Method and device for determining the profile of a frame groove

Abstract

The invention relates to a method for determining the outer shape (34) of a frame groove (12c) in a bezel (12a) of a spectacle frame (12). The method comprises illuminating the bezel (12a), capturing a plurality of images of the illuminated bezel (12a) from different predetermined viewing angles, and evaluating the captured images and determining the spatial curve of the frame groove (12c) based on the evaluated images. The method is characterized in that the bezel (12a) is illuminated along the entire circumference of the bezel (12a) by directional illumination. Furthermore, the method is characterized in that the evaluation of the captured images comprises assigning each portion contained in the captured images to a corresponding surface element (F1, . . . , F4) of the frame groove (12c) based on at least one of the following characteristics: a shading of the corresponding portion, a brightness of the corresponding portion, and a phase angle of the illumination of the corresponding portion. Furthermore, the invention relates to a device (10), a computer program, a method for grinding a spectacle lens, and a computer-implemented method for determining the geometric shape of a spectacle lens.

Description

Technical Field

[0001] The invention relates to a method and a device for determining the profile of a frame recess in a bezel of a pair of spectacles according to the preambles of the respective independent patent claims. The invention also relates to a computer program product and the use of a device for determining the vertex power and / or the use of a device for creating a spatially resolved power map. The invention therefore belongs in particular to the field of spectacles technology. Background Art

[0002] In order to grind a spectacle lens for fitting purposes, i.e. in order to adapt the spectacle lens to the dimensions of the bezel of a corresponding spectacle frame, the outer shape of the bezel, in particular the outer shape of the groove in the bezel, must be measured so that the dimensions and the curve of the outer shape are available as digital information. The outer shape is subsequently used to grind the spectacle lens. Conventionally, the groove or bezel is measured by means of a tactile tracer, the measuring needle of which dips into the groove of the bezel, mechanically probes the groove of the bezel and travels along a so-called track in the process. The information determined in this process is usually provided in the form of three-dimensional coordinates of the groove. Typically, the required measurement accuracy is about 1 / 10 millimeter or 100 µm in order to ensure that the spectacle lens fits precisely in the spectacle frame or in the groove.

[0003] For a long time, the industry has been looking to simplify the workflow for users, especially to avoid the frequent need to recalibrate the probes of tactile tracers. There have been some approaches to replace tactile measurement with optical measurement for this purpose, but none have been able to provide the required measurement accuracy and reliability within an economical budget.

[0004] In this case, the requirements when determining the outer shape of the frame recesses include in particular the ability to determine the outer shape of the two frame recesses of the spectacle frame independently of one another, since tolerances in the manufacturing process of the spectacle frame can lead to an asymmetry of the two rims that exceeds the grinding tolerances of the lenses. In particular, it is necessary to register asymmetries that lead to deviations of more than 0.1 mm between the rims. The absolute geometric inaccuracy of the frame outer shape must not exceed 0.1 mm, so that a lens ground based on a determined outer shape sits flush in the frame, that is to say does not wobble.

[0005] Furthermore, it must be possible to detect the bezel independently of the material and color of the frame and independently of the scatter or reflectivity of the frame material.

[0006] DE 19937647 C1 discloses a device and a method for determining the outer shape of a spectacle frame, in which case an attempt is made to record different views of a groove by means of a camera and a device for generating a shaking motion. In this case, the shaking motion can be generated a) by a stepper motor or b) by a rotatable mirror mounted below the camera and below the spectacle frame to be measured. A similar device is known from DE 4019866 A1.

[0007] FR 2763707 A has disclosed a device for determining the vertex power and frame shape of a pair of spectacles, within the scope of which the pair of spectacles is illuminated from behind by a flat light source (e.g. a translucent screen). The 2D shape of the rim is determined based on the black contour visible in the camera image, wherein some needles are introduced into the grooves in a motor-driven manner or by means of springs, and the position of the needles is determined in order to determine the shape of the grooves.

[0008] DE 4224640 A discloses an apparatus intended for optically reconstructing a groove formed in a "V" form. In this case, two parallel light beams are directed at the groove from two different angles by means of two light sources and appropriate optics, and the reflected or scattered light is recorded by a camera. The three distinct points of the "V" are reconstructed by triangulation. In each case, the frame groove is moved into the measurement field by means of a rotation device.

[0009] US 9316489 B2 discloses a device intended for reconstructing a spectacle frame as a 3D model by means of a laser line, a camera, a linear drive and a rotary table. A corresponding profile section is generated in each image recording. The entire 3D model is generated from the entire set of profile sections.

[0010] DE 19727226 A1 discloses a device for determining the spatial shape of a groove, within the scope of which a light beam or laser beam is guided to the center of the groove by means of deflection mirrors. The spatial shape of the groove is determined by triangulation. In this case, the mirror holder must also be moved by means of a rotary table. A corresponding profile section is generated in each image recording.

[0011] DE 19919311 C1 describes a method and a device for contactless scanning of a groove in a spectacle frame opening, the groove being scanned 360° in a first plane above or below the opening plane of the spectacle frame or spectacle lens or lens former and also in lowered and raised planes.

[0012] Chapter 8 of Bernd Jähne's book: "Digital Image Processing", Springer-Verlag, Berlin, January 1, 2002, describes the so-called aperture problem in motion analysis.

[0013] Furthermore, various tactile tracers are known which can determine the outer shape of two rims of a pair of spectacles with an accuracy of about 0.01 mm and a total measuring time of about 30 seconds (for 16 000 measuring points, 2 frame openings). In this process, a rotatably mounted stylus is guided along a groove. In doing so, a force is exerted on the groove and therefore on the rim, as described, for example, in “http: / / www.ftpuso.com / brochurepdf / Huvitz / HFR-8000.pdf”, retrieved on September 30, 2019.

[0014] There is no known method or device that is able to determine the overall shape and contour of a groove in a short time, that is to say using only a few image recordings, with a target accuracy of 0.1 mm. Summary of the invention

[0015] Starting from DE 19937647 C1, it is an object of the present invention to provide a method and a device which facilitate reliable, simple and cost-effective determination of the outer shape of a spectacle frame recess.

[0016] This object is achieved by a method, a device, a computer program product and a use having the features of the respective independent claims. Preferred embodiments are subject matter of the dependent claims.

[0017] In a first aspect, the invention relates to a method for determining the outer shape of a frame groove in a bezel of a pair of spectacles. The method comprises illuminating the bezel, capturing a plurality of images of the illuminated bezel from different predetermined viewing angles, and evaluating the captured images and determining the spatial curve of the frame groove based on the evaluated images. The method is characterized in that the bezel is illuminated along the entire circumference of the bezel by directional illumination. Further, the method is characterized in that the evaluation of the captured images comprises assigning each portion contained in the captured images to a corresponding surface element of the frame groove based on at least one of the following characteristics: a shading of the corresponding portion, a brightness of the corresponding portion, and a phase angle of the illumination of the corresponding portion.

[0018] In another aspect, the invention relates to a device for contactlessly determining the outer shape of a frame groove in a bezel of a pair of spectacles. The device comprises an illumination device designed to illuminate the bezel, an image capture device configured to capture a plurality of images of the illuminated bezel from different predetermined viewing angles, and an image evaluation unit configured to evaluate the captured images and to provide the evaluated images for determining a spatial curve of the frame groove. The device is characterized in that the illumination device is designed to illuminate the bezel along the entire circumference of the bezel with directional illumination, and the evaluation of the captured images comprises assigning each portion contained in the captured image to a corresponding surface element on the frame groove based on at least one of the following characteristics: a shading of the corresponding portion, a brightness of the corresponding portion, and a phase angle of the illumination of the corresponding portion.

[0019] In another aspect, the invention relates to a computer program product comprising a non-volatile computer-readable storage medium on which an executable program code is stored, the executable program code being configured to prompt a device to perform a method according to the invention.

[0020] In another aspect, the invention relates to the use of a device for determining the vertex power or a device for creating a spatially resolved power map for determining the profile of a frame recess in a bezel of a pair of spectacles or a spectacle frame.

[0021] In another aspect, the invention relates to a method for grinding a spectacle lens for insertion into a spectacle frame, comprising the method according to the invention for determining the outer shape of a frame recess.

[0022] In another aspect, the invention relates to a computer-implemented method for determining the geometry of a spectacle lens, comprising a method according to the invention for determining the outer shape (34) of a frame recess (12c) in a bezel (12a) of a spectacle frame.

[0023] In this case, the bezel is the surface of the spectacle frame that faces the spectacle lens inserted into the corresponding bezel of the spectacle frame and is adjacent to the spectacle lens. In other words, the bezel is the surface that delimits the opening of the spectacle frame, into which the spectacle lens can be inserted. In this case, the circumference of the bezel is preferably the inner circumference of the bezel. The inner contour of the bezel visible in the image is also referred to as the bezel contour.

[0024] A frame groove, also referred to simply as a groove, is in this case a contour present in the bezel, which preferably extends perpendicularly to the bezel. For example, the frame groove may be in the form of a notch and / or a milled depression in the bezel. For example, the cross section of the frame groove may have exactly or approximately a V-shape, such information about the approximate or assumed cross-sectional shape of the frame groove being able to be taken into account or incorporated when determining the outer shape of the frame groove.

