Measuring system for generating 3D images

Abstract

The invention relates to a measuring system (100) for generating 3D images comprising - at least one illumination unit (4) for emitting multispectral light, - a wavelength-selective beam splitter (1) to which the illumination unit (4) is directed and which is configured to split the light emitted from the illumination unit (4) into at least two light beams of different spectral components and to deflect the light beams in a wavelength-dependent manner into at least two different directions, and - at least one wavelength-sensitive image sensor (2), in particular a video sensor, which is capable of recording for different spectral components, - wherein the image sensor (2) and the illumination unit (4) are directed to the beam splitter (1) and the recording device of the image sensor (2) corresponds to the direction of light emission from the illumination unit (4), - wherein the light emitted from the illumination unit (4) comprises those spectral components for which the image sensor (2) is capable of recording and for which the beam splitter (1) is designed and - wherein the image sensor (2), the beam splitter (1) and the illumination unit (4) are arranged rigidly with respect to one another in terms of movement and rotation.

Description

[0001]Measuring system for generating 3D images The invention relates to a measuring system for generating 3D images according to claim 1, Stereoscopic recording system for creating three-dimensional images of cavities according to claim 11,and a method for creating three-dimensional images of cavities according to claim 12. Prior art WO 2021016644 A1 discloses a method and a device for simultaneous 3D surface reconstruction of cavities or body openings or the like from up to three different viewing directions using up to three separately operated imaging systems. Each of the imaging systems is individually aligned in a desired measurement direction or equipped with mirrors for beam deflection. The image information recorded by the individual imaging systems is transferred into a common coordinate system and combined into a 3D image of the recorded surface of a cavity by means of registration with respect to a reference image. A disadvantage of such an approach is the complex hardware structure with multiple individual imaging systems.which is, on the one hand, expensive and, on the other hand, difficult to miniaturize due to its large number of individual parts and prone to damage. Furthermore, the procedure for creating three-dimensional images with three individual imaging systems, whose image recordings must be synchronized in time, is complex to implement. On the one hand, each of the individual imaging systems must first be calibrated individually and then the individual imaging systems must be calibrated to each other in order to locate the measurement points recorded by the imaging systems in a common measurement space. The creation of a dense 3D image is also complicated by the fact that the images from the individual imaging systems are spatially offset from one another. Thus, for example, when measuring boreholes, areas in the common 3D image are given,which cannot be filled with 3D measurement values. The object of the invention is therefore to eliminate the aforementioned disadvantages and to provide a measuring system or a stereoscopic recording system as well as a method with which the creation of three-dimensional images of cavities can be realized with a compact, robust hardware structure and, at the same time, with low time and computing expenditure. The invention solves this problem with a measuring system for generating 3D images according to claim 1. According to the invention, the measuring system comprises the following components: - at least one illumination unit for emitting multispectral light, - a wavelength-selective beam splitter, towards which the illumination unit is directed and which is designedto split the light emitted by the lighting unit into at least two light beams of different spectral components and to redirect the light beams in at least two different directions depending on the wavelength, and - at least one wavelength-sensitive image sensor, in particular a video sensor, which is capable of recording different spectral components, - wherein the image sensor and the lighting unit are directed towards the beam splitter and the recording direction of the image sensor corresponds to the direction of the light emission of the lighting unit, - wherein the light emitted by the lighting unit comprises those spectral components for which the image sensor is capable of recording and for which the beam splitter is designed and - wherein the image sensor,the beam splitter and the illumination unit are arranged in a manner that is motionally and rotationally rigid with respect to one another. For the sake of simplicity and clarity, an image sensor or a beam splitter is understood to mean a wavelength-sensitive image sensor or a wavelength-selective beam splitter, respectively. The terms "wavelength-selective" and "wavelength-sensitive" refer to the wavelength of light, i.e., the wavelength of electromagnetic radiation visible to the human eye. The image sensor can be sensitive to a wide wavelength range, for example, the visible light range, or to narrow wavelength ranges or bands. The image sensor and the illumination unit of a measuring system according to the invention are directed towards the beam splitter, and the recording direction of the image sensor corresponds to the direction of light emission from the illumination unit. This means,The illumination unit emits a mixture of light that falls on the beam splitter, which is then split into separate spectral components and redirected in different directions depending on the wavelength, thus exiting the beam splitter in different directions. The direction in which this mixture of light is emitted by the illumination unit corresponds to the recording direction of the image sensor. The individual spectral components that exit the beam splitter separately are reflected by the surroundings of the measuring system and then enter the beam splitter separately in the corresponding directions, where they are mixed and emitted as a mixture of light. Since the recording direction of the image sensor corresponds to the direction of the light output of the illumination unit,This light mixture emerging from the beam splitter falls on the image sensor. A measuring system according to the invention is characterized by a compact structure comprising an illumination unit, a wavelength-selective beam splitter, and a wavelength-sensitive image sensor. Depending on the design of the wavelength-selective beam splitter, up to three different measurement perspectives can be realized with just one measuring system. The different measurement perspectives are advantageously achieved by beam deflection with beam splitting and a spectral separation of the emitted light from the illumination unit after beam deflection / beam splitting, as well as by a correspondingly wavelength-selective image recording of the light reflected from the surface of the measurement object from up to three different