3D scanner for optical scanning of an object using light patterns

Description

[0001] The present invention lies in the technical field of two-dimensional (2D) optical scanning of the surface of an object using light patterns to generate a digital three-dimensional (3D) representation of the object. In particular, the invention relates to a 3D scanner suitable for scanning the surface of an object using light patterns. A preferred application is in the field of dental technology for optically scanning the surfaces of dental objects, particularly in the form of a desktop 3D scanner.

[0002] 3D scanners for optically scanning objects are well known in practice. Such 3D scanners essentially function by projecting a two-dimensional light stripe pattern onto the surface of the object to be optically scanned. Part of the light beam is reflected by the surface of the object and captured by one or more cameras to capture images of the light beam reflected from the surface at a non-zero angle to the projection direction. Using a suitable projection geometry, such as triangulation, information about the spatial position of individual points on the object's surface can be obtained, allowing a digital 3D representation of the object to be computationally determined.Typically, to derive the coordinates of the object's surface, the object and the projection unit are moved relative to each other so that new images can be acquired, from which the 3D coordinates for other surface areas of the object are derived. Such a method is generally known to those skilled in the art, for example, under the term "striped light projection" and has been described numerous times in patent literature (see, for example, DE 19608632 A1). 3D scanners based on striped light projection are commercially available.

[0003] US 2018 / 188020 A1 discloses a 3D scanner comprising a camera in "rolling shutter" mode and a projector, wherein the frame rate of the projection unit corresponds to 3 times the frame rate of the camera, wherein during one exposure cycle of the camera, four different patterns are projected at four different positions with respect to the readout direction, offset in time and synchronously with the line exposure.

[0004] In optical surface scanning, the light beam reflected from an object is captured by at least one camera with an image sensor and converted into an electrical signal. The image sensor (area sensor) comprises a plurality of light-sensitive semiconductor elements (pixels) in a two-dimensional array that forms the image field, with each semiconductor element representing its own photodetector, through which electrons can be generated from the incident photons. During the exposure time, the electrons released by the incident light in the semiconductor material are collected. The amount of charge per pixel is then read out using the sensor readout technology used, with the generated charge being converted into a voltage signal by a converter transistor. The exposure time can be controlled by a mechanical or electronic shutter.

[0005] 3D scanners are increasingly using silicon-based image sensors manufactured using CMOS technology, whose performance has recently been continually improved and even surpasses that of CCD sensors. Such image sensors can be read out in various operating modes. In the so-called global shutter mode, all pixels of the image sensor are exposed at the same time. At the end of the exposure time, the charge accumulated in each pixel is transferred to a storage capacitor assigned to the respective pixel, where it is temporarily stored or converted into a proportional voltage, which is then temporarily stored. The temporarily stored signals are then read out line by line. Image sensors for global shutter mode, however, have a more complex structure, as the required storage capacitors require additional space in the respective pixel.The so-called rolling shutter mode is advantageous, in which the individual pixel rows of the image sensor are exposed sequentially one after the other, and each pixel row is read out at the end of the exposure. Exposure is controlled in a conventional manner by erasing the charge accumulated in the pixels to be exposed by connecting the pixel to a reset potential before the start of exposure, and by accumulating the charge generated in the light-sensitive element during the specified exposure time. After the exposure time has elapsed, the charge accumulated in the light-sensitive element is transferred to the converter transistor, where it is converted into a corresponding voltage signal, which is transferred via the column line to the signal processing circuit, where it is amplified and processed.In rolling shutter mode, the individual pixel lines of the image field are exposed with a time delay (staggered in time), whereby the time required to read out the pixel line determines the time offset that occurs between the individual pixel lines of the image field.

[0006] In practical applications, particularly in dental technology, it is important that the optical scanning of the surface of objects to generate a digital 3D representation of the objects is carried out within the shortest possible time period, but still with sufficient accuracy, so that, for example, a dental restoration can be produced as time- and cost-efficiently as possible without reducing the essential requirements for quality.