[0025] In this case, the outer shape of the frame groove is preferably a three-dimensional space curve reflecting the contour and / or size of the frame groove. For example, the space curve can be provided by three-dimensional coordinates, for example as a set of points in space, in particular as a set of space curves. Here, the space point is preferably a triple of three-dimensional coordinates defining the position of the space point in a predetermined coordinate system. The predetermined coordinate system can preferably represent a relative coordinate system, for example relative to the optical center of the image capture unit and / or relative to a position defined by a marking element.

[0026] In this case, the fact that the lens ring is illuminated along the entire circumference does not mean that the lens ring is sequentially swept by the illumination along its circumference, but rather that the entire circumference of the lens ring is captured by the illumination. In particular, this can be achieved by the illumination being incident on the lens ring from the outside. It is not necessary to illuminate the lens ring from the inside, that is to say using a light source arranged in the lens ring opening. In this case, preferably the entire circumference of the lens ring is illuminated simultaneously. The fact that the illumination may not completely penetrate the lens holder groove in some cases does not prevent illumination along the entire circumference. In this case, illuminating the lens ring simultaneously along its entire circumference means that the illumination is directed to the entire lens ring.

[0027] In this case, directional illumination is illumination incident on the eye ring from a given direction. Particularly preferably, the illumination can originate from a point light source and / or from a planar illumination from a predetermined direction. The directional illumination is selected so that it forms a shadow as precise as possible on the frame recess. For example, the directional illumination can be obtained by the spectacle frame being incident only from a direction relative to the image capture device, for example by incident light from a direction in which the image evaluation unit is also arranged relative to the spectacle frame. To this end, the illumination device can provide directional illumination by, for example, one or more point light sources, which illuminate the eye ring from a predetermined direction and form a shadow as precise as possible on the frame recess. For example, the point light source can be in the form of an incandescent lamp and / or an LED and / or a laser diode and / or a laser. Alternatively or in addition, one or more light sources can be provided, which contribute to the transmitted light illumination of the eye ring. In other words, a light source can optionally be provided for incident illumination from the opposite direction of the spectacle frame, that is to say from the side of the spectacle frame facing away from the image capture device. For this purpose, the image capture device can have a planar emitter, for example a display and / or an OLED lighting element. In particular, this can be advantageous for determining the outer shape of the frame recess in an at least partially transparent spectacle frame. In particular, surface light sources such as displays and / or OLED lighting elements can be used for this purpose.

[0028] Here, the shadow of the corresponding part is a characteristic of the corresponding part caused by the different brightness of the lighting incident on the part or by the shadow of the corresponding part caused by the lighting. Different parts that are exposed to different degrees of lighting and / or have different degrees of shadow caused by the lighting have different levels of shadow and / or different levels of brightness. Therefore, different parts can be identified in the captured image, preferably based on the shadow (in the image), and the parts are preferably assigned to surface elements (in space) of the frame groove. For example, the transition between different brightness levels and / or shadow levels can be evaluated as a transition between adjacent parts, which can then also be evaluated as a transition between adjacent surface elements. Based on these surface elements, a plurality of spatial points (in space) can then be determined, which characterize the relative position and size of the frame groove or in particular the cross section of the frame groove. For example, the cross section of the frame groove can be obtained at each position of the frame groove along the contour of the frame groove by five spatial points, and the surface elements extend between these spatial points. In simplification, the contour extending at the bottom of the groove can be characterized by a single spatial point in the profile cross section. In another simplification, the cross section of the groove profile can be characterized by only three spatial points, one of which represents the cross section of the groove bottom and the other two represent the corresponding transitions between the groove side and the bezel. According to other embodiments, more than five spatial points can also be detected in order to model the cross-sectional profile of the frame groove.

[0029] For example, in this case, a surface element is a continuous area of ​​a virtual and / or mathematical representation of the shape assigned to a portion in a captured image, which portion is identified or evaluated as a continuous portion due to uniform or approximately uniform properties in terms of shading and / or brightness and / or phase angle of transmitted and / or reflected illumination and is thereby delimited from other surface elements.

[0030] In particular, the shading is a brightness distribution in the captured image. For example, the shading may be present as a linear brightness distribution or may comprise a linear brightness distribution. Preferably, also different image filters may be used, for example further image channels may be generated and / or a segmentation and / or evaluation of the image may be performed based on the image channels thus generated. For example, the illumination device may provide a phase-shifted illumination, so that in the case of images captured using mutually phase-shifted illumination, a phase image and / or an amplitude image and / or a curvature image may also be calculated.

[0031] For example, the phase angle of the illumination of the respective part is the relative phase angle of the captured illumination relative to other parts in the captured image. By varying the phase of the illumination, for example, the spatial contour of a groove or a bezel can be detected particularly accurately. For capturing the phase angle, it can be particularly advantageous to illuminate an at least partially transparent spectacle frame using transmitted light illumination, since in this case the shadow can be determined particularly reliably.

[0032] In this case, different predetermined viewing angles from which the images are captured are advantageous for determining the exact relative positions of surface elements and / or spatial points in space, for example by triangulation. In particular, the set of spatial curves is determined based on spatial points. For example, in this case, the spatial curves extend through spatial points located at corresponding cross-sectional positions of the frame groove or through a contour along the contour of the frame groove. Therefore, the spatial curves preferably extend parallel to the contour or the frame groove, so that the set of spatial curves includes the corresponding surface elements and corresponds to the contour of the frame groove. Furthermore, it may be advantageous to capture a plurality of images in which directional illumination is incident on the lens ring from different predetermined viewing angles, as this allows a particularly reliable determination of the shadow.

[0033] In this case, the device for determining the vertex power is, for example, a vertex power measuring device or a lens meter. Such a device can be used to measure the vertex power of a spectacle lens, the optical power of the vertex of which can be determined in the form of a single value.

[0034] For example, the image capture device may include one or more sensors and / or cameras. For example, a separate sensor and / or separate camera may be provided for each lens ring. Alternatively or in addition, a sensor and / or camera capable of capturing both lens rings simultaneously may be provided.

[0035] The image evaluation unit and the geometric evaluation unit can be in the form of a computing unit and / or can be implemented in a computing unit. In this case, the image evaluation unit and the geometric evaluation unit can be assigned to a single computing unit or to separate computing units.

[0036] The invention offers the advantage that the outer shape of a frame recess in a spectacle frame can be determined optically using only a few image recordings. The invention also offers the advantage that the outer shape of a frame recess can be determined particularly accurately and reliably due to the directional lighting used, since in particular a high optical contrast can be achieved. The invention also offers the advantage that a particularly reliable automatic recognition of surface elements can be achieved by evaluating shadows in the captured images. The invention thus offers the advantage of providing a reliable, simple and sensitive option for determining the outer shape.

[0037] For the sake of completeness, attention is drawn to the fact that the invention does not require mechanical probing of the frame recess. The contour of the frame recess is thus preferably determined contactlessly. This offers the further advantage that a particularly high degree of accuracy can be achieved, since there is no risk of deformation of the spectacle frame due to the mechanical probing; this is advantageous in particular in the case of very thin bezels. Since, in contrast to, for example, mechanical tracers, no manual adjustment means are required, a particularly high degree of reliability and credibility as well as a low sensitivity to operating errors can also be achieved. In addition, the invention facilitates a particularly rapid determination of the contour, since preferably only a few image recordings are required.

[0038] Further, the invention provides the advantage that no mechanical positioning unit requiring high-precision calibration is needed. This can avoid significant adjustment costs and reduce sensitivity to errors. This can also reduce maintenance costs.

[0039] Furthermore, the invention offers the advantage of being able to determine the three-dimensional shape of the profile section or cross section of the frame groove, thus allowing this to be taken into account when grinding the lens. As a result, the spectacle lens can be adjusted to a particularly accurate match to the spectacle frame during the grinding process in order to obtain the best possible retention of the spectacle lens in the frame. In addition, determining the three-dimensional shape helps to reliably and accurately grind spectacle lenses for significantly curved frames, in which case a shape determined only in two dimensions may not be sufficient. Further, the glass design can be optimized based on the shape of the frame groove and the rim, so that the lens of a pair of finished spectacles also meets optional aesthetic requirements. This can be achieved, for example, by the lens of a pair of finished spectacles terminating as exactly as possible at the front edge of the rim.

[0040] A glass design is understood to mean the spatial embodiment or geometry of a spectacle lens which the spectacle lens exhibits upon completion. For example, the glass design may depend on the prescription power, the material of the spectacle lens and the outer shape of the spectacle frame into which the lens is to be inserted. In addition, the glass design may depend, for example, on other parameters, such as the relative position of the viewpoint and / or the expected viewing habits of the user, which parameters may then be taken into account when creating the glass design.

[0041] The invention also offers the advantage that the contour of the frame recess can be determined independently of the ambient light by using an illumination device, so that reliable and reproducible results can be obtained even under different ambient light conditions.