beam paths, thus representing virtual camera positions. Thus, in comparison to the known prior art, it is advantageously not necessaryto provide individual measuring systems for the different desired measurement perspectives. Since the illumination unit, the image sensor, and the beam splitter are coordinated with each other in terms of their spectral components, it is possible, on the one hand, to fully utilize the deflection or separation by the beam splitter and to achieve deflection in as many different directions as possible, i.e., to achieve as many virtual camera positions as possible. On the other hand, it is also possible to use all spectral components of the light mixture incident on the image sensor for further processing and, thanks to the wavelength sensitivity of the image sensor, to use them as individual images from different virtual camera positions. The problem of time synchronization is advantageously eliminated with such a measuring system, since the images from the image sensor always contain all individual spectral components of the light incident on the image sensor.for which the image sensor is capable of recording and which represent the individual measurement perspectives or virtual camera positions. A particularly detailed recording of weakly or non-textured measurement object surfaces can be achieved if the illumination unit comprises a light pattern projector for projecting a directed random image, wherein it is provided in particular that- that the individual pixels of the random image have one of at least two different color values ​​and / or intensity values, and - the surroundings of each pixel within the random image can be clearly assigned to one or more pixels of the image sensor. Such a projected light pattern can advantageously be specifically optimized for the respective selected stereo correspondence finding and calibration algorithm, as well as for the respective measurement object surface and measurement object topography. In order to be able to transmit the light within the illumination unit or between the illumination unit and the wavelength-selective beam splitter over longer distances with simultaneously high intensity, and at the same time to achieve efficient thermal decoupling between the image sensor and an illuminant of the illumination unit, it can be provided,that the illumination unit comprises an optical waveguide. In order to ensure in a particularly simple manner that the light emitted by the illumination unit comprises those spectral components for which the image sensor is capable of recording and for which the beam splitter is designed, it can be provided that the illumination unit comprises at least two illumination means, in particular light-emitting diodes or laser diodes, for emitting light with at least two different spectral components and a further wavelength-selective beam splitter, wherein the illumination means are directed towards the further beam splitter in such a way that a light mixture can be created from the light emitted by the illumination means by the further beam splitter. By using a further wavelength-selective beam splitter, those spectral components that are additionally emitted by the illumination means are advantageouslyfor which the image sensor is not capable of recording, are separated and not transmitted by the illumination unit to the beam splitter. Alternatively, in a measuring system according to the invention, it can also be provided that the illumination unit comprises at least two illumination means, in particular light-emitting diodes or laser diodes, for emitting light with at least two different spectral components, wherein the different spectral components can be emitted individually one after the other. The individual illumination means of the illumination unit of the measuring system are switched on and off one after the other during a measuring process. In order to be able to use stereoscopy in a measuring system according to the invention, it can be provided that the measuring system comprises a further wavelength-sensitive image sensor, in particular a further video sensor,wherein the recording direction of the further image sensor is substantially parallel to the recording direction of the image sensor. In order to be able to analyze a particularly wide or large wavelength range, it can be provided in a measuring system according to the invention that the image sensor and / or the further image sensor comprises a plurality of sensor pixels, in particular arranged in rows and columns in the form of a pixel grid, wherein a gray value and / or brightness value can be determined by means of the individual sensor pixels. In particular, it is provided that the image sensor and / or the further image sensor is designed as a gray value video sensor. With two image sensors and a light pattern projector in a calibrated arrangement, three three-dimensional measuring arrangements can be realized.which are each directed at the measurement object surface from different viewing directions. If only the two image sensors are calibrated to each other and the measurement object surface is sufficiently textured, a light pattern projection can be dispensed with, and multispectral illumination is sufficient to achieve three viewing directions of the measurement object surface. Three different measurement perspectives or virtual camera positions can advantageously be achieved in a measuring system according to the invention if the beam splitter and / or the additional beam splitter is a cross-dichroic prism or a dichroic prism, in particular a Philips prism. In order to be able to obtain color information from an image sensor with photocells that can only detect brightness values ​​such as black, grayscale, and white, the image sensor and / or the additional image sensor can have the following: - a plurality of sensor pixels,which are arranged in rows and columns in the form of a pixel grid, and - an optical filter grid arranged correspondingly and repetitively to the pixel grid, in particular an optical filter grid according to the Bayer pattern, which is capable of receptive wavelengths for at least two, in particular three, different wavelengths. The individual spectral sensitivity ranges can advantageously be specifically adapted to the properties of the respective measurement object surface and / or to the spectral characteristics of the image sensor used. In order to be able to use image sensors in a measurement system according to the invention that are capable of receptive spectral components red, green, and blue, it can be provided that the wavelength-selective beam splitter is designed to split the light emitted by the illumination unit into three light beams of the spectral components red, green, and blue and to redirect the light beams in three different directions.wherein the directions are preferably at an angle of 90 degrees to one another. A further object of the invention is to provide a recording system for creating three-dimensional images of cavities that is both compact and robust. The invention achieves this object with a stereoscopic recording system comprising a measuring system according to the invention and a control and