For example, a suitable light microscope in reflected light geometry could be used for material testing. A height profile of the sample object can be created for this purpose.

[0028] Using the optical system 100, it is possible to magnify small structures of a sample object fixed by a sample holder 113. For example, the optical system 100 could implement a wide-field microscope in which a sample is illuminated across its entire surface. In some examples, the imaging optics 112 can generate an image of the sample object on a detector 114. The detector 114 can then be configured to capture one or more images of the sample object. Viewing through an eyepiece is also conceivable.

[0029] In some examples, imaging optics 112 with a large aperture can be used. For example, the imaging optic 112 could have a numerical aperture of not less than 0.2, optionally not less than 0.3, and further optionally not less than 0.5. For example, the imaging optic 112 could have an immersion objective.

[0030] The optical system 100 also includes an illumination module 111. The illumination module 111 is designed to illuminate the sample object, which is fixed to the sample holder 113. For example, this illumination could be implemented using Köhler illumination. This involves the use of a condenser lens and a condenser aperture diaphragm. This results in a particularly homogeneous intensity distribution of the light used for illumination in the plane of the sample object. For example, partially incoherent illumination can be implemented. The illumination module 111 could also be configured to illuminate the sample object in dark-field geometry.

[0031] In the example of the FIG. 1 The lighting module 111 is configured to enable angle-variable illumination. This means that different illumination geometries of the light used to illuminate the sample can be implemented using the lighting module 111. These different illumination geometries can correspond to illuminating the sample from different directions. Angle-variable illumination is therefore sometimes referred to as angular space-structured illumination.

[0032] In the various examples described herein, different hardware implementations are possible to provide the different lighting geometries. For example, the lighting module 111 could include several adjustable lighting elements configured to locally modify or emit light. A controller 115 can control the lighting module 111 or the lighting elements to implement a specific lighting geometry.

[0033] For example, the control unit 115 could be implemented as a microprocessor or microcontroller. Alternatively or additionally, the control unit 115 could, for example, comprise an FPGA or ASIC. The control unit 115 can also, alternatively or additionally, control the sample holder 113, the imaging optics 112, and / or the detector 114.

[0034] FIG. 2 illustrates aspects relating to the lighting module 111. In FIG. 2 The figure shows that the lighting module 111 has a multitude of adjustable lighting elements 121 in a matrix structure. The matrix structure is oriented in a plane perpendicular to the path of the light rays (lateral plane; spatial coordinates x, y).

[0035] Instead of a matrix structure, it would also be possible in other examples to use other geometric arrangements of the adjustable elements, for example ring-shaped, semicircular, etc.

[0036] In one example, the adjustable lighting elements 121 could be implemented as light sources, for instance as LEDs. Then, for example, different LEDs with varying luminous intensities could emit light to illuminate the sample. This allows for the implementation of a specific lighting geometry. In another implementation, the lighting module 111 could be implemented as a spatial light modulator (SLM). The SLM can perform spatially resolved manipulation of a condenser pupil, which can have a direct impact on the imaging.

[0037] FIG. 3 illustrates aspects relating to an exemplary lighting geometry 300. In FIG. 3 The provided luminous intensity 301 for the various adjustable elements 121 of the lighting module 111 along the axis XX' is FIG. 2 The lighting geometry 300 is shown. It is dependent on the position along the XX' axis and is therefore structured.

[0038] FIG. 4 illustrates aspects relating to an exemplary lighting geometry 300. FIG. 4 The lighting geometry 300 is illustrated abstractly from the lighting module 111 used. In the example of the FIG. 4 A lighting geometry of 300 is used, in which one side is illuminated (black color in FIG. 4 ) and the other side is not illuminated (white color in FIG. 4 ). In FIG. 5 Another exemplary lighting geometry is shown (with corresponding color coding, as already mentioned in relation to FIG. 4 (described).

[0039] FIG. 6 illustrates aspects relating to an exemplary transfer function 400 (where in FIG. 6 Black represents a value of +1 and white a value of -1; the coordinates u x and u y are defined in the spatial frequency domain and correspond there to the spatial coordinates x and y). The transfer function 400 can be used to determine, based on an image which, for example, has the lighting geometry 300 according to the example of the FIG. 4 The goal was to determine a result image after the data was collected. The result image can include a height profile of the sample object.