[0042] Furthermore, the invention offers the advantage that the outer shape of the frame groove can be detected very quickly. Particularly preferably, only a few image recordings are required to register the outer shape. Furthermore, the invention can offer the advantage that very large or long sections of the outer shape of the frame groove can be registered as correspondingly continuous sections or surface elements. For example, at least one surface element can extend over at least 10° of the inner angle of the bezel.

[0043] Furthermore, the invention preferably comprises determining a set of spatial curves corresponding to the contour of the frame recess based on the surface elements assigned in the respective images. This provides the advantage that the contour can be determined and / or reproduced particularly accurately and reliably.

[0044] Preferably, the method further comprises determining a plurality of spatial points based on the connection of the surface elements correspondingly assigned in the respective captured images before determining a set of spatial curves, the set of spatial curves being determined in such a way that each spatial curve passes through at least some of the spatial points. Thus, the spatial curves can be determined particularly accurately.

[0045] Preferably, the directional illumination comprises at least one of the following features: collimated illumination, structured illumination, transmitted light illumination, incident light illumination, ultraviolet illumination and polarized illumination. This provides the advantage of being able to provide a specifically defined illumination, which provides a plurality of parameters, which allows the captured image to be evaluated. In this case, the directional illumination is, for example, illumination incident on the lens ring from one direction. The directional illumination can preferably be provided by a point light source. In this case, the directional illumination has the effect of forming a precise shadow on the frame groove, thus having a precisely assignable shadow of these parts. For example, the lighting device can include one or more incandescent lamps, LEDs and / or laser diodes. Preferably, the lighting device includes at least one of the following lighting elements: a slide projector, a digital stripe projector, a laser line projector and a laser dot pattern projector. For example, the lighting device for directional illumination can include, for example, one or more LEDs. The position of the LED illumination is preferably positioned at most 90° outside the camera plane so as to fully scatter the light at the lens ring into the field of view of the lighting device. Optionally, the lighting device can include one or more light sources for transmitted light illumination, which are optionally arranged on the side of the lens ring facing away from the image capture device. For example, these can be in the form of planar emitters. This offers the advantage that incident light and transmitted light measurements can be combined for at least partially transparent spectacle frames.

[0046] Preferably, the evaluation of the image may include filtering of the captured image and / or implementation of other preprocessing steps, in particular in an image channel. In particular, the image channel may be an output image of a phase shifting process of a fringe image, that is, the output image is based on an input image, which is recorded by illuminating the mirror ring with structured illumination, such as fringe projection, and subsequently evaluated using a phase shifting algorithm. In such a fringe image, the basic intensity, amplitude and phase images are usually available as separate image channels. This is advantageous, for example, for further processing and / or evaluation of the image and / or simplifies and / or improves the geometric evaluation. In this context, the use of a "micro phase shifting" method is particularly advantageous; see M. Gupta and SK Nayar, "Micro Phase Shifting", 2012 IEEE Conference on Computer Vision and Pattern Recognition, Providence, RI, 2012, pp. 813-820, doi: 10.1109 / CVPR.2012.6247753.

[0047] The evaluation and / or capture of an image may preferably include the addition and / or subtraction of a plurality of images. For example, the average value of the pixel brightness of a plurality of captured images may be calculated, and one or more images composed of such an average value may be used for further image evaluation.

[0048] According to another preferred embodiment, the device can be designed to provide collimated illumination of the lens ring, for example by laser line projection. For this purpose, preferably optical features are projected onto the lens ring, which can be correlated in the images captured from various viewing angles and simplify the determination of the profile. This offers the advantage that the illumination can be provided in a particularly defined manner.

[0049] For example, the collimated illumination can be provided by a laser diode or by a laser.Alternatively or additionally, when starting from a divergent or convergent emitting light source, the collimated illumination can be obtained by lenses and / or curved mirrors.

[0050] For example, structured illumination can be obtained by projecting and / or imaging a pattern onto the lens ring. For example, a line pattern and / or a dot pattern can be projected and / or imaged onto the lens ring, thereby facilitating reliable determination of the shape. For example, the structured illumination can be provided by an illumination device, which includes a slide projector, in particular for projecting static patterns, and / or a digital stripe projector and / or a laser line projector and / or a laser dot pattern projector. For example, the image evaluation unit can be configured to determine the image position of the projected pattern features in the image. For example, the geometric evaluation unit can be configured to calculate or determine the spatial position of the corresponding feature from the image position of the corresponding or associated image feature in each case, and to determine, for example, the shape therefrom. Optionally, at least two different arrangements of structured illumination are projected onto the lens ring, wherein at least one image is captured from each of the different viewing angles in each arrangement, so that each part of the lens ring is captured at least once as an image in a different structure of the structured illumination. For example, an image in which certain parts of the lens ring are located in a dark area of ​​the projected stripe illumination can be captured first, and an image in which the same parts are located in a bright area of ​​the projected lens ring illumination can be captured second. In order to change the arrangement of the structured illumination, in particular the phase of the stripe illumination can be changed. This means that the periodically varying luminous intensity of the fringe pattern is varied at least partially perpendicularly to the direction of extension of the fringe, that is to say not parallel to the direction of extension of the fringe, so that the fringe is moved or "migrated through the image". For example, micro-phase shifting, which is known to the person skilled in the art from the prior art, can be used in this process (see, for example, M. Gupta et al. "Micro Phase Shifting", IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1-8, June 2012). Such a phase shift allows all parts of the mirror circle captured by the projected fringe illumination to be captured at least once in the dark part of the illumination and at least once in the bright part of the illumination. In this case, the phase is the phase of the at least approximately sinusoidal intensity curve of the fringe pattern.

[0051] The illumination using a pattern can preferably be combined with the image capture of an image using collimated illumination, for example so that the principle of edge detection can be reliably applied. This provides the advantage of being able to facilitate particularly reliable determination of the profile. For example, if a digital fringe projector is used for pattern projection, a series of phase-shifted fringe images can be captured, preferably using a known phase-shifted fringe projection method. From these images, an average image (so-called basic intensity) and / or a gradient image of the phase gradient and / or an image of the amplitude value can be calculated, for example, using known methods (e.g., averaging the corresponding pixels on all images in each case). Therefore, the amplitude image channel of one or more fringe images can be used for edge detection. Edge detection can be particularly advantageously applied to it because, compared to conventional gray value images, other information useful for edge detection is additionally provided. In particular, the contrast in the amplitude image is particularly detailed, as is the curvature intensity that becomes visible by deriving and evaluating the curvature information from the phase image.

[0052] For example, transmitted light illumination can be obtained by illuminating the lens ring from the side of the lens ring opposite to the image capture device. Transmitted light illumination can be advantageous, in particular in the case of partially transparent spectacle frames, wherein the illumination can at least partially penetrate the spectacle frame, in particular at least partially through the lens ring. The advantage is that a reliable determination of the profile is facilitated even if the lack of reflectivity of the spectacle frame makes determination using only incident light illumination impossible or only possible but difficult.

[0053] Incident light illumination can be achieved in particular by illuminating the lens bezel from the same side of the lens bezel on which the image capture device is also arranged. For example, the optical axes of the light source and the image capture device can lie exactly or approximately in one plane. Incident light illumination offers the advantage of being able to capture light reflected and / or scattered by the lens bezel. Incident light illumination can be particularly advantageous for illuminating an opaque lens bezel which cannot be penetrated or at least cannot be penetrated to a sufficient extent.

[0054] For example, ultraviolet lighting may be provided, for example by a UV lamp and / or by a laser. Ultraviolet lighting may be advantageous in particular for being able to reduce reflection effects, in particular on metallic surfaces.

[0055] Polarized illumination can be achieved, for example, by using a light source that emits in a polarized manner and / or by means of a polarization filter. Polarized illumination is preferably linearly polarized illumination. This offers the advantage that the reflective properties of the illumination can be adapted to the lens ring and / or can be varied. In this case, a specific polarization may also be advantageous for being able to obtain an enhancement of intensity differences or brightness differences between different parts or segments of the lens ring and / or for being able to obtain the effect of suppressing destructive reflections.

[0056] Preferably, the illumination comprises a switchable illumination of the lens ring, so that lens rings of transparent or translucent materials and lens rings of non-transparent materials, such as metals, can be recorded in such a way that the various surface segments are distinguishable in the captured image. In this case, in the case of translucent materials, preferably transmitted light illumination (so-called backlighting) is used, while in the case of opaque or non-transparent materials (such as metals), preferably incident light illumination is used, because optionally transmitted light in the case of translucent materials and reflected light in the case of opaque materials provide a higher optical contrast. The intensity of the illumination, that is to say the lighting intensity, can be particularly advantageously adapted to the degree of absorption of the material - in the case of pronounced absorption, the lighting intensity can be increased, while in the case of pronounced reflectivity or scattering, the lighting intensity can be reduced.

[0057] Preferably, the captured image is at least partially evaluated by image segmentation. This offers the advantage that a particularly reliable automatic image evaluation can be achieved. In this case, the image segmentation optionally comprises the following steps: bezel segmentation or frame segmentation, contour detection of the inner bezel, edge detection and edge completion of the groove contour by mathematical optimization methods.