processing unit, - wherein the control and processing unit is connected to the illumination unit and is in data communication with the at least one image sensor and / or the further image sensor and is designed to - control the illumination unit to emit light, - assign a separate virtual camera position to each of the individual images created with the image sensor, each of which contains a single spectral component of the light incident on the image sensor for which the image sensor is capable of recording,- to determine the position and orientation of the image sensor and the illumination unit relative to each other in advance as part of an extrinsic camera calibration, and to determine intrinsic camera parameters of the image sensor, in particular the focal length and / or the lens distortion, as part of an intrinsic camera calibration, - to carry out a geometric calibration of the at least two virtual camera positions relative to each other in advance by determining a transformation into a common coordinate system for each of the at least two virtual camera positions, wherein these transformations each - assign exactly one position in the coordinate system to a position of a measurement space point determined in one of the virtual camera positions, and - if necessary, assign the same position in the coordinate system to the position of the same measurement space point determined in at least one other virtual camera position,and - to transfer the images created in the individual virtual camera positions into a common coordinate system using the determined transformations and thus combine them into a three-dimensional image. With a stereoscopic recording system according to the invention, it is possible to create three-dimensional images of cavities or a 3D surface reconstruction of the cavity wall from at least two different measurement perspectives using only a single measuring system, whereby the previously mentioned advantages of the measuring system can also be utilized. The hardware of the stereoscopic recording system can therefore advantageously be kept compact, so that the structure of the stereoscopic recording system as a whole can be miniaturized. With such a compact, miniaturized structure, it is advantageously possiblehand-held or stationary 3D reconstructions and 3D modeling or measurements, e.g. for general 3D measuring tasks in cavities and for the inspection and measurement of bores, cavities or difficult-to-access workpiece zones that would not be accessible for known measuring arrangements due to their size, but also for the preparation of medical diagnoses, for the production and adjustment of various medical aids such as hearing aids, hearing protection, dental crowns, etc. The object of the invention is further to provide a method with which three-dimensional images of cavities can be created using a stereoscopic recording system according to the invention, and which, at the same time, is computationally simple and rapid. The invention solves this problem with a method for creating three-dimensional images of cavities according to patent claim 12. According to the invention,that the stereoscopic recording system comprises - an illumination unit for emitting multispectral light, - a wavelength-selective beam splitter, towards which the illumination unit is directed, and - at least one wavelength-sensitive image sensor, in particular a video sensor, which is capable of recording various spectral components, - wherein the image sensor and the illumination unit are directed towards the beam splitter and the recording direction of the image sensor corresponds to the direction of the light output of the illumination unit, - wherein the light emitted by the illumination unit comprises those spectral components for which the image sensor is capable of recording and for which the beam splitter is designed, and - wherein the image sensor, the beam splitter and the illumination unit are arranged in a movement- and rotation-resistant manner relative to one another,- wherein the position and orientation of the image sensor and the illumination unit relative to each other are determined in advance as part of an extrinsic camera calibration, and intrinsic camera parameters of the image sensor, in particular the focal length and / or the lens distortion, are determined as part of an intrinsic camera calibration, - wherein multispectral light is emitted from the illumination unit onto the beam splitter and split by the beam splitter into at least two light beams of different spectral components, the light beams being deflected in at least two different directions depending on the wavelength, - wherein images of the light incident on the image sensor from the beam splitter are created with the image sensor, the images each containing a single spectral component of the light incident on the image sensor for which the image sensor is capable of recording,- wherein a separate virtual camera position is assigned to each of the respective recordings of the at least two individual spectral components of the incident light, - wherein a geometric calibration of the at least two virtual camera positions relative to each other is carried out in advance by determining a transformation into a common coordinate system for each of the at least two virtual camera positions, wherein these transformations each - assign exactly one position in the coordinate system to a position of a measurement space point determined in one of the virtual camera positions, and - assign the same position in the coordinate system to a position of the same measurement space point determined in at least one other virtual camera position,and - the images created in the individual virtual camera positions are transferred into a common coordinate system using the determined transformations and thus combined to form a three-dimensional image. In addition to the previously mentioned advantages of the stereoscopic recording system or the measuring system, which can also be used in a method according to the invention, registration is advantageously not required in such a method, since the merging of the individual 3D point clouds or the measurement points of the images created in the individual virtual camera positions into a common three-dimensional coordinate system is automatically achieved by calibrating the at least two virtual camera positions to each other. In order to be able to use as many measurement perspectives or virtual camera positions as possible in a method according to the invention, it can be provided that- that the measuring system comprises a further wavelength-sensitive image sensor, in particular a further video sensor, wherein the recording direction of the further image sensor is substantially parallel to the recording direction of the image sensor; - that, as part of the extrinsic camera calibration, the position and orientation of the further image sensor and the illumination unit relative to one another, as well as the position and orientation of the image sensor and the further image sensor relative to one another, are determined; and - that, as part of the intrinsic camera calibration, intrinsic camera parameters of the further image sensor, in