[0007] In contrast, the object of the present invention is to provide an improved 3D scanner with which digital 3D representations of objects can be created within a very short period of time with sufficient image quality.

[0008] These and other objects are achieved according to the invention by a 3D scanner according to the independent patent claim. Advantageous embodiments of the invention are set out in the subclaims.

[0009] According to the invention, a 3D scanner for optically scanning the surface of an object is shown. The 3D scanner is preferably used in dental technology for optically scanning dental objects such as the upper and lower jaws, preferably as a desktop scanner.

[0010] The 3D scanner comprises one or more scanning units, each having a projection unit and one or more cameras. The projection unit is used to project images (frames) containing light patterns, preferably striped light patterns. The frame rate of the projection unit corresponds to N times the frame rate of the at least one camera, where N is a natural number greater than 1 (N≥1). The frame rate of the projection unit refers to the number of projected (individual) images (frames) per second, which is typically specified in fps (frames per second). The frame rate of the camera refers to the refresh rate for recording (individual) images (frames) based on sensor- or camera-internal timing. Accordingly, the refresh rate of the camera is specified in fps (frames per second).The scanning unit is configured so that the projection unit projects images containing light patterns onto the surface of an object and the images reflected from the surface are recorded by one or more cameras.

[0011] Each camera has an electro-optical sensor or image sensor (area sensor), in particular a CMOS image sensor, with a plurality of image points (pixels). The pixels are arranged in rows and columns of an image field, with each pixel comprising at least one light-sensitive element for generating electrical charge from light incident along an exposure direction, and a converter transistor for converting a charge generated by the light-sensitive element into a voltage signal at an output of the converter transistor.

[0012] The two-dimensional arrangement of pixels corresponds to the image field (array) of the image sensor. The pixels are arranged in a plurality of pixel rows, each extending along a first direction (x) and parallel to one another in a second direction (y) perpendicular thereto, i.e., in a row or row-like sequence. The image area of ​​the image sensor is rectangular or square.

[0013] The cameras are each set up so that the pixels of the image sensor are exposed and read out in rolling shutter mode, in which the exposure of the pixel rows takes place line by line in staggered time or with a time offset and the pixel rows are read out sequentially immediately after their end of exposure.

[0014] Depending on the image area of ​​the image sensor, the images projected by the projection unit are each in the shape of a rectangle or square. Each projected image has a first direction (x') and a second direction (y') perpendicular thereto, wherein the first direction (x') of the projected image is assigned to the first direction (x) of the image sensor and the second direction (y') of the projected image is assigned to the second direction (x) of the image sensor. Each projected image is assigned uniquely (i.e. 1:1) to the image area of ​​the image sensor. In other words, the projected images are each assigned to the image area of ​​the image sensor in such a way that an image area of ​​a projected image is used to expose a corresponding sensor area of ​​the image area of ​​the image sensor.Thus, the pixel rows of the image sensor are assigned to line-shaped image areas of the projected image (1:1) in an analogous manner, so that a pixel row is exposed by a corresponding line-shaped image area. Any image area is assigned 1:1 to the corresponding sensor area.

[0015] The images always have the same dimensions along the second direction (y'), with the dimensions being measured from a first image edge to a second image edge along the second direction (y'). When viewed vertically, the first image edge can be understood as the upper edge, and the second image edge as the lower image edge.

[0016] The pixel rows of the image sensor are exposed sequentially and staggered in time in respective passes and read out line by line. For the purposes of the present invention, the single exposure and readout of all pixel rows along the row-like sequence of pixel rows is referred to as a "pass." The images each have a dimension SF in a direction associated with the line-by-line readout of the pixel rows, i.e., along the row-like arrangement of the pixel rows. Along this direction, which is associated with the line-by-line readout of the pixel rows, image regions have a dimension SR and light patterns have a dimension SP. In the direction perpendicular to this, i.e., in the direction of the pixel rows, the dimensions of the images, image regions, and light patterns are each selected such that all pixels of a single pixel row are exposed.