[0058] For example, the segmentation of the ring or the frame can be achieved by foreground / background segmentation, for which transmitted light illumination may be particularly advantageous. In this case, an image of the black outline of the frame is preferably generated by placing the frame between the illumination and the camera and the frame protruding from the background. In the image, the pixels of the black outline of the frame are darker than the background. Therefore, the frame can be segmented by a threshold method, the result of which is a binary image. In order to further improve the contrast, for example, a difference image between the black outline image and a blank background image without the frame can be formed before thresholding and binarization. The outline of the inner ring can preferably be derived from the binary image by conventional edge extraction methods, for example by morphological operators. In order to further improve the accuracy, sub-pixel accurate edge detection can preferably be performed in the area of ​​the ring outline. For example, the sub-pixel canny method in the document "Stable Segmentation of 2D Curves", AWJ Fritzgibbon, PhD thesis, 1997, University of Edinburgh can be used.

[0059] The edge detection of the edges belonging to the groove contour can be realized in particular in the binary image of the segmented bezel or frame, and preferably the edges extending parallel or almost parallel to the outer contour line of the inner bezel can be detected. Here, the direction vectors which can be calculated, for example, by forming a gradient or determining a straight line fit are preferably assigned to the points or segments of the corresponding line. The edge segments whose direction vectors deviate more significantly from the given maximum angle are optionally removed from the further calculations.

[0060] For example, completing the edge by a mathematical optimization method may include minimization of a cost function.

[0061] In this case, it is preferred to use a model of the contour line to be determined which contains free parameters. For example, a V-shaped cross-section of the frame groove can be assumed and can be included in the optimization method. The distance from the edge to the model is preferably included as a data item in the cost function. For example, the least squares method can be used for this purpose. Preferably, a set of parameters for which the cost function is minimized is determined within the optimization loop. Optionally, other terms, such as the evaluation of the smoothness of the corresponding contour line, can be included in the cost function for regularization purposes. For example, the model can include a complex-valued Fourier series, in which the free parameters determine the coefficients of the series and the real and imaginary parts encoding the x and y values ​​of the contour points, respectively. Other available contour models can be based on, for example, polynomial splines, NURBS or conic sections.

[0062] Preferably, capturing images from different predetermined viewing angles comprises capturing images from different predetermined observation angles about the rotation axis. This offers the advantage that different viewing angles can be employed with particularly little expense and / or in a particularly reliable manner. To this end, for example, the spectacle frame can be arranged at different angles relative to the image capture device. For example, the spectacle frame can be pivoted to different positions for this purpose. Preferably, the rotation axis for pivoting the spectacle frame extends parallel to the upper edge of the spectacle frame. However, alternatively or additionally, the position of the image capture device can also be changed in order to facilitate capturing images from different viewing angles and / or observation angles. It is also possible to use multiple image capture devices, which facilitate capturing images from different predetermined observation angles or viewing angles without pivoting and / or shifting the spectacle frame and / or the image capture device.

[0063] Optionally, the spectacle frame is arranged at different angles with respect to the image capture device at different viewing angles. This provides the option of capturing the spectacle frame from different angles using the image capture device. From some different angles, then optionally parts of the spectacle frame that may not be visible from different angles and viewing angles can be viewed. Thus, viewing the spectacle frame from different viewing angles can be achieved in a simple and effective manner.

[0064] Optionally, the spectacle frame is pivotable about a rotation axis, which extends parallel to the upper edge of the spectacle frame. This allows the frame recess to be visible from different angles or viewing angles even if the image capture device is arranged outside the spectacle bezel. In other words, it is not necessary to position the image capture device inside the bezel, that is to say in the area provided for the insertion of the spectacle lens. Instead, it is sufficient to observe the bezel from the outside using the image capture device. This can reduce the complexity of the device for contactless determination of the outer shape of the frame recess in the bezel, and furthermore, the image capture device can capture a larger area of ​​the bezel or the frame recess at the same time. Optionally, the entire bezel can be captured at the same time in this case.

[0065] Optionally, images are captured from different viewing angles based on the changed position of the image capture device. In other words, for the purpose of capturing images from different viewing angles, the position and / or orientation of the image capture device relative to the spectacle frame is optionally changed, and / or the position of the spectacle frame relative to the image capture device is optionally changed. This facilitates simple implementation and technical realisation of capturing images from different viewing angles.

[0066] Optionally, multiple image capture devices are used to capture images from different predetermined viewing angles. This provides the option of capturing images of the rim from multiple viewing angles simultaneously. Optionally, this avoids the requirement to cause a change in the relative position and / or orientation of the eyeglass frame with respect to the image capture device. This can reduce the amount of time required to determine the profile of the frame recess.

[0067] Optionally, a plurality of images are captured in such a way that the image capture device captures both eye circles of the spectacle frame in the image area. This can save time, since it is not mandatory to capture both eye circles in two separate measurements. Likewise, image evaluation of both eye circles can be achieved in one process, without having to evaluate a plurality of different images of the respective eye circles.

[0068] Preferably, a marking element is further provided, which is arranged in a unique predetermined position in such a way that a unique positioning of the eye ring relative to the marking element can be determined by the captured image. This provides the option of determining a unique position of the eyeglass frame and thus the outer shape of the frame recess relative to the image capture device by means of the marking element. To this end, the marking element preferably has a precisely defined position relative to the image capture device. Thus, the relative position of the eyeglass frame relative to the image capture device can be determined based on the relative position of the eyeglass frame relative to the marking element.

[0069] Preferably, the marking element has at least one distinctive feature which can be located very accurately in the captured image from the image capture device and, if the image capture device comprises a plurality of cameras, in the images of all the cameras involved and has a fixed spatial relationship with the eye ring. This provides the advantage of facilitating simple orientation and / or determination of the position in the captured image. In particular, the image position of this marking element is determined in order to determine the relative position and orientation of the eye ring with respect to the camera involved. This contributes to the realization of the tilting device in a cost-effective manner, since a repeatable and accurate positioning of the eyeglass frame in the device can be omitted. Further, this contributes to the use of cost-effective mechanical components, such as cost-effective holding devices. The marking element can preferably have a partially transparent design to allow one or more distinctive features to be determined both in incident light illumination and in transmitted light illumination. In this case, a distinctive feature is a feature whose position can be uniquely determined very reliably and very accurately in the captured image.

[0070] Preferably, the device comprises a holding device, which is designed to hold the pair of spectacles in both a fixed position and a fixed orientation relative to the lighting device and relative to the image capture device. The image capture device, which can optionally be rearranged in a plurality of different positions relative to the rim, is not an obstruction, since the fixed position and fixed orientation positioning is associated with a corresponding fixed arrangement of the image capture device in one of the various viewing angles. Preferably, the holding device is also designed to hold the pair of spectacles in at least two different orientations relative to the lighting device and relative to the image capture device, the at least two different orientations defining different predetermined viewing angles of the image capture device relative to the rim. Particularly preferably, the image capture device is designed to rotate or pivot the spectacle frame about a rotation axis in order to allow the image capture device to capture images of the illuminated rim from at least two different predetermined viewing angles, particularly preferably from a plurality of different predetermined viewing angles.

[0071] Preferably, the holding device comprises a spectacles holder with a spring element for stably holding the pair of spectacles in different pivot positions. This facilitates easy fastening and removal of the spectacles and securely holds the pair of spectacles while determining the outer shape of the frame recess.

[0072] Preferably, the device according to the invention is further designed to determine at least one of the following two properties of the spectacle lenses of a pair of spectacles held in the device: the vertex power and the spatially resolved power map. In other words, the device is preferably designed as a vertex power measuring device and / or as a device for measuring a spatially resolved power map and can be used, among other functions, to determine the profile of one frame recess or to determine the profile of two frame recesses. This offers the advantage that the same equipment can be used for a variety of purposes, and thus the amount of equipment required for adjusting and / or completing a pair of spectacles can be reduced. In particular, this offers the advantage that, particularly preferably, the same illumination device, the same image evaluation unit and the same holding device can be used for different purposes.

[0073] Preferably, the contours of two frame recesses, that is to say the respective frame recess in each frame ring, can be determined within the measuring range. Particularly preferably, both contours can be detected simultaneously. This contributes to a particularly rapid determination of the contours. For example, the image evaluation unit can be configured to capture both frame rings in each image recording. For example, to this end, the image capture device can be designed to capture both frame rings in the image area. Alternatively or in addition, the image capture device can include a plurality of sensors and / or cameras in order to capture each frame ring with a separate sensor and / or a separate camera. Preferably, the image capture device includes at least two cameras. Particularly preferably, each of the two cameras is arranged such that the optical axis of the respective camera extends through one of the two frame rings or contours or through an opening in a pair of spectacles spanned by the respective frame ring.

[0074] Preferably, the image capture device comprises at least one "high dynamic range" (HDR) camera. Alternatively or additionally, the image capture comprises recording a plurality of images of each arrangement or pivot position of the spectacle frame using in each case a range of lighting conditions (e.g. different lighting intensities), such that both the transition from the rim to the groove and the groove bottom are clearly visible in at least one of the images having a suitable lighting condition from the range of lighting conditions.