particular the focal length and / or the lens distortion, are determined. In this way, it is advantageously possible to obtain up to three different measurement perspectives and three 3D data sources, independent of the downstream beam deflection.so that, together with the light wavelength-selective beam deflection, up to nine independent 3D data sources are available. According to a further advantageous variant of a method according to the invention, it can be provided that the illumination unit comprises at least two illumination means, in particular light-emitting diodes or laser diodes, for emitting light with at least two different spectral components, and that the different spectral components are emitted individually one after the other. In order to be able to analyze a particularly broad or large wavelength range and to manage without a color image sensor, it can be provided in a method according to the invention that the image sensor and / or the further image sensor comprises a plurality of sensor pixels, in particular arranged in rows and columns in the form of a pixel grid, wherein a gray value and / or brightness value is determined by means of the individual sensor pixels, wherein it is provided in particular thatthat the image sensor and / or the further image sensor is designed as a grayscale video sensor and that the individual pixels of the recordings of the light incident on the image sensor from the beam splitter contain at least one grayscale value and / or brightness value. Further advantages and embodiments of the invention will become apparent from the description and the accompanying drawings. The invention is schematically illustrated below in the drawings using particularly advantageous, but non-limiting, embodiments and is described by way of example with reference to the drawings. The following schematically show: Fig. 1 an embodiment of a measuring system according to the invention, Figs. 2 and 3 embodiments of wavelength-selective beam splitters,Fig. 4 shows an embodiment of an optical filter grid of a wavelength-sensitive image sensor. Fig. 1 schematically shows the structure of a measuring system 100 according to the invention, as it can be used in a stereoscopic recording system according to the invention for creating three-dimensional images of cavities. A measuring system 100 according to the invention can be used for a time-synchronous 3D scene reconstruction from up to three independent measuring perspectives or virtual camera positions. The up to three different measuring perspectives or virtual camera positions are multispectrally based with only a single measuring system 100 with corresponding multispectral illumination.Light wavelength-selective or light wavelength-dependent beam deflection and light wavelength-sensitive image acquisition are realized. The measuring system 100 in Fig. 1 comprises a wavelength-selective beam splitter 1 for the wavelength-selective beam deflection. In the exemplary embodiment in Fig. 1, the light wavelength-selective beam deflection is realized with a cross-dichroic prism (see Fig. 2a), which acts selectively for the wavelength ranges red R, green G, and blue B and splits white light W into these spectral components or combines these spectral components into white light W. Such a cross-dichroic prism is described, for example, in US 6407868 B1. The cross-dichroic prism splits light emitted onto the prism into three light beams of the spectral components red R, green G and blue B and redirects these three light beams in three different directions,which are at a 90-degree angle to each other. This results in three orthogonally aligned, RGB light wavelength-dependent viewing directions. As an alternative to a cross-dichroic prism, a dichroic prism (see Fig. 2b) can be used, such as the one developed by Philips, which is therefore known as the Philips prism, and is now manufactured, for example, by www.foctek.net, last accessed on November 30, 2022. Such a prism is described, for example, in US 3202039 A and available at https: / / www.foctek.com / upfiles / file / 13730058320.pdf, last accessed on November 29, 2022. The dichroic prism also splits light emitted onto the prism into three light beams of the spectral components red R, green G, and blue B and redirects these three light beams in three different directions. However, the directions are not at an angle of 90 degrees to each other,but at an angle greater or smaller than 90 degrees. A light beam in Fig. 2b enters a first prism P1, and the blue spectral component B is reflected by the dichroic filter F1. Longer wavelengths pass through and enter a second prism P2 and are split by a second filter F2. The filter F2 reflects the red spectral component R. The green spectral component G passes through the two prisms P1 and P2 and exits opposite the entrance point of the light W. An air gap is arranged between the prisms P1 and P2, and the angles of the prisms P1 and P2 are chosen such that the blue spectral component B and the red spectral component R are deflected by total internal reflection in the prism. Cross-dichroic and dichroic prisms can also be used.to combine light of different wavelengths. In wavelength-sensitive image acquisition, various possibilities for utilizing stereoscopy exist in a measuring system 100 according to the invention. The following components can be used: - a wavelength-sensitive image sensor 2 and a light pattern projector 3, wherein the image sensor 2 can be integrated, for example, in a video camera 20 and thus forms a video camera / projector 3D measuring system, which is advantageously particularly well-suited for weakly textured or non-textured measurement object surfaces, - two wavelength-sensitive image sensors 2 and an unstructured scene illumination, wherein the image sensors 2 can be integrated, for example, individually or jointly in a video camera 20. Such a measuring system 100 is particularly advantageous, for example, for highly textured measurement object surfaces, or - two wavelength-sensitive image sensors 2 and a light pattern projector 3,The image sensors 2 can, for example, be individually or jointly integrated into a video camera 20, thus forming a stereo camera / projector 3D measuring system, which is particularly advantageous for, for example, weakly textured or non-textured measurement object surfaces. Thus, in a measuring system 100 according to the invention, it is possible to use two image sensors 2, so that a spatial impression of the recorded areas can be achieved with the aid of stereoscopy or images recorded from different viewing angles. Alternatively, as previously mentioned, an image sensor 2 and a light pattern projector 3 can be used, which is realized in the embodiment in Fig. 1. The light emitted by the light pattern projector 3 or the scene illumination contains those spectral components,for which the respective wavelength-selective beam splitter 1 used is designed and for which the respective wavelength-sensitive image sensor 2 is sensitive. In the exemplary embodiment in Fig. 1, the three measurement perspectives in three separate spectral ranges or the three virtual camera positions 61, 62, 63 are realized with only a single measuring