[0017] Each projected image has at least one light pattern, preferably a striped light pattern. The light pattern is rectangular or square depending on the shape of the image, wherein the dimension of the light pattern in the first direction (x') corresponds to that of the projected image. The dimension of the light pattern SP in the second direction (y') can differ from that of the projected image. Thus, the light pattern can extend completely or only partially across the image, i.e. the light pattern can also extend over only a partial area of ​​an image, wherein the position within the image is selectable. Correspondingly, the light pattern has a first light pattern edge and a second light pattern edge along the second direction (y'). If a light pattern completely fills an image, the first light pattern edge borders on the first image edge and, at the same time, the second light pattern edge borders on the second image edge.If the light pattern dimension SP is smaller than the image dimension SF, . i) the first light pattern edge is adjacent to the first image edge and at the same time the second light pattern edge is not adjacent to the second image edge, or ii) the first light pattern edge is not adjacent to the first image edge and at the same time the second light pattern edge is not adjacent to the second image edge, or iii) the first light pattern edge is not adjacent to the first image edge and at the same time the second light pattern edge is adjacent to the second image edge.

[0018] The association between the image and the image area of ​​the image sensor is also given if a light pattern does not completely fill a projected image and no light pattern is projected into the remaining areas of the image.

[0019] As already explained, a frame rate of the projection unit is N times a frame rate of the at least one camera, where N is a natural number greater than 1. In a direction associated with the line-by-line readout of the pixel lines, images have an image dimension SF, image areas have an image area dimension SR, and light patterns have a light pattern dimension SP.

[0020] The 3D scanner further comprises a control unit for controlling the at least one scanning unit. The control unit is configured to project different light patterns onto the surface of an object, provided that in a respective pass for successive (single) exposure of all pixel rows of the image sensor, all pixel rows are always exposed to the same light pattern. The control unit is configured to computationally determine a digital 3D representation of an object from the reflected images or light patterns captured by the one or more cameras.

[0021] According to one embodiment of the invention, for which independent protection is claimed, a set of N+1 images is projected onto the surface of the object per pass, with the light patterns being projected such that the light pattern dimension SP satisfies the following: SP ≥ SF * 2 / (N+2). In other words, the light pattern dimension SP is at least equal to the image dimension SF multiplied by 2 divided by (N+2).

[0022] In this embodiment, a light pattern is projected into each image of a set belonging to a run such that an image area with the image area dimension SR is covered, where SR = SF * 2 / (N+2) applies for the image area dimension SR. The light pattern with the light pattern dimension SP, with SP ≥ SF * 2 / (N+2), thus covers the image area in any case and can be as large as the image area or extend beyond it. In addition, the image area is offset in each subsequent image of a respective set by SF * 1 / (N+2) along the direction associated with the line-by-line readout of the pixel rows. The image area dimension SR extends analogously from a first (e.g., upper) image area edge to a second (e.g., lower) image area edge. Thus, the image area moves successively in the projected images of a set from the first (e.g. upper) image edge to the second (e.g.lower) image edge, whereby in the first image the first image area edge borders on the first image edge and in the last image the second image area edge borders on the second image edge.

[0023] In this embodiment, in at least one set in one to a maximum of N-1 consecutive images, ending with the (N+1')th image, in addition to the one light pattern belonging to the pass, one or more further light patterns belonging to subsequent passes are projected. The one or more further light patterns are projected between the one light pattern and the first edge of the image. The one or more further light patterns are directly adjacent to the one light pattern. If two or more further light patterns are projected, the further light patterns are projected closer to the one light pattern the closer in time a further pass follows.

[0024] Thus, by exposing a portion of the image sensor's image area with another light pattern while another portion of the image area is still exposed to the same light pattern, thus interleaving the exposures with different light patterns, it is advantageously possible to create digital 3D representations of objects within a very short period of time while still maintaining high image quality. This is a major advantage of this design.