[0075] Optionally, the image capture device comprises at least one camera, wherein the optical axis of the at least one camera extends through one of the two bezels or profiles or through an opening of a pair of spectacles spanned by the respective bezel. In this case, the at least one camera is directed toward the opening of the at least one bezel. Thus, the bezel can be observed from the outside and, therefore, there is no need to position the camera inside the bezel, that is, in the area provided for inserting a spectacle lens.

[0076] Preferably, the distances between the space curves in the set of space curves are interpolated, so that the shape can still be fully determined, for example, in the case where the shape in the image is partially hidden. This is particularly advantageous in the case where the tilting or pivoting of the spectacle frame around the horizontal axis is limited, because in this case the shape at the left and right rims can be hidden in all rotational positions. In the case where the rotation around the vertical axis is limited, this preferably applies to the upper / lower rims.

[0077] Preferably, within the optimization loop there is an iterative application of the image evaluation unit and the geometric evaluation unit, the result of the geometric evaluation unit being used in each case in order to be included in the image evaluation unit as a further input. In other words, the steps of evaluating the captured image, determining a plurality of spatial points and determining the set of spatial curves are repeatedly performed within an iterative loop. The advantage of the image evaluation unit optionally extended in this way is that the detection result is made more precise - this allows the result of the edge detection to be improved, in particular in areas with low contrast.

[0078] Preferably, the geometry evaluation unit is further configured to determine a plurality of spatial points based on the connection of correspondingly assigned surface elements in the respective captured images before determining the set of space curves, and to determine the set of space curves in such a way that the space curves each pass through at least some of the spatial points. This offers the advantage of allowing a particularly precise determination of the space curves.

[0079] It is to be understood that not only the features and embodiments mentioned above and the features and embodiments explained below are to be regarded as disclosed in the respectively explicitly mentioned combination, but also other technically advantageous combinations and embodiments are comprised in the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Further details and advantages of the invention shall now be explained in more detail based on the following examples and preferred embodiments with reference to the attached drawings,

[0081] In the attached picture:

[0082] FIG1 shows a device 10 for contactlessly determining the contour of a frame groove;

[0083] Figure 2 A holding device 28 according to a preferred embodiment is shown;

[0084] FIG. 3A to FIG. 3C Three different perspective views of the holding device 28 of the device 10 according to the preferred embodiment are shown;

[0085] Figure 4A and Figure 4B shows an image of an eyeglass frame captured in an exemplary manner;

[0086] Figure 5 A schematic diagram showing the outer shape of the frame groove 12c in the lens ring determined or reconstructed in an exemplary manner;

[0087] Figure 6 shows exemplary representations of capturing images from different viewing angles; and

[0088] FIG. 7A to FIG. 7C An exemplary explanation related to shape modeling is shown. DETAILED DESCRIPTION

[0089] Figure 1A and Figure 1B The side view ( Figure 1A ) and floor plan ( Figure 1B ) shows a schematic diagram of a device 10 for contactlessly determining the outer shape of a frame groove in a ring of an eyeglass frame 12 according to a preferred embodiment.

[0090] The device 10 comprises an illumination device 14, which is in the form of a display according to the illustrated embodiment. For example, the display may comprise a commercially available computer display or may be designed as such. The illumination device 14 is designed to provide directional illumination to the rims of the spectacle frame 12 as transmitted light illumination along the entire circumference of the rims 12a. In this case, the spectacle frame 12 comprises two rims 12a and two temples 12b. According to other embodiments, the device 10 may comprise other illumination devices and / or light sources, which are designed to provide incident light illumination, so that light is incident on the spectacle frame 12 from the side of the spectacle frame 12 facing the image capture device.

[0091] Further, the device 10 comprises an image capture device 16, by means of which a plurality of images of the illuminated rim 12a can be captured. In this case, the spectacle frame 12 is arranged between the image evaluation unit 16 and the illumination device 14, so that the directional illumination provided by the illumination device 14, as transmitted light illumination, at least partially passes through the rim 12a or rims 12a of the spectacle frame 12 and is captured by the image capture device 16. According to the embodiment shown, the image capture device 16 comprises two optical cameras 18, which are arranged in such a way that their optical axes extend through the opening spanned by the rim 12a of the spectacle frame 12. In this case, the cameras 18 are arranged in such a way that the rim 12a of the spectacle frame 12 is arranged within the focal length range 100 of the cameras 18, so that the image capture device 16 can capture a sharp or focused image representation of the rim 12a. In this case, the two cameras 18 are arranged next to each other in such a way that their lateral spacing corresponds approximately to the distance between the two centers of the opening spanned by the rim 12a. However, other arrangements of the camera 18 that facilitate imaging the lens ring 12a are possible.

[0092] According to other embodiments, the image capture device may also include only one camera or sensor, or more than two cameras or sensors, by means of which an image of the rim 12a may be captured. Alternatively, according to other embodiments, the device 10 may be designed to determine the contour of only one of the frame grooves of the two rims 12a of the spectacle frame 12. Using such a device, the contours of the two grooves of the rim 12a may then be captured one after the other.

[0093] Further, the image capture device 16 is pivotable in such a way that the image capture device 16 can capture images of the eyeglass frame 12a from different predetermined viewing angles. Alternatively or in addition, the device 10 may include a suitable holding device by which the eyeglass frame 12a can be arranged in different orientations so as to include images from different predetermined viewing angles using the image capture device 16.

[0094] According to the embodiment shown, the device 10 further comprises a computing unit 20, in which an image evaluation unit 22 and a geometry evaluation unit 24 are integrated or implemented. According to another embodiment, the image evaluation unit 22 and the geometry evaluation unit 24 can also be formed in a separate computing unit. In this case, the computing unit has a data communication link with the image capture device in order to receive and evaluate the captured images.

[0095] The image evaluation unit 22 evaluates the image captured and provided by the image capture device 16 and assigns each portion contained therein to a corresponding surface element of the frame recess. The assignment is performed based on the shading and / or brightness of the corresponding portion.

[0096] The geometry evaluation unit 24 then determines a plurality of spatial points in space, that is to say a plurality of spatial points in the object space of the spectacle frame 12, based on the connection of the surface elements in the images captured from different viewing angles, and determines a set of spatial curves. This connection can preferably be realized by triangulation. Each spatial curve extends through at least some of the spatial points. In this case, the set of spatial curves corresponds to the contour of the frame groove.

[0097] Furthermore, the device 10 comprises a housing 26 for enclosing the other components of the device 10. Preferably, the housing 26 is opaque and is designed so that the components of the device 10 arranged within the housing 26 are shielded from the influence of ambient light. This provides the advantage that disruptive background lighting can be avoided when determining the shape, and thus allows to improve the accuracy and / or reliability of the determination. In addition, the shape can be determined independently of the brightness and / or lighting of the surrounding environment.

[0098] Preferably, the device 10 according to the illustrated embodiment is a device 10 that is also designed to determine the vertex power and / or create a spatially resolved power map. In other words, the device 10 can be designed to determine both the vertex power and / or the spatially resolved power map and also the profile of the recess in the bezel 12a. This provides the advantage that, for example, the optician can perform multiple tasks or functions using a single piece of equipment.

[0099] The spectacle frame or the spectacle frame without lenses 12 can be arranged in the device 10 for determining the outer shape of the spectacle ring. In other measurements, the spectacle frame 12, this time including spectacle lenses, can again be arranged in the device to determine the vertex power and / or the spatially resolved optical power map.

[0100] Preferably, the camera is in the form of a high-resolution camera with at least 10 megapixels. The device is preferably configured to measure the optical power of two spectacle lenses inserted into the pair of spectacles simultaneously over the entire visual surface of the spectacle frame 12. The cameras are preferably precisely calibrated to each other, for example with an accuracy of 1 / 10 mm or more, and are pointed at a holding device for holding the spectacle frame 12. In this case, the display of the lighting device 14 can preferably be configured to display an optical pattern so that one or more images of the pattern can be captured through the spectacle lenses by the image capture device 16 or by the camera. Preferably, the camera is adjusted in such a way that the relative focus position is accurately set on the spectacle lenses of the spectacle frame 12.

[0101] Preferably, the eyeglass frame 12 can be held in position very precisely mechanically by a holding device and can be shifted and / or pivoted or turned so that the image capture device can capture images of the illuminated rim from at least two predetermined viewing angles, thus allowing the outer shape 34 of the recess 12c in the rim 12a to be determined.

[0102] Preferably, the device facilitates at least one of the following three specified mechanical movements of the spectacle frame or the spectacle frame 12 relative to the camera or relative to the image capture device 16, for example by means of the holding device 28:

[0103] A) Translation of the eyeglass frame relative to the camera (e.g. horizontal or vertical)

[0104] B) The eyeglass frame is rotated past the camera (e.g., with respect to a horizontally extending axis of rotation)

[0105] C) Combination of rotation and translation

[0106] In this case, it is conceivable that the camera moves relative to the frame and / or the frame moves relative to the camera. In the case of rotation (options B and / or C), the axis of rotation can be freely positioned in space - that is, it does not necessarily have to be parallel to any other line or surface.