system 100 comprising a light pattern projector 3, an image sensor 2, and a wavelength-selective beam splitter 1 as an optical component for light wavelength-dependent beam deflection. The spectral components for which the beam splitter 1 and the image sensor 2 are selective or sensitive are the wavelength ranges red R, green G, and blue B. In the exemplary embodiment in Fig. 1, the camera lens 21 of a video camera 20, into which the image sensor 2 is integrated, is directed onto the wavelength-selective beam splitter 1. The image sensor 2 has a plurality of sensor pixels,which are arranged in rows and columns and thus form a pixel grid. The image sensor 2 in Fig. 1 is an image sensor capable of recording the red, green, and blue wavelength ranges, or an RGB image sensor. The video camera 20 thus represents an RGB video camera. In order to make the image sensor 2 of the video camera 20 sensitive to the red, green, and blue wavelength ranges, the image sensor 2 in the exemplary embodiment is provided with a repetitively arranged optical filter grid corresponding to the pixel grid. A suitable filter grid of this type is transparent to up to four different light wavelengths. The RGB image sensor or RGB video sensor in Fig. 1 is equipped with an optical filter grid according to the Bayer pattern (see Fig. 3), i.e., an RGB Bayer filter mask, in order to make it sensitive to the red,green and blue sensitive. Such a filter grid based on the Bayer pattern is described, for example, in DE 2608998 B2. However, the Bayer pattern filter arrangement, which is standard for RGB color video sensors, is mentioned here only as an example of a suitable filter grid. As an alternative to the Bayer pattern filter grid, other filter grids such as "Complementary Color Mosaic Filter" or "Primary Color Vertical Stripe Filter" can also be used, as described, for example, in, support / knowledge-base / frequently-asked-questions / what-are-color-filters / 14989 / , last accessed on November 29, 2022. In the first exemplary embodiment, the measuring system 100 further comprises a light pattern projector 3 in the form of an RGB light pattern projector with a projection lens 31, a collimation lens 33, and a slide 32 with a random image. The collimation lens 33—in the simplest variant, a collimation lens—collects diverging light emitted from a light source, directs the light "parallel," and, in the case of projection lenses with exit pupil positions that lie behind the light pattern slide plane in the projection direction, directs it directly via the light pattern slide and the projection lens onto the measurement object surface.Additionally, a focusing lens can be inserted into the beam path after the collimating lens, which focuses the light into the exit pupil of projection lenses, which lies in front of the light pattern slide plane in the projection direction. A combination of collimating lens and focusing lens is referred to as a "condenser." The pattern of the projected light pattern of the light pattern projector 3 is designed to optimally utilize the imaging properties of all optical components used, and to impart a known texture to the scene surface for the 3D measurement process. The random image has pixels that exhibit two or more differing color values ​​or intensities, and within the random image, the surroundings of each pixel can be uniquely assigned to one or more pixels of the image sensor 2.The background to using such a random image is that the environment of each pixel, for example a 3x3 or 5x5 environment within the entire random image, should be clearly assignable to the respective corresponding pixel in the three-dimensional image. Due to the uniqueness of the environment, a spatial point can be contained at different locations, possibly in two or three images created in each of the virtual camera positions. As already mentioned, the light source of the light pattern projector 3 is tuned in the emitted wavelength ranges to the cross-dichroic prism and the image sensor 2 of the RGB video camera. In Fig. 1, an illumination unit 4 is directed as an RGB light source onto the light pattern projector 3.In the exemplary embodiment, the lighting unit 4 comprises a further beam splitter 11 in the form of a further cross-dichroic prism, as well as three lighting devices - namely a red LED 41, a green LED 42 and a blue LED 43 - which are each attached to the corresponding sides of the further cross-dichroic prism and serve as light sources for the spectral components red, green and blue. By aligning the lighting devices to the further beam splitter 11, the further beam splitter 11 can create a light mixture from the light emitted by the lighting devices. Alternatively, without the use of a further beam splitter 11 or a cross-dichroic prism, a white light source such as an LED, incandescent lamp or gas discharge lamp with a corresponding spectral distribution can also be used as the light source for the light pattern projector 3 or the scene lighting.In the exemplary embodiment, the created light mixture in the form of RGB light is coupled out on the corresponding side of the cross-dichroic prism via an optical fiber 5, which is directed onto the collimation lens 33 of the light pattern projector 3 or opens into it. The RGB light from the RGB light source is thus coupled into the light pattern projector 3 via the optical fiber 5 and the collimation lens 33. The RGB light can be transmitted over long distances with high intensity using an optical fiber 5, while efficient thermal decoupling is provided between the RGB video camera and the RGB light source. The light pattern projector 3 and the video camera 20 or its image sensor 2 are arranged so as to be mechanically rigid, i.e. movement- and rotation-resistant, to one another, and also mechanically rigid, i.e. movement- and rotation-resistant, to the corresponding side of the beam splitter 1 orcross-dichroic prism and the further beam splitter 11 or further cross-dichroic prism. A measuring system 100 as shown in Fig. 1 and described above, together with a control and processing unit, can form a stereoscopic recording system with which three-dimensional images of cavities can be created. The control and processing unit controls the illumination unit 4 for emitting light, and optionally also controls the image sensor 2 for creating images. The control and processing unit further handles the data processing steps of a method according to the invention, described in detail below, within the framework of extrinsic and intrinsic camera calibration, geometric calibration, and the merging of the images created in the virtual camera positions 61, 62, 63 to form a three-dimensional image of the surroundings of the measuring system 100.The control and processing unit is in data communication with the image sensor and assigns a separate virtual camera position 61, 62, 63 to the images created by the image sensor 2, each of which contains the red spectral component R, the green spectral component G, or the blue spectral component B of the light exiting the beam splitter 1 and incident on the image sensor. This means that the first virtual camera position 61 corresponds to the red spectral component R, the second virtual