[0025] In general, each set contains at least one image that belongs to at least one other set. Therefore, images can also belong to two or more sets or runs. A set of images can contain two or more images that belong to one or more other sets.

[0026] Preferably, the light patterns in the N+1 images of a respective scan are projected such that the light pattern dimension SP is: SP = SF * 2 / (N+2). In other words, the light pattern dimension SP corresponds to the image area dimension SR, so that the light pattern and the image area are of equal size. This enables particularly simple control of the 3D scanner.

[0027] According to one embodiment of the invention, for which independent protection is claimed, in the at least one scanning unit, the frame rate of the projection unit corresponds to twice the frame rate of the at least one camera, wherein a set of three images is projected onto the surface of the object per pass, wherein the light patterns are projected such that the light pattern dimension SP is: SP ≥ SF * 1 / 2. In addition, a light pattern is projected into each image of a respective set such that an image area with the image area dimension SR is covered, where SR = SF * 1 / 2. In this case, the image area in each subsequent image of a respective set is offset by SF * 1 / 4 along the direction associated with the line-by-line readout of the pixel rows. Furthermore, in at least one set, in the third image, another light pattern belonging to an (immediately) subsequent pass is projected.The further light pattern is projected between the one light pattern and the first (e.g. upper) edge of the image.

[0028] Accordingly, a set of three images is projected onto the surface of the object for each pass of the image sensor, each set consisting of a first image, a second image, and a third image. In a third image of a respective pass, a light pattern with half the size of the third image, based on the direction assigned to the line-by-line readout of the pixel rows, is projected into the second (e.g., lower) half of the third image, while a different, further light pattern with half the size of the third image is projected into the first (e.g., upper) half of the third image. The third image of a respective pass is the first image of an immediately subsequent pass.

[0029] Preferably, in the immediately above embodiment, a set of three images is projected onto the surface of the object per pass, with the light patterns being projected such that the light pattern dimension SP is: SP = SF * 1 / 2. This enables particularly simple control of the 3D scanner.

[0030] Alternatively, in every second image of a set, the light pattern dimension SP and the image dimension SF are the same, and the light pattern extends completely over the second image.

[0031] Preferably, in a first pass for exposing the image sensor, a light pattern with a light pattern dimension SP, where SP = SF * 1 / 2, is projected into a first half of the first image, while no light pattern is projected into a second half of the first image, wherein the division into the first image and the second image is related to the direction associated with the line-by-line readout of the pixel lines.

[0032] Preferably, in a final pass for exposing the image sensor, a light pattern having a light pattern dimension SP, where SP = SF * 1 / 2, is projected into a second half of the first image, while no light pattern is projected into a first half of the first image, the division into the first image and the second image being related to the direction associated with the line-by-line readout of the pixel lines.

[0033] Preferably, during each pass, a central pixel row of the image sensor is exposed only with the second image.

[0034] The light pattern can basically be designed in any way and have a regular sequence of light and dark areas, with a striped light pattern being preferred.

[0035] According to a further embodiment of the invention, for which independent protection is claimed, the 3D scanner has an object holder for two or more object holders, each of which serves to place an object. Preferably, the object holders are each rotatable about one or more rotation axes. Preferably, the control unit is configured so that the surfaces of two or more objects are optically scanned simultaneously.

[0036] According to a further embodiment of the invention, for which independent protection is claimed, the object holders each have a memory chip with stored data about the object holder. In the case of an object holder arranged on the object holder, the memory chip is connected to the control unit for data transmission. The control unit is configured so that the data from the memory chip is read (preferably automatically).

[0037] According to a further embodiment of the invention, for which independent protection is claimed, the projection unit and the one or more cameras are synchronized such that an exposure time of the image sensor corresponds to a multiple of an RGB cycle of the projection unit.