[0107] According to a particularly preferred embodiment, the spectacle frame including its rim 12a is pivoted about a horizontal rotation axis 200 arranged parallel to a connecting line perpendicular to the optical axes of the two cameras. This helps to keep the rim 12a both within the focal length and within the vertically and horizontally limited field of view of the cameras, even when rotated.

[0108] A plurality of images of the lens ring are captured from a plurality of viewing angles or viewing angles. Particularly preferably, the device and the method are configured in such a way that the lens ring is not shadowed in different positions or arrangements in the field of view of the camera. In addition, a very precise mechanical fixing device (see, for example, Figure 2 ) is advantageous, by means of which a change in the position and / or orientation of the spectacle frame can be controlled, so that the position and / or orientation of the spectacle frame is controllable. In particular, the device is preferably embodied in such a way that the spectacle frame is fixed by the holding device in such a way that the position and / or orientation of the spectacle frame 12 or the spectacle frame does not change, even in the case of a change in the position and / or orientation relative to the holding device.

[0109] Figure 2 A schematic diagram of a holding device 28 according to a preferred embodiment is shown, by means of which a spectacle frame 12 or a spectacle frame can be held or fixed in the device 10 .

[0110] The holding device 28 comprises three holding arms 30, by means of which the spectacle frame 12 can be held or fixed in a predetermined position and / or orientation relative to the image capture device 16. According to the illustrated embodiment, one holding arm 30 fixes the nose bridge of the spectacle frame 12, while the other two holding arms 30 each fix one temple 12 b. In this case, the spectacle frame 12 can also be simply placed on the holding arms 30. However, the holding arms 30 are preferably designed in such a way that they firmly fix the pair of spectacles and stabilize them against sliding.

[0111] For example, the holding arms 30 may be variable in their height or length so that the spectacle frame 12 can be pivoted or rotated at least partially around the rotation axis 200 within a specified angular range. For example, the length of the rear holding arm 30 can be extended and / or the front holding arm 30 can be shortened so that the spectacle frame 12 is tilted forward or vice versa. To this end, the holding arms 30 may, for example, be in the form of telescopic arms. Alternatively or in addition, the entire holding device 28 may, for example, be designed to be rotatable or pivotable or tiltable in order to facilitate the pivoting of the pair of spectacles 12 around the rotation axis 200.

[0112] In addition, according to other embodiments, the rotation axis 200 can be arranged at different positions and / or extend in different directions. In this case, the holding device 28 is then preferably configured to pivot the spectacle frame 12 around the corresponding rotation axis 200.

[0113] Furthermore, the holding device 28 according to the embodiment shown comprises a marking element 32 which is fixed in its position and orientation relative to the holding device 28. In addition, the device 10 is preferably configured and calibrated in such a way that the position and orientation of the marking element relative to the image capture device is precisely known (at each viewing angle) so that the exact position of the outer shape of the rim 12a in the image captured by the image capture device 16 can be determined as accurately as possible, even if the relative position and / or orientation of the spectacle frame 12 relative to the image capture device 16 is not initially precisely determined. In other words, the marking element 32 can preferably be used to provide a reference point for determining the exact position and / or orientation of the outer shape of the spectacle frame 12.

[0114] FIG. 3A to FIG. 3C Three different perspective views of the holding device 28 of the device 10 according to the preferred embodiment are shown.

[0115] In this case, the spectacle frame 12 is fastened in the holding device 28 in such a way that the three holding arms 30 of the holding device 28 hold the spectacle frame 12. According to the embodiment shown, the two rear holding arms 30 that hold the temples 12b of the pair of spectacles are of rigid form. In contrast, the front holding arms 30 are of a movable design and allow the rims 12a to be raised and / or lowered to facilitate capturing images of the rims 12a from different viewing angles. Accordingly, the main clamp for fixing the spectacle frame 12 is implemented in a clamping manner on the nosepiece of the spectacle frame, and this allows the spectacle frame to be fixed even if the center of gravity of the spectacle frame 12 changes. Further, if necessary, the two temples 12b of the spectacle frame are clamped so that the flexible temples, for example in the case of Titanflex spectacles, do not cover the field of view of the image capture device 16.

[0116] In the present exemplary embodiment, the holding device 28 holds the spectacle frame in a horizontal relative position in front of the camera 18 of the image capture device 16. From this relative position, the spectacle frame can be rotated, for example by a stepper motor, about a horizontal axis of rotation by one or more predetermined angles, upwards and downwards by precisely defined angles. During this time, the image capture device records the spectacle frame and its bezel for each discretely adopted position.

[0117] In order to improve the reconstruction of the mirror ring geometry, the algorithm can preferably be assisted by various illumination settings of the mirror ring. For example, illumination including fringe projection can be used.

[0118] The marking element 32 is designed as an integral component of the front holding arm 30 , so that a fixed position and orientation of the marking element 32 relative to the holding device 28 is achieved in a simple manner.

[0119] Figure 3C Furthermore, an illumination device 14 is shown, which provides illumination for the contactless determination of the outer shape of the frame recess 12 c and for the determination of the vertex power and / or the spatially resolved optical power map.

[0120] Figure 4A and Figure 4B Captured images of a spectacle frame are shown in an exemplary manner, which images show the bezel 12a of the spectacle frame 12. In the process, the bezel 12a is illuminated using transmitted light illumination, so that the frame groove 12c in the corresponding bezel 12a clearly appears in the image. Figure 5B shows two enlarged views of a cross section of a lens ring 12a, which more clearly show the frame groove 12c. In particular, because the frame groove 12c produces a shadow under lighting due to the depression, it can be recognized here that the frame groove 12c can be identified based on the shadows of different parts. Therefore, the shadows in the captured image can be identified and detected as parts by different levels of shadows or different levels of brightness. Therefore, parts with the same shadow can preferably be assigned to surface elements very accurately.

[0121] The images captured by the image capture device from different predetermined viewing angles can be respectively as follows: Figure 4A and Figure 5 B, and is then provided to an image evaluation unit. The evaluation of the captured image is explained in the following in an exemplary manner based on a preferred embodiment.

[0122] Image evaluation based on known image segmentation methods is designed to detect image regions belonging to individual parts or surface segments of the eyeglass bezel or eyeglass frame groove. Suitable methods are known in the prior art and are described, for example, in the online encyclopedia Wikipedia under the term "image segmentation".

[0123] For example, the following program can be used in this context:

[0124] - lens ring segmentation, in particular with the aid of images recorded using transmitted light and foreground / background segmentation;

[0125] - Edge detection within the segmented mirror ring, preferably an edge extending almost parallel to the mirror ring;

[0126] - The edges are completed with the help of an optimization method within the scope of which a cost function is minimized.

[0127] As a result of the image evaluation, there may be, for example, a set of surface elements F1 to F4 and / or space curves K1, K2, K3, K4 and K5 and, for each image in the image stack, a point list of space points or a pixel chain of a polygon. These are transmitted, for example, to a geometry evaluation unit.

[0128] For example, edge detection can be performed with the aid of a suitably trained neural network. In this process, a human expert annotates, for example, the bezels and edges of a plurality of frames that are as different as possible. Subsequently, a neural network is trained based on these data, and the bezels and edges in the image that belong to the spatial curves K1 to K5 are determined in the image, for example, in the case of unknown frames. In this case, for example, a two-stage method is advantageous, in which, in each case, the bezels are first determined in the entire image and a set of image windows are determined along the bezels based on the bezels. The use of image windows of fixed size is advantageous. The edges belonging to the spatial curves K2, K3, K4 in the image are thus determined in each case within an image window.

[0129] Other machine learning methods can also be used instead of neural networks, such as principal component analysis (PCA) and support vector machines (SVM).

[0130] Figure 5 A schematic diagram of a determined or reconstructed profile of a frame groove 12c in a bezel is shown. A V-shaped profile section of the profile is assumed in an exemplary manner and used as a basis for determining the shape edge profile in order to determine or reconstruct the profile 34. In cross section, the profile is defined or determined by spatial points P1 to P5. In this case, spatial points P1, P2, P4 and P5 characterize the bezel, while P3 determines the bottom of the notch or depression of the frame groove 12c. Therefore, the "V-shape" of the groove is particularly spanned by the spatial points P2, P3 and P4 depicted in an exemplary manner. Although the exemplary spatial points P1 to P5 are depicted only for one cross-sectional position along the profile, each cross-sectional position naturally has a corresponding spatial point. In this case, the relative positions of the spatial points P1 to P5 are determined by connecting corresponding points in images captured from different predetermined viewing angles. This can be achieved by triangulation.

[0131] Further, the outer shape is determined by space curves K1 to K5, which extend through corresponding space points P1 to P5 and through all other space points arranged at corresponding positions of the cross section. In other words, the space curve K1 extends through the space point P1 and through all points corresponding to the point P1 at other positions along the lens ring in the cross section of the lens ring. Corresponding descriptions apply to the space curves K2 to K5 and the space points P2 to P5. In addition, according to the preferred embodiment, surface segments F1 to F4 extending between corresponding adjacent space curves are modeled. For example, the surface segment F1 extends between the space curves K1 and K2, etc.