camera position 62 corresponds to the green spectral component G, and the third virtual camera position 63 corresponds to the blue spectral component B, which each exit separately from the beam splitter 1 and illuminate the surroundings of the measuring system 100 or are reflected by the surroundings of the measuring system 100, enter the beam splitter 1, and are incident on the image sensor 2 in a bundled manner by the beam splitter 1.In the geometric arrangement described above, in a method according to the invention for creating three-dimensional images of cavities, the three orthogonal 3D measurement paths "red", "green", and "blue", i.e. the three virtual camera positions 61, 62, 63, are calibrated independently of one another intrinsically and extrinsically using calibration methods known from the prior art. Such calibration methods are described, for example, in Fabio Remondino, Clive Fraser: Digital camera calibration methods: Considerations and comparisons. In: ISPRS Commission V Symposium 'Image Engineering and Vision Metrology'. ISPRS, September 25, 2006. During intrinsic calibration, parameters for correcting the optical distortions, the resulting optical focal lengths, and the positions of the points of intersection of the optical axes through the image plane / slide plane of the image sensor 2 / the light pattern projector 3 are determined.Using these parameters, a corrected (ideal) image is calculated, which would be created with a pinhole camera / pinhole projector. During extrinsic calibration, the positions and orientations of the image sensor 2 and the light pattern projector 3—as a pinhole camera model or a pinhole projector model—are determined relative to each other, i.e., with 6 degrees of freedom, using the pinhole camera / pinhole projector parameters determined by the intrinsic calibration. In this way, the position and orientation of the image sensor 2 and the illumination unit 4 relative to each other are determined, as well as the intrinsic camera parameters of the image sensor 2, such as the focal length, lens distortion, etc., are determined. As a further process step, a geometric calibration of the virtual camera positions 61, 62, 63 relative to each other is performed.Such a calibration procedure is described, for example, in Jacek Komorowski and Przemyslaw Rokita, "Extrinsic Camera Calibration Method and Its Performance Evaluation," Proceedings of the 2012 international conference on Computer Vision and Graphics, September 2012, DOI:10.1007 / 978-3-642-33564-8_16. For each of the virtual camera positions 61, 62, and 63, a transformation into a common coordinate system is determined. These transformations are characterized by the fact that they assign exactly one position in the coordinate system to a position of a measurement space point determined in one of the virtual camera positions 61, 62, and 63. Positions of the same measurement space point determined in one or more other virtual camera positions are assigned the same position in the coordinate system.During "geometric" calibration, the individual positions and orientations of up to three different measurement directions generated with beam splitter 1 or the cross-dichroic prism are determined relative to each other after the extrinsic and intrinsic calibration. This makes it possible to directly transform the 3D point clouds from the individual measurement directions into a common 3D model or a common three-dimensional coordinate system. Using the individual RGB partial images—i.e., the images captured at the individual virtual camera positions 61, 62, 63—and the known projection pattern, a 3D point cloud can be calculated for the three respective viewing directions, for example, using correlation-based correspondence search. The 3D point clouds can be transferred into a common coordinate system, thus creating a three-dimensional image of the captured area.A method for correlation-based correspondence search is described, for example, in Lazaros, Nalpantidis; Sirakoulis, Georgios Christou; Gasteratos, Antonios (2008), “Review of Stereo Vision Algorithms: From Software to Hardware”. International Journal of Optomechatronics. 2 (4): 435–462. If, as described above, two wavelength-sensitive image sensors 2 and an illumination unit 4, for example with a light pattern projector 3, are optionally provided, then during the extrinsic camera calibration not only the position and orientation of the image sensor 2 and the illumination unit 4 relative to one another are determined, but also the position and orientation of the further image sensor and the illumination unit 4, as well as the position and orientation of the image sensor 2 and the further image sensor relative to one another.As part of the intrinsic camera calibration, the intrinsic camera parameters of the additional image sensor, in particular the focal length and / or the lens distortion, are also determined. In this way, it is possible to obtain three different measurement perspectives and three 3D data sources independently of the downstream beam deflection, so that together with the light wavelength-selective beam deflection, up to nine independent 3D data sources are then available. A measuring system 100 according to the invention or a stereoscopic recording system according to the invention can be arranged on a carrier which has a preferred feed direction and can be pushed or inserted into a cavity manually or automatically. A measuring system 100 according to the invention orA stereoscopic recording system according to the invention and a method according to the invention can be used, for example, in all fields of application that are characterized by recording areas that are difficult to access and therefore require miniaturized measuring or recording systems and possibly have a surface with little structure. This applies, for example, to fields of application in industry or robotics or, for example, in the cosmetic or surgical field, for example when recording three-dimensional images of body orifices such as the ear canal of humans or animals. In the field of surgery, for example, an application is also possible, for example for recording three-dimensional images of internal organs. Although embodiments of a measuring system or method according to the invention have been...a stereoscopic recording system according to the invention and a method according to the invention have been described above in connection with an image sensor 2 which is sensitive to the wavelength ranges red R, green G and blue B. Alternatively, however, an image sensor and / or another image sensor which is sensitive to a broad wavelength range - for example the range of visible light - can also be used without restriction, for example a grayscale video sensor whose sensor pixels each determine a grayscale value and / or brightness value and do not have a color filter. In this case, the individual light sources or illumination means of the illumination unit are switched on and off one after the other during a measuring process, i.e. the illumination means are not operated simultaneously. In this case, a grayscale video image is recorded for each image, which is then used to calculate a 3D reconstruction.