[0040] In the example of the FIG. 6 The transfer function 400 has a symmetry axis 405, which corresponds to a symmetry axis 305 of the illumination geometry 300. This makes it possible to choose the transfer function 400 to match the illumination geometry 300. As a result, the image can exhibit particularly strong contrast.

[0041] In FIG. 6 The diameter of the detector aperture of the imaging optics 112 is also shown. Because partially incoherent illumination is used, the transfer function is non-zero up to twice the size of the detector aperture of the imaging optics 112.

[0042] FIG. 7 It also illustrates aspects relating to a transfer function of 400. The example of the FIG. 7 This basically corresponds to the example of FIG. 6 However, in the example of the FIG. 6 the size of the detector aperture is larger than in the example of the FIG. 6 (see horizontal dashed lines; N A (refers to the size of the detector aperture).

[0043] However, the transfer function 400 is scaled accordingly to the one in FIG. 7 compared to FIG. 6 enlarged detector aperture. For example, the transfer function 400 could be as shown in the example of the FIG. 6 to serve as a reference transfer function. Then, for example, the controller 115 could be configured to use the transfer function 400 according to the example of the FIG. 7 to determine based on a scaling of this reference transfer function to the enlarged aperture of the imaging optics 112.

[0044] Based on such techniques, it is possible to use a particularly large aperture for the imaging optics 112, which may be desirable in certain applications for imaging using the optical system 100.

[0045] From the examples of FIGs. 6 und 7 It is evident that the range within which the transfer function 400 assumes non-zero values ​​can be determined by the size of the aperture of the imaging optics 112.

[0046] Such techniques are based on the understanding that even for transfer functions scaled according to the aperture size of the imaging optics, it may be possible to determine the resulting image with a meaningful contrast – for example, a height profile of the sample object. In some cases, the contrast in the resulting image may not provide a quantitative description of the sample object's topography, but it may provide a qualitative description. In particular, the qualitative description of the sample object's topography can be provided consistently across the entire image. This can be advantageous, especially compared to reference techniques where different spatial gradients of the sample object's topography – for example, at opposite edges of the sample object – are represented with different contrast signs in the resulting image.

[0047] FIG. 8 illustrates aspects relating to a transfer function 400 (where in FIG. 8 Black represents a value of +1 and white a value of -1; the coordinates u x and u y are defined in the spatial frequency domain and correspond there to the spatial coordinates x and y). The transfer function 400 can be used to determine, based on an image which, for example, has the lighting geometry 300 according to the example of the FIG. 5 The recording was used to determine a result image. In FIG. 8 The diameter of the detector aperture of the imaging optics 112 is also shown.

[0048] From the FIGs. 6-8 It is evident that the transfer function 400 can be determined as a function of the angularly variable illumination geometry 300. In particular, it is possible for the geometry of the transfer function 400 in the spatial frequency domain to model the illumination geometry 300 in the spatial domain. Such techniques can achieve a particularly strong contrast in the resulting image, i.e., a high signal-to-noise ratio, for example, for the elevation profile.

[0049] FIG. 9 illustrates aspects relating to different transfer functions 400 (different transfer functions are in FIG. 9 (represented by the solid line, the dashed line, the dotted line, and the dashed-dotted line).

[0050] The in FIG. 9 The transfer functions shown (400) can, for example, be used for different lighting geometries (in FIG. 9 (The lighting geometries are not shown).

[0051] In FIG. 9 The transfer function is 400 along an axis. u x The spatial frequency domain is represented. In some examples, the transfer function might only vary along one coordinate of the spatial frequency domain; in other examples, however, it might vary along two orthogonal axes. u x and u y are available.

[0052] In the example of the FIG. 9 For example, a transfer function 400 is formed as a monotonically increasing linear function (solid line). Furthermore, in the example, the FIG. 9 Another transfer function 400 is formed as a monotonically increasing sigmoid function (dashed line). In the example of the FIG. 9 Another transfer function, 400, is formed as a convolved, monotonically decreasing, linear function (dotted line). In the example of the FIG. 9 Another transfer function 400 is designed as a step function (dotted line).

[0053] Such examples of transfer functions 400 are purely illustrative, and other examples may use differently shaped transfer functions or superpositions of those in the example. FIG. 9 The transfer functions shown are 400. However, the transfer functions used in the various examples described herein may exhibit certain features or characteristics that allow for a particularly accurate determination of the result. Such features of the transfer functions used are described below.