[0132] Based on the space curves K1 to K5 , in particular based on the surface segments F1 to F4 , the outer shape 34 of the frame recess 12 c can then be reproduced or determined and made available as information for adapting the corresponding spectacle lens, for example for grinding.

[0133] Figure 6 Schematic diagram showing capturing multiple images of the illuminated mirror ring 12a from different predetermined viewing angles according to a preferred embodiment.

[0134] In this case, the image capture device 16 is directed toward the spectacle frame 12 or the rim 12a from two different viewing angles, the difference between the two viewing angles being that the image capture device 16 is directed toward the spectacle frame 12 from different positions or orientations. In this case, the different positions or orientations or viewing angles correspond to different angular positions about the rotation axis 200, the different angular positions comprising the angle α.

[0135] Despite Figure 6 1 , two image capture devices 16 are depicted, but this may be a single image capture device 16 which is arranged at different positions relative to the spectacle frame 12 in both cases. The different positions of the image capture device 16 relative to the spectacle frame 12 can be achieved by the movability of the spectacle frame 12 and / or the image capture device 16. Here, the optical axes of the image capture devices 16 are respectively depicted as dashed lines. The image capture device 16 preferably comprises two cameras which are pointed at the respective lens rings 12a of the spectacle frame of the spectacle frame 12.

[0136] The following is based on Figure 6 An exemplary method of the functioning of a geometry evaluation unit for calibration purposes and for determining the profile of a frame recess of a spectacle frame 12 is explained.

[0137] In a first step there is an intrinsic calibration, in which the image capture device 16 is specifically calibrated so that the assignment of camera pixels of the cameras of the image capture device 16 to lines of sight is known in the coordinate system of the respective camera. Exemplary lines of sight are denoted by S1 and S2.

[0138] Furthermore, the position and orientation of the rotation axis 200 are calibrated relative to the cameras of the image capture device 16 which are located at different positions and / orientations about the rotation axis 200. This can be achieved within the scope of a calibration step with the aid of a calibration object having a known geometry, for example, with the aid of a plate with a chessboard pattern printed on it, or with the aid of a calibration object with a mirror frame (see, for example, Figure 2 ) is realized by using marker elements 32 with a fixed spatial relationship, whose spatial position and orientation relative to the image capture device 16 can be determined with the help of a single recorded image. In summary, after this step, the position of the optical centers Z1 and Z2 and the axis A in space is known. After this step, it is known to assign the camera pixels of the two images, that is to say the images from two different viewing angles, to the lines of sight in space.

[0139] Furthermore, in order to determine the connection of the camera pixels of the image from the first viewing angle to the camera sight plane of the image from the second viewing angle, an epipolar surface is determined. This method is known in the prior art as epipolar surface geometry and is explained, for example, in the Wikipedia online encyclopedia. The entire epipolar surface group always extends through the two optical centers Z1 and Z2. The connection of the pixels of the first image to the pixel groups on the second image is determined (usually, these pixel groups form epipolar curves; if additional image distortion correction is used within the intrinsic calibration, they form epipolar lines; if additional image correction is used, they are located on the image columns of the second camera image).

[0140] Furthermore, in the first camera image, ie the camera image from the first viewing angle, pixels Q1 are selected which belong to any sought spatial point P on the circle on the spatial curve Kj. The polygon chain assigned to the spatial curve provided by the image evaluation unit is used as a basis for selecting the pixels Q1.

[0141] Then, the epipolar surface and a set of points M2 of the pixels of the second camera image, that is to say the image from the second viewing angle, are determined, which are respectively assigned to the pixel Q1 or connected to it, so that in particular the epipolar surface contains the line of sight S1. In particular, the epipolar surface contains the spatial point P sought.

[0142] Subsequently, the pixel Q2 is determined from the point group M2 of the polygon chain based on the space curve Kj in the second image. In this process, for example, intersection points can be determined where the kernel curve or kernel surface of M2 intersects the polygon chain. If image rectification is used, the representation of the polygon chain can already be selected based on the image columns, so the relevant pixels can be read directly.

[0143] Furthermore, there is the triangulation of the sights S1 and S2. Knowing Q1 and Q2, with the aid of the calibration information from the intrinsic calibration, the sight S1 can be determined from Q1, and the sight S2 can be determined from Q2. In this case, the point in space with the smallest distance to S1 and S2 forms an approximation to the point P sought.

[0144] The described method can then be carried out iteratively for all spatial points of the space curve in order to determine the entire shape.

[0145] The epipolar surface is efficiently determined by using image distortion correction within an intrinsic calibration range and using image rectification within a range for determining the epipolar surface, so that the epipolar surfaces of pixels assigned to an image column in a first image are identical to each other, and conversely, the epipolar surfaces of camera pixels assigned to an image column in a second image are also identical.

[0146] If more than two predetermined viewing angles are used, construction of epipolar images is preferably recommended for more efficient computation.

[0147] For all spatial points P of the spatial curve, a triangulation of the line of sight is performed together, the minimum distance criterion being included as an optimization objective in the form of a cost function term in an iterative optimization. Additional terms can preferably be included in the cost function, such as a smoothness condition for the spatial curve. Since the grooves are formed within the scope of the manufacturing process, certain assumptions can be made, for example, with regard to smoothness. In this case, the tangent vectors at the spatial curve can be numerically calculated and a measure of the difference between adjacent tangent vectors can be incorporated in the cost function. Statistical measurements are also possible, for example with the aid of a principal component analysis of a plurality of known groove shapes in eyeglass frames.

[0148] In order to estimate the accuracy of the relative position of the profile curve of the groove profile, according to the embodiment explained by way of example, the following data are used:

[0149] Camera image size [pixels]: 3840 × 2748 (10.6 MP)

[0150] Field of view [mm]: 150 × 107.3

[0151] Object side pixel size [mm]: 0.039

[0152] Due to defocus and the optionally occurring low contrast of outlines produced by differences in shading, sub-pixel accurate detection is not estimated for the accuracy of edge detection. In this case, the object side pixel size is estimated as a measure of inaccuracy.

[0153] In contrast, in the case of contour lines with high contrast, sub-pixel accurate detection techniques can be used to achieve higher accuracy. In industrial metrology, sub-pixel accuracies of approximately 1 / 5 to 1 / 20 pixel can be estimated (see, for example, ZEISS O-SELECT: object pixel size approximately 35 µm; specified measurement accuracy 4.5 µm; see “https: / / www.zeiss.de / messtechnik / produkte / systeme / optische-messtechnik / o-select.html” and “https: / / applications.zeiss.com / C1257A26006EFF9E / 0 / A6DDC43030E4AAFBC1257F0100378040 / $FILE / O-SELECT_DE_60_020_0003I.pdf”), retrieved on September 30, 2019.

[0154] A shape determination technique with sub-pixel accuracy is the sub-pixel Canny edge filter [see: Gioi, Gregory Randall, Rafael Grompone: "Sub-pixel Edge Detector: An Implementation of the Canny / Devernay Algorithm", 2017, "https: / / www.ipol.im / pub / art / 2017 / 216 / article_lr.pdf"]. For other techniques, see Wikipedia: https: / / en.wikipedia.Org / wiki / Edge_detection#Subpixel , retrieved September 30, 2019.

[0155] Additionally, the three-dimensional error is estimated during the triangulation process.

[0156] When the frame is rotated through an angle of ß = 30 degrees in order to determine the spatial points of the profile curve by triangulation, the following relationship is obtained to determine the error Δz in the direction of the camera optical axis, denoted below as "z direction":

[0157]

[0158] or

[0159]

[0160] In the case of a lateral error in the profile inspection at ß = 30 degrees and Δx = 40 μm, a deviation in the z direction of Δz = 1.73 • Δx = 69 μm is therefore obtained.

[0161] Furthermore, in practice, the contour or profile of the groove may deviate from the V-shape assumed in an exemplary manner to such an extent that the tip of the V-shape has a radius of curvature.

[0162] A CT scan of a metal frame is used below as an example. Fig. 7A In an exemplary manner, a cross section of a bezel of a spectacle frame captured by CT scanning is shown, in which case the tip or bottom of the frame groove has a radius of curvature r of 0.15 mm. The V-angle has an angle of approximately 80°. The measurement of the length of the distance 300 represents the distance between the two upper ends of the "V", which yields 0.9 mm.

[0163] When the groove bottom is observed from two different predetermined viewing angles, for example along the lines S1 and S2, the center of the gray value transition or brightness transition is the same as the center of the groove bottom in this model (see Figure 7B ). Similar remarks apply to all gray value transitions or brightness transitions used in the measurement. Overall, this results in a polygonal chain model of the outer shape of the frame groove, with the nodes 500 of the polygonal chain being located at the center of the arc 600 of the frame groove 12c, as shown in FIG. Figure 7C As shown in the exemplary manner.