[0038] Preferably, the 3D scanner is designed in the form of a desktop scanner for dental technology, whereby the surface of dental objects, such as the upper and lower jaw, can be optically scanned.

[0039] The various embodiments of the invention can be implemented individually or in any combination. In particular, the features mentioned above and explained below can be used not only in the specified combinations, but also in other combinations or on their own, without departing from the scope of the present invention.

[0040] The invention is explained in more detail below using exemplary embodiments, with reference to the accompanying figures. They show, in simplified form and not to scale: Fig. 1 a schematic representation of essential components of the 3D scanner, Fig. 2 a cross-sectional view of an embodiment of the 3D scanner in the form of a desktop scanner for dental technology, Fig. 3 a cross-sectional view of the 3D scanner of Figure 2 with external computer, Fig. 4 a cross-sectional view of the 3D scanner of Figure 2 with an articulator, Fig. 5A a schematic perspective view of the object recording of the 3D scanner of Figure 2 , Fig. 5B A cross-sectional view of the object holder of the 3D scanner of Figure 2 , Fig. 6 a time-location diagram of an embodiment for optical scanning in the 3D scanner, Fig. 7 the time-location diagram of Figure 6with exposure and readout times, Fig. 8 a time-location diagram of another embodiment for optical scanning in the 3D scanner.

[0041] First, be Figure 1 considered, in which essential components of the 3D scanner, which is designated overall by the reference number 1, are illustrated by means of a schematic representation.

[0042] Accordingly, the 3D scanner 1 comprises a control unit 2 for controlling and monitoring the various functions of the 3D scanner 1 by means of a computer program, as well as at least one scanning unit 3, which is connected to the control unit 2 via data transmission. The scanning unit 3 comprises a projection unit 4 for projecting images (frames) containing light patterns, preferably striped light patterns. The control unit 2 serves in particular to control and monitor the projection unit 4 and the two cameras 5. Although in Figure 1two cameras 5 are illustrated by way of example, it is understood that equally only a single camera 5 or more than two cameras 5 can be provided for the at least one scanning unit 3.

[0043] The projection unit 4 serves to project an image with a light pattern onto the surface 7 of an object 8, preferably a dental object, using a light beam 6. The light beam 6 is partially reflected by the surface 7 of the object 8, wherein the reflected light beam 6' is captured by the two cameras 5 at an angle other than zero to the projection direction. The light pattern is preferably a striped light pattern with a periodic arrangement of bright lines and dark spaces in between (striped light projection). Using known projection geometry, such as triangulation, the spatial position of each individual point on the surface 7 of the object 8 can be calculated from the reflected light beam 6', so that a digital 3D image of the object 8 can be generated computationally.

[0044] In order to capture the entire surface 7 of the object 8, the object 8 and the scanning unit 3 can be moved relative to each other, e.g., by rotating the object 8 around one or more rotation axes, so that further images can be recorded, from which the 3D coordinates for other surface areas of the object 8 can then be derived. This is known to the person skilled in the art, so it need not be discussed in detail here. Figure 1 Not shown is an object holder for storing and positioning the object 8 relative to the scanning unit 3.

[0045] The cameras 5 each have an image sensor 9 with a plurality of pixels 10 (see insert in Figure 1 ). The pixels 10 are arranged in a two-dimensional array 13 in pixel rows 11 and pixel rows 12. In the schematic representation in Figure 1Each pixel row 11 contains six pixels 10, and the array 13 has ten pixel rows 11. It should be understood that in reality, the number of pixels 10 per pixel row 11 and the number of pixel rows 11 in an array 13 is significantly higher. The pixel rows 11 are arranged one above the other (stacked) in a direction (y) perpendicular to the row direction (x), thus forming the array 13.