[0054] From the examples of transfer functions 400 in FIG. 9 It is evident that it is possible to form the transfer functions 400 for spatial frequencies within the aperture of the imaging optics 112 without local extrema, i.e., without local maxima or minima that would be smaller than the absolute extrema (i.e., the amplitudes of +1 and -1 respectively in the example of the FIG. 9 This can be achieved through a monotonically increasing or decreasing transfer function, or through a step function.

[0055] Avoiding local extreme values ​​in this way can have particularly beneficial effects on reducing signal noise and artifacts in the resulting image. Sometimes, due to structural reasons, there may be a discrepancy between the actual aperture and the nominal aperture of the imaging optics. In such cases, the position of the local extreme values ​​of the transfer function may be incorrectly positioned in the spatial frequency domain relative to the actual aperture. This causes frequencies included in the images to be erroneously amplified strongly due to the incorrect positioning of these local extreme values ​​in the spatial frequency domain, which can lead to artifacts in the resulting image. By using a transfer function without local extreme values ​​within the detector aperture, as described in the various examples herein, the following can be achieved:By using a dual detector aperture, such erroneous amplification of frequencies contained in the images due to a shifted local extremum of the transfer function can be avoided. This results in uniform propagation of the frequencies contained in the captured images.

[0056] From the in FIG. 9 The examples of transfer functions 400 shown further demonstrate that implementations are possible in which the transfer function for spatial frequencies within the aperture of the imaging optic 112 or within twice the aperture of the imaging optic 112 assumes no values ​​or essentially no values ​​equal to zero, i.e., only finite non-zero values. In general, it can sometimes be desirable to avoid the transfer function for spatial frequencies within the aperture of the imaging optic 112 or within twice the aperture of the imaging optic 112 assuming comparatively small values—for example, with respect to a maximum of all magnitude values ​​of the transfer function for spatial frequencies within the corresponding range.For example, it would be possible for the transfer function for spatial frequencies within the aperture or the double aperture of the imaging optic 112 to have no absolute values ​​<5% of a maximum of all absolute values ​​of the transfer function 400 for spatial frequencies within the aperture of the imaging optic 112, optionally no values ​​<2%, and further optionally no values ​​<0.5%. Such behavior can be provided, for example, by a step function.

[0057] Such techniques are based on the understanding that values ​​of zero for the transfer function 400 can correspond to the suppression of the corresponding frequencies contained in the images. However, it is often desirable that no suppression of corresponding frequencies contained in the images occurs within the aperture of the imaging optics 112 or within the double aperture of the imaging optics 112.

[0058] From the example of FIG. 9 It is further evident that the transfer functions shown there assume values ​​of zero for spatial frequencies outside the double aperture of the imaging optic 112. In general, it may be possible to use transfer functions that assume values ​​of essentially zero outside the spatial frequencies transmitted by the imaging optic 112, i.e., typically outside the single aperture or the double aperture with partially phase-incoherent illumination. For example, it would be possible that the transfer functions used for spatial frequencies outside the single or double aperture of the imaging optic have no absolute values ​​>5% of a maximum of all absolute values ​​of the transfer functions for spatial frequencies inside the single or double aperture of the imaging optic, optionally no values ​​greater than 2%, and further optionally no values ​​greater than 0.5%.This prevents artifacts or noise from being amplified in the resulting image.

[0059] FIG. 10 This is a flowchart of an exemplary procedure. First, a sample object is fixed in step 1001, for example, using a sample holder. The sample object could, for example, be a material sample of a bulk material. Step 1001 is optional.

[0060] Then, in module 1002, the sample object is illuminated with several pairs of angle-variable illumination geometries. A corresponding illumination module can be controlled accordingly. For example, it would be possible to illuminate the sample object with several pairs of two complementary illumination geometries, which could be semicircular and correspond to different semicircles.

[0061] In 1003, multiple images of the sample object are acquired using imaging optics and a detector, for example, a CMOS or CCD sensor. 1003 can include the corresponding control of the detector. Each image contains a representation of the sample object. Different images with different illumination geometries from 1002 are associated with each image.

[0062] In some examples, two pairs of images can be captured, each associated with complementary, semicircular lighting directions. In other examples outside the claimed invention, however, only two or three images could be captured.

[0063] A difference could then be calculated, e.g. according to I links − I rechts I links + I rechts I oben − I unten I oben + I unten where Ilinks and Irechts the images are each associated with a left- or right-oriented semicircular lighting geometry and wherein Ioben and lunten the images are each associated with a top- or bottom-oriented semicircular lighting geometry.