[0164] List of Reference Numerals

[0165] 10. Equipment

[0166] 12. Glasses frames

[0167] 12a The rims of glasses

[0168] 12b The temples of glasses

[0169] 12c Frame groove

[0170] 14 lighting fixtures

[0171] 16 Image Capture Device

[0172] 18 Camera

[0173] 20 Computational Units

[0174] 22 Image evaluation unit

[0175] 24 Geometry Evaluation Device

[0176] 26 Housing

[0177] 28 Keep equipment

[0178] 30 Holding arm

[0179] 32 Marking components

[0180] 34 Appearance

[0181] 100 Focus area of ​​the image capture device

[0182] 200 Rotation axis

[0183] 300 "V" connection route between the upper end

[0184] 400 Gray value transition position

[0185] 500 nodes

[0186] 600 arc

[0187] S1, S2 sight

[0188] Z1, Z2 Optical center of image capture device 16

[0189] Q1 Pixels in the first camera image

[0190] Q2 Pixels in the second camera image

[0191] P is a spatial point on the mirror circle.

Claims

1. A method for determining the outer shape (34) of a frame recess (12c) in a rim (12a) of an eyeglass frame (12), the method comprising the following steps: - illuminating the lens ring (12a); - capturing a plurality of images of the illuminated lens ring (12a) from different predetermined viewing angles by means of an image capture device; - evaluating the captured image by means of an image evaluation unit and determining a spatial curve (K1, . . . , K5) characterizing the frame groove (12c) based on the evaluated image; and - determining the outer shape (34) of the frame groove (12c) based on these space curves (K1, . . . , K5); It is characterized in that - the lens ring (12a) is illuminated simultaneously along the entire circumference of the lens ring (12a) by directional lighting; - the evaluation of the captured image comprises identifying parts contained in the captured image and assigning the parts contained in the captured image to surface elements (F1, . . . , F4) of the frame groove (12c) respectively based on the brightness and / or shading of the corresponding parts, wherein identifying the parts comprises: -Mirror circle segmentation; - detecting edges bounded by these sections and surface elements (F1, ..., F4) within the segmented mirror ring (12a); and - completing these edges with the help of a first optimization method within the scope of which a cost function is minimized; and - Determining these space curves (K1, ..., K5) comprises: - based on the surface elements (F1, ..., F4), determining the spatial points (P1, ..., P5) which characterize the frame groove (12c) and through which the spatial curves (K1, ..., K5) extend, wherein the spatial points (P1, ..., P5) are located on the edge bounded by the portions and the surface elements (F1, ..., F4); and -The relative positions of these spatial points (P1, ..., P5) in space are determined by triangulation, wherein the triangulation is performed jointly for all spatial points (P1, ..., P5) of the spatial curves of the multiple spatial curves (K1, ..., K5), and a minimum distance criterion as an optimization objective in the form of a term of a cost function is included in the iterative optimization within the scope of the second optimization method.

2. The method of claim 1, wherein: Completion of these edges consists in performing triangulation of sight lines jointly for all spatial points of the spatial curve of the contour (34).

3. The method according to claim 1 or 2, wherein: The bezel (12a) is illuminated so that illumination is incident on the bezel from the outside.

4. The method according to claim 1 or 2, wherein: Directional illumination is incident on the eyeglass frame from only one direction relative to the image capture device.

5. The method according to claim 1 or 2, wherein: Directional illumination impinges on the spectacle frame as incident light from a direction in which the image evaluation unit is arranged relative to the spectacle frame.

6. The method according to claim 1 or 2, wherein: The directional lighting is provided by a lighting device through one or more point light sources, which illuminates the lens ring from a predetermined direction and produces a shadow that is as accurate as possible on the frame groove.

7. The method according to claim 1 or 2, wherein: The directional lighting includes at least one of the following features: Collimated, transmitted, incident, UV and polarized illumination.

8. The method according to claim 1 or 2, wherein: The plurality of images are captured in such a way that the image capture device captures two rings of the eyeglass frame in the image area.

9. The method of claim 1, wherein: These edges are edges extending almost parallel to the lens ring (12a).

10. A device (10) for contactlessly determining the outer shape (34) of a frame recess in a ring (12a) of an eyeglass frame (12), comprising: - an illumination device (14) designed to illuminate the lens ring (12a); - an image capture device (16) configured to capture a plurality of images of the illuminated lens ring (12a) from different predetermined viewing angles; - an image evaluation unit (24) configured to evaluate the captured image, to determine spatial curves (K1, ..., K5) characterizing the frame recess (12c) based on the evaluated image, and to determine an outer shape (34) of the frame recess (12c) based on these spatial curves (K1, ..., K5); It is characterized in that The lighting device is designed to illuminate the lens ring (12a) simultaneously along the entire circumference of the lens ring (12a) using directional lighting; - the image evaluation unit (24) is further configured such that the evaluation of the captured image comprises identifying parts contained in the captured image and assigning the parts contained in the captured image to surface elements (F1, ..., F4) of the frame groove, respectively, based on the brightness and / or shading of the corresponding parts, wherein the identifying of the parts comprises: -Mirror circle segmentation; - detecting edges bounded by these sections and surface elements (F1, ..., F4) within the segmented mirror ring (12a); and - completing these edges with the help of a first optimization method within the scope of which a cost function is minimized; and The image evaluation unit (24) is further configured such that the determination of the space curves (K1, . . . , K5) comprises: - based on the surface elements (F1, ..., F4), determining the spatial points (P1, ..., P5) that characterize the frame groove (12c) and through which the spatial curves (K1, ..., K5) extend, wherein the spatial points (P1, ..., P5) are located on the edge bounded by the portions and the surface elements (F1, ..., F4); and -The relative positions of these spatial points (P1, ..., P5) in space are determined by triangulation, wherein the triangulation is performed jointly for all spatial points (P1, ..., P5) of the spatial curves of the multiple spatial curves (K1, ..., K5), and a minimum distance criterion as an optimization objective in the form of a term of a cost function is included in the iterative optimization within the scope of the second optimization method.

11. The device of claim 10, wherein: The image evaluation unit (24) is further configured such that the completion of the edges comprises jointly performing a triangulation of the line of sight for all spatial points of the spatial curve of the contour (34).

12. The device according to claim 10 or 11, wherein: The device is designed so that when the bezel (12a) is illuminated, the illumination is incident on the bezel from the outside.

13. The apparatus of claim 10 or 11, wherein: The device is designed so that when the rim (12a) is illuminated, the directional illumination is incident on the spectacle frame from only one direction relative to the image capture device.

14. The apparatus of claim 10 or 11, wherein: The device is designed such that when the rim (12a) is illuminated, the directional illumination impinges on the spectacle frame as incident light from a direction in which the image evaluation unit is arranged relative to the spectacle frame.

15. The apparatus of claim 10 or 11, wherein: The lighting device includes one or more point light sources for providing the directional lighting, and is designed to illuminate the lens ring from a predetermined direction and produce a shadow as accurate as possible on the frame groove.

16. The device (10) according to claim 15, wherein The one or more point light sources are respectively designed as one of the following elements: an incandescent lamp, an LED, and a laser.

17. The device (10) according to claim 15, wherein: The one or more point light sources are each designed as a laser diode.

18. The apparatus (10) of claim 15, further comprising one or more light sources to facilitate transmitted light illumination of the lens ring.

19. The apparatus (10) of claim 16, further comprising one or more light sources to facilitate transmitted light illumination of the lens ring.

20. The apparatus of claim 10 or 11, wherein: The lighting device is designed to provide directional lighting that is at least one of the following types of lighting: collimated lighting, transmitted light lighting, incident light lighting, ultraviolet lighting, and polarized lighting.

21. The apparatus of claim 10 or 11, wherein: The image capture device is designed to capture both rims of the eyeglass frame in the image area when capturing the plurality of images.

22. The device (10) as claimed in claim 10 or 11, further comprising a holding device (28), wherein the holding device (28) is designed to hold a pair of glasses in a fixed position and orientation relative to the lighting device (14) and relative to the image capture device (16).

23. The device (10) of claim 22, wherein: The holding device (28) is further designed to hold the pair of glasses (12) in at least two different orientations relative to the lighting device (14) and relative to the image capture device (16), wherein the at least two different orientations define different predetermined viewing angles of the image capture device (16) relative to the rim (12a).

24. The device (10) according to claim 23, wherein The device is further designed to determine at least one of the following two properties of a spectacle lens mounted in a spectacle frame (12) held in the device (10): a vertex power and a spatially resolved power map.

25. The apparatus of claim 10 or 11, wherein: The image capture device comprises at least one camera, and wherein an optical axis of the at least one camera extends through one of the two bezels or profiles or through an opening of the pair of glasses spanned by the respective bezel.

26. The apparatus of claim 10, wherein: These edges are edges extending almost parallel to the lens ring (12a).

27. A method for grinding eyeglass lenses for installation in eyeglass frames, wherein: The outer shape (34) of a frame groove (12c) in a rim (12a) of the spectacle frame (12) comprises a method for determining the outer shape (34) of the frame groove according to any one of claims 1 to 9.

28. A computer-implemented method for determining the geometry of a spectacle lens, comprising a method for determining the outer shape (34) of a frame recess (12c) in a bezel (12a) of a spectacle frame according to any one of claims 1 to 9.

Images