Claims

1. A measuring system (100) for generating 3D images, comprising - at least one illumination unit (4) for emitting multispectral light, - a wavelength-selective beam splitter (1) onto which the illumination unit (4) is directed and which is designed to split the light emitted by the illumination unit (4) into at least two light beams of different spectral components and to redirect the light beams in at least two different directions depending on the wavelength, and - at least one wavelength-sensitive image sensor (2), in particular a video sensor, which is capable of recording different spectral components, - wherein the image sensor (2) and the illumination unit (4) are directed onto the beam splitter (1) and the recording direction of the image sensor (2) corresponds to the direction of the light emission of the illumination unit (4), - wherein the light emitted by the illumination unit (4) comprises those spectral components,for which the image sensor (2) is capable of recording and for which the beam splitter (1) is designed, and - wherein the image sensor (2), the beam splitter (1), and the illumination unit (4) are arranged in a movement- and rotation-resistant manner relative to one another.

2. Measuring system (100) according to claim 1, characterized in that the illumination unit comprises a light pattern projector (3) for projecting a directed random image, wherein it is provided in particular that - the individual pixels of the random image have one of at least two different color values ​​and / or intensity values, and - the environment of each pixel within the random image can be uniquely assigned to one or more pixels of the image sensor.

3. Measuring system (100) according to claim 1 or 2, characterized in that the illumination unit (4) comprises an optical waveguide (5).

4. Measuring system (100) according to one of the preceding claims, characterized in thatthat the lighting unit (4) comprises at least two lighting means, in particular light-emitting diodes or laser diodes, for emitting light with at least two different spectral components and a further wavelength-selective beam splitter (11), wherein the lighting means are directed onto the further beam splitter (11) in such a way that, a light mixture can be created from the light emitted by the illumination means by the further beam splitter (11).

5. The measuring system (100) according to one of claims 1 to 3, characterized in that the illumination unit (4) comprises at least two illumination means, in particular light-emitting diodes or laser diodes, for emitting light with at least two different spectral components, wherein the different spectral components can be emitted individually one after the other.

6. The measuring system (100) according to one of the preceding claims, characterized in that the measuring system (100) comprises a further wavelength-sensitive image sensor, in particular a further video sensor, wherein the recording direction of the further image sensor is substantially parallel to the recording direction of the image sensor (2).Measuring system (100) according to one of the preceding claims, characterized in that the image sensor (2) and / or the further image sensor comprises a plurality of sensor pixels, in particular arranged in rows and columns in the form of a pixel grid, wherein a gray value and / or brightness value can be determined by means of each individual sensor pixel. In particular, it is provided that the image sensor (2) and / or the further image sensor is designed as a gray value video sensor.

8. Measuring system (100) according to one of the preceding claims, characterized in that the beam splitter (1) and / or the further beam splitter (11) is a cross-dichroic prism or a dichroic prism. 9.Measuring system (100) according to one of the preceding claims, characterized in that the image sensor (2) and / or the further image sensor has the following: - a plurality of sensor pixels arranged in rows and columns in the form of a pixel grid, and - an optical filter grid arranged correspondingly and repetitively with the pixel grid, in particular an optical filter grid according to the Bayer pattern, which is capable of recording at least two, in particular three, different wavelengths.

10. Measuring system (100) according to one of the preceding claims, characterized in that the wavelength-selective beam splitter (1) is designed to split the light emitted by the illumination unit (4) into three light beams of the spectral components red (R), green (G), and blue (B), and to direct the light beams in three different directions. to redirect, wherein the directions are preferably at an angle of 90 degrees to one another.