[0064] Then, in 1004, a result image is determined, which represents a height profile of the sample object. The determination of the result image in 1004 is based on several transfer functions that describe the imaging of the sample object using the corresponding optical system for the respective illumination geometries. The result image is also determined based on the images acquired in 1003. According to the claimed invention, this involves calculating differences and, if necessary, normalizing several images acquired in 1003 that are associated with different illumination geometries.

[0065] For example, the procedure could be according to FIG. 10 Furthermore, this includes scaling a reference transfer function to the size of the aperture of the imaging optics. This means that the reference transfer function can be adapted to the size of the aperture of the imaging optics.

[0066] In summary, the preceding section described techniques for determining a result image with strong contrast, even with comparatively large apertures of the imaging optics used, which, for example, encodes the height of a sample object. These techniques are based on taking the aperture size of the imaging optics into account. For example, a predefined reference transfer function can be scaled according to the aperture size of the imaging optics. The reference transfer function can therefore also be called an artificial transfer function because it may deviate from the transfer function theoretically expected based on the illumination geometry.

[0067] The techniques described herein, relating to the transfer function according to the claimed invention, are combined with a reflected light geometry. This is illustrated by the example of the FIG. 11 depicted. FIG. 11 illustrates aspects relating to optical system 100. In particular, illustrates FIG. 11 Aspects relating to the arrangement of the illumination module 111 and the detector 114 in relation to a sample holder 113, which fixes a sample object 181. The sample object 181 comprises individual impurities 180, for example lint.

[0068] In the example of the FIG. 11 are the illumination module 111 and the detector or the imaging optics (in FIG. 11 (not shown) arranged in reflected light geometry, i.e. on the same side of the sample holder 113. In the example of the FIG. 11 The lighting elements 121 are implemented by light sources. For example, the lighting elements 121 could be implemented by light-emitting diodes, etc.

[0069] In some examples, a central aperture could also be provided, which makes it possible to illuminate the sample object 180, 181 in dark field geometry.

[0070] By means of an optical system 100 according to the example of the FIG. 11 It may be possible to perform a material analysis of the sample objects 180, 181. For example, anomalies – such as impurities 180 – can be detected. The sample objects 180, 181 do not need to be transparent. With such techniques, a height profile of the sample objects 180, 181 can be determined by means of angle-variable illumination. The impurities 180 then become clearly visible. FIG. 12 illustrates aspects relating to the optical system 100. FIG. 12 This is a top view of the illumination module. It shows that the illumination module 111 has four illumination elements 121 arranged concentrically with respect to the detector 114, serving as light sources. Variations of the illumination module of the 111 would also be possible; compare FIGs. 13 and 14 .

[0071] Based on the resulting image, which is determined according to the techniques described herein using the transfer function 400, certain applications can then be implemented. The result, according to the claimed invention, is indicative of the height profile of the sample object 180, 181.

[0072] For example, it would be possible to determine areas of equal height based on the resulting image. In general, based on a resulting image determined, for example, using the techniques described herein and based on the transfer function, anomalies can be detected that correspond, for example, to impurities 180.

[0073] This means that certain deviations from the norm (anomalies) can be identified in the resulting image. The norm can be determined, for example, based on reference ranges in the resulting image, or it can be predefined. Such techniques are based on the understanding that anomalies can be particularly well identified using the height profile of the sample. For example, the anomalies could correspond to impurities 180, such as lint, which are arranged on the sample and thus represent a deviation from the general topography of the sample 181.

[0074] Different techniques can be used to detect the anomalies in the various examples described herein. In In some examples, it might be possible to identify the anomalies as extreme values ​​in the contrast of the result image. For instance, if the result image describes the elevation profile of sample object 181, it may be possible to recognize the extreme values ​​as local deviations from the mean elevation of the sample object's topography. Such extreme values ​​could, for example, be determined based on a threshold comparison. The threshold considered in the threshold comparison could be determined with respect to an absolute maximum and / or with respect to a mean of the contrast values ​​of the result image.

[0075] Alternatively or additionally, the anomalies can also be determined using machine learning techniques. For example, artificial neural networks could be used to classify local deviations from the norm.