11. A stereoscopic recording system for creating three-dimensional images of cavities, comprising a measuring system (100) according to one of claims 1 to 10, and a control and processing unit, - wherein the control and processing unit is connected to the illumination unit (4) and is in data communication with the at least one image sensor (2) and / or the further image sensor and is designed to - control the illumination unit (4) to emit light, - assign a separate virtual camera position (61, 62, 63) to the individual images created with the image sensor (2), each of which contains a single spectral component of the light incident on the image sensor (2) for which the image sensor (2) is capable of recording,- to determine the position and orientation of the image sensor (2) and the illumination unit (4) relative to one another in advance as part of an extrinsic camera calibration, and to determine intrinsic camera parameters of the image sensor (2), in particular the focal length and / or the lens distortion, as part of an intrinsic camera calibration, - to carry out a geometric calibration of the at least two virtual camera positions (61, 62, 63) relative to one another in advance by determining a transformation into a common coordinate system for each of the at least two virtual camera positions (61, 62, 63), wherein these transformations each - assign a position of a measuring space point determined in one of the virtual camera positions (61, 62, 63) to exactly one position in the coordinate system, and - if necessary, to the respective positions determined in at least one other virtual camera position (61, 62,63) determines the same position in the coordinate system, and - converting the images created in the individual virtual camera positions (61, 62, 63) into a common coordinate system using the determined transformations and thus combining them to form a three-dimensional image.

12. A method for creating three-dimensional images of cavities with a measuring system according to one of claims 1 to 10, in particular a stereoscopic recording system according to claim 11, comprising, - an illumination unit (4) for emitting multispectral light, - a wavelength-selective beam splitter (1) onto which the illumination unit (4) is directed, and - at least one wavelength-sensitive image sensor (2), in particular a video sensor, capable of recording various spectral components, - wherein the image sensor (2) and the illumination unit (4) are directed onto the beam splitter (1) and the recording direction of the image sensor (2) corresponds to the direction of light emission from the illumination unit (4), - wherein the light emitted by the illumination unit (4) comprises those spectral components for which the image sensor (2) is capable of recording and for which the beam splitter (1) is designed, and - wherein the image sensor (2), the beam splitter (1), and the illumination unit (4) are arranged in a movement- and rotation-resistant manner relative to one another,- wherein the position and orientation of the image sensor (2) and the illumination unit (4) relative to each other are determined in advance as part of an extrinsic camera calibration, and intrinsic camera parameters of the image sensor (2), in particular the focal length and / or the lens distortion, are determined as part of an intrinsic camera calibration, - wherein multispectral light is emitted by the illumination unit (4) onto the beam splitter (1) and split by the beam splitter (1) into at least two light beams of different spectral components, the light beams being deflected in at least two different directions depending on the wavelength, - wherein images of the light incident on the image sensor (2) from the beam splitter (1) are created with the image sensor (2), the images each containing a single spectral component of the light incident on the image sensor (2) for which the image sensor (2) is capable of recording,- wherein a separate virtual camera position (61, 62, 63) is assigned to each of the respective recordings of the at least two individual spectral components of the incident light, - wherein a geometric calibration of the at least two virtual camera positions (61, 62, 63) relative to one another is carried out in advance by determining a transformation into a common coordinate system for each of the at least two virtual camera positions (61, 62, 63), wherein these transformations each - assign exactly one position in the coordinate system to a position of a measurement space point determined in one of the virtual camera positions (61, 62, 63), and - assign the same position in the coordinate system to a position of the same measurement space point determined in at least one other virtual camera position (61, 62, 63), and, - the images created in the individual virtual camera positions (61, 62, 63) are converted into a common three-dimensional coordinate system using the determined transformations and thus combined to form a three-dimensional image. 13.Method according to claim 12, characterized in that - the measuring system (100) comprises a further wavelength-sensitive image sensor, in particular a further video sensor, wherein the recording direction of the further image sensor is substantially parallel to the recording direction of the image sensor (2), and - that within the scope of the extrinsic camera calibration, the position and orientation of the further image sensor and the illumination unit (4) relative to one another, as well as the position and orientation of the image sensor (2) and the further image sensor relative to one another, are determined, and - that within the scope of the intrinsic camera calibration, intrinsic camera parameters of the further image sensor, in particular the focal length and / or the lens distortion, are determined. 14.Method according to claim 12 or 13, characterized in that the illumination unit (4) comprises at least two illumination means, in particular light-emitting diodes or laser diodes, for emitting light with at least two different spectral components, and in that the different spectral components are emitted individually one after the other.

15. Method according to one of claims 12 to 14, characterized in that the image sensor (2) and / or the further image sensor comprises a plurality of sensor pixels, in particular arranged in rows and columns in the form of a pixel grid, wherein a gray value and / or brightness value is determined by means of each individual sensor pixel. In particular, it is provided that the image sensor (2) and / or the further image sensor is designed as a gray value video sensor and that the individual pixels of the recordings of the light incident on the image sensor (2) from the beam splitter (1) contain at least one gray value and / or brightness value.

Images