[0076] Alternatively or additionally, techniques can also be used to detect anomalies by exploiting a deviation from a spatial pattern of sample object 180. For example, sample object 181 could be a fabric sample that exhibits a certain periodicity of topography due to its tissue structure. In In some examples, it would be possible that such periodicity or the spatial pattern of the topography of the sample object is known as a priori information. In In other examples, it would also be possible that the spatial pattern is determined by a global frequency analysis of sample object 181 as a reference. Then it would be possible to identify the anomalies as local deviations from this spatial pattern.

[0077] Such local deviations from the spatial pattern can be detected in various ways. For example, a reference image could be generated based on the resulting image, in which the spatial pattern is suppressed. The anomalies would then be particularly noticeable in the reference image. Deviations from the spatial pattern of the sample object can be detected, for example, by filtering in the spatial frequency domain. Alternatively or additionally, autocorrelation can be performed, which emphasizes changes in periodicity. Deviations from a reference image of the spatial pattern can also be detected.

[0078] FIG. 15 illustrates experimental results for the various techniques described herein.

[0079] Image 2001 contains an image of a textile fabric. Image 2001 was created using reflected light geometry with an illumination module 111 according to FIG. 12 recorded.

[0080] Image 2002 was obtained using techniques described in DE 10 2014 112 242 A1. Image 2002 illustrates a contamination of the textile fabric. It is clearly visible in image 2002 that the upper edge of the contamination exhibits a white contrast and the lower edge a dark contrast, while the central area of ​​the contamination has a contrast corresponding to the average contrast of image 2002. Therefore, while image 2002 does exhibit a contrast that encodes the height of the sample object, it does not depict a height profile but is instead proportional to the gradient of the height profile.

[0081] This differs from result image 2003. Result image 2003 was determined according to the techniques described herein, i.e., using a suitable transfer function. In result image 2003, the contaminant is also marked with an arrow. It is clearly visible that the contaminant exhibits a brighter contrast in its center than along its edges. The contrast of the contaminant in its central region is also brighter than the average contrast of result image 2003. Thus, result image 2003 indicates the height profile of the sample. The various contaminants or lint particles stand out with particularly bright contrast in result image 2003.

[0082] This can be used in various examples to detect anomalies in the result image 2003. A simple implementation would allow for the identification of local extreme values ​​of the contrast in the result image 2003. For example, a threshold comparison could be performed, where the relevant threshold could be determined, for instance, in relation to a global maximum or a global minimum of the contrast values ​​in the result image 2003. A corresponding evaluation is shown in Figure 2004.

[0083] The textile fabric depicted in images 2001-2004 also exhibits a characteristic spatial pattern. This spatial pattern corresponds to the weave of the yarn used. Alternatively or additionally to the threshold comparison described above, it may also be possible to detect the anomalies based on this spatial pattern. For example, these anomalies could be identified as deviations from the spatial pattern of the sample object. This could be achieved, for instance, by performing filtering in the spatial frequency domain or, for example, by conducting one- or two-dimensional autocorrelation. In other examples, a reference image depicting the spatial pattern could also be used.

[0084] FIG. 16 Figure 1011 is a flowchart of an exemplary procedure. In this procedure, anomalies are detected in a result image. For example, the result image could be generated by a procedure according to the example of... FIG. 10be determined.

[0085] Naturally, the features of the embodiments and aspects of the invention described above can be combined with one another. In particular, the features can be used not only in the combinations described, but also in other combinations or individually, without leaving the scope of the invention.

[0086] The scaling of the amplitudes of the various transfer functions described herein is purely illustrative. For example, transfer functions with an amplitude of +1 or -1 have often been presented in the various examples described herein; however, it may also be possible to use transfer functions with other amplitudes in other examples.

[0087] Furthermore, various implementations involving the illumination of the sample object with partially incoherent light have been described. In these cases, the bandwidth of the transmitted spatial frequencies is equal to twice the aperture of the imaging optics. However, other illumination techniques could be used in other examples, resulting in a different bandwidth for the transmitted spatial frequencies. This can be taken into account in the various examples described herein by, for instance, appropriately scaling a reference transfer function up to the theoretical maximum of the transmitted spatial frequencies.

[0088] Furthermore, various examples relating to the detection of anomalies in the result image have been described above. In general, other material testing techniques could also be used based on the result image.

Claims

1. Optical system (100), comprising: - an illumination module (111), which is set up for illuminating a sample object (180, 181) with multiple pairs of angle-variable illumination geometries (300) which are associated with different illumination directions, - an imaging optical unit (112), which is set up for producing an imaged representation of the sample object (180, 181) that is illuminated with the multiple pairs of angle-variable illumination geometries (300) on a detector (114), - the detector (114), which is arranged together with the illumination module (111) in incident light geometry and which is set up for capturing, on the basis of the imaged representation, multiple pairs of images of the sample object (180, 181) in the multiple pairs of angle-variable illumination geometries, and - a controller (115), which is set up for determining a result image, which represents a height profile of the sample object, on the basis of multiple transfer functions (400) and a pairwise subtraction of the multiple pairs of images by way of digital post-processing of the multiple pairs of images, wherein each of the multiple transfer functions (400) is determined on the basis of a corresponding pair of the multiple pairs of angle-variable illumination geometries (300), and wherein, when determining the result image by means of an inverse Fourier transform, a summation is performed over the product of each of the multiple transfer functions and the respective corresponding spatial frequency domain representation of the pairwise subtractions of the respective corresponding pair of images.

2. Optical system (100) according to Claim 1, wherein the illumination module (111) is set up to illuminate the sample object (180, 181) in dark field geometry.

3. Optical system (100) according to Claim 1 or 2, wherein the controller (115) is furthermore set up for detecting anomalies in the result image, wherein the anomalies comprise extremal values in the contrast of the result image.

4. Optical system (100) according to any of the preceding claims, wherein the controller (115) is furthermore set up for detecting anomalies in the result image, wherein the anomalies comprise deviations from a real-space pattern of the sample object (180, 181).

5. Optical system (100) according to Claim 4, wherein the controller (115) is set up to determine the deviations from the real-space pattern of the sample object (180, 181) based on at least one of the following techniques: Fourier space filtering; autocorrelation; deviations with respect to a reference image of the real-space pattern.

6. Optical system (100) according to any of the preceding claims, wherein the transfer functions (400) have for spatial frequencies within the single aperture or twice the aperture of the imaging optical unit (112) no absolute values smaller than 5% of a maximum of all absolute values of the transfer functions (400) for spatial frequencies within the single aperture or twice the aperture of the imaging optical unit (112), optionally no values smaller than 2%, further optionally no values smaller than 0.5%.

7. Optical system (100) according to any of the preceding claims, wherein the transfer functions (400) for spatial frequencies within the aperture of the imaging optical unit (112) have no local extremal values.

8. Optical system (100) according to any of the preceding claims, wherein the transfer functions (400) are step functions.

9. Optical system (100) according to any of the preceding claims, wherein the transfer functions (400) are monotonically increasing or monotonically decreasing functions, optionally linear functions or sigmoid functions.

10. Optical system (100) according to any of the preceding claims, wherein the transfer functions (400) have an axis of symmetry (405) that corresponds to an axis of symmetry (305) of the at least one pair of illumination geometries (300).

11. Optical system (100) according to any of the preceding claims, wherein the transfer functions (400) have for spatial frequencies outside the single aperture or twice the aperture of the imaging optical unit (112) no absolute values greater than 5% of a maximum of all absolute values of the transfer function (400) for spatial frequencies within the single aperture or twice the aperture of the imaging optical unit (112), optionally no values greater than 2%, further optionally no values greater than 0.5%.

12. Optical system (100) according to any of the preceding claims, wherein the controller (115) is set up for determining the result image based on a Tikhonov regularization with inverse Fourier transform.

13. Method, which comprises: - illuminating a sample object (180, 181) with multiple pairs of angle-variable illumination geometries (300) which are associated with different illumination directions, - imaging the sample object (180, 181) illuminated with the multiple pairs of angle-variable illumination geometries (300) on a detector (114), on a detector in reflected-light geometry, - on the basis of the imaged representation: capturing multiple pairs of images of the sample object (180, 181) in the multiple pairs of angle-variable illumination geometries by the detector, and - on the basis of multiple transfer functions (400) and a pairwise subtraction of the multiple pairs of images: determining a result image by way of digital post-processing of the multiple images, wherein each of the multiple transfer functions (400) is determined on the basis of a corresponding pair of the multiple pairs of angle-variable illumination geometries (300), and wherein, when determining the result image by means of an inverse Fourier transform, a summation is performed over the product of each of the multiple transfer functions and the respective corresponding spatial frequency domain representation of the pairwise subtractions of the respective corresponding pair of images.