[0046] The cameras 5 are each set up so that the image sensor 9, which is a CMOS image sensor, for example, is exposed and read out in rolling shutter mode. Here, the pixel rows 1 are exposed and read out line by line in a staggered manner, i.e. with a time offset. The process begins with the exposure of a first pixel row 11 of the array 13, followed by a later start of the exposure of the immediately adjacent second pixel row 11, and so on, until the exposure of the last pixel row 11 of the array 13 begins. The readout of a pixel row 11 takes place immediately after the end of the exposure of the pixel row 11. The single exposure and readout of all pixel rows 11 corresponds to one pass.In a subsequent pass, exposure of the first pixel row 11 of the 2D array 13 is started again, followed by a later start of exposure of the immediately adjacent second pixel row 11, and so on, until exposure of the last pixel row 11 of the 2D array is started. Preferably, but not necessarily, the time offset of the start of exposure of two immediately adjacent pixel rows 11 corresponds to the time period for reading out one pixel row 11. This is done in conjunction with the . Figures 6 and 7 explained in more detail.

[0047] In the following, a design of the 3D scanner is presented by Figure 1 in the form of a desktop scanner for the dental sector. In Figure 2is an example of a 3D scanner 1 for the dental field in the form of a desktop scanner, illustrated schematically in a (vertical) cross-sectional view. The 3D scanner 1 comprises a base housing 14, here for example C-shaped, which surrounds an interior area 15 that is open at the front and sides. The C-shaped base housing 14 is composed of a flat base section 16, a flat cover section 17 and a flat rear wall 18 connecting the base and cover sections. If the 3D scanner 1 is located on a horizontal surface (e.g. on a table), the base and cover sections 16, 17 are aligned horizontally, while the rear wall 18 is aligned vertically. The interior area 15 is freely accessible from the front (i.e. opposite the rear wall 18) and from both sides in order to load the 3D scanner 1 with dental objects to be optically scanned.

[0048] Attached to the front of the rear wall 18 is an object holder 19, to which one or more object holders 20, here, for example, two object holders 20, can be attached. The object holder 19 is, for example, plate-shaped. The two object holders 20 each serve to store a dental object to be scanned and, for this purpose, each have a plate 21 on the top of which a dental object can be placed. The object holder 19 is located entirely within the interior area 15.

[0049] On the underside of the cover section 17, two separate scanning units 3 are arranged, corresponding to the number of object holders 20, wherein one scanning unit 3 is assigned to one object holder 20 and the other scanning unit 3 is assigned to the other object holder 20. Each scanning unit 3 comprises a projection unit 4 and here, for example, four cameras 5, which in Figure 2not shown in detail. By means of the projection unit 4 of a respective scanning unit 3, images containing light patterns, preferably striped light patterns, can be projected onto the surface of a dental object placed on the associated object holder 20. The light beam emitted by the projection unit 4 is partially reflected by the surface of the dental object, and the reflected light beam is captured by the cameras 5 of the associated scanning unit 3 at an angle other than zero to the projection direction, so that a digital 3D representation of the dental object can be computationally generated. The projection units 4 have, for example, commercially available white light LEDs as light sources.

[0050] In order to capture the entire surface of the dental objects located on the object holders 20, the dental objects can be moved relative to the scanning units 3. For this purpose, the object holder 19 is coupled to a turntable arranged on the rear wall 18, such that the object holder 19 can be rotated about a rotation axis perpendicular to the rear wall 18 (e.g., horizontal). In addition, the object holders 20 can be rotated about a rotation axis directed perpendicular to the object holder 19. Thus, a dental object placed on an object holder 20 can be moved in space about two mutually perpendicular axes of rotation, such that the entire surface of the dental object can be optically scanned by the associated scanning unit 3. The rotational movements can be controlled via stepper motors, which enable reliable, precise, and fast adjustment of various rotational positions.

[0051] In the Figure 2The illustrated 3D scanner 1 includes two object holders 20, each with an associated scanning unit 3. It is understood that the 3D scanner 1 can have a smaller or larger number of object holders 20, each with an associated scanning unit 3. Two object holders 20 with associated scanning units 3 advantageously enable the simultaneous optical scanning of the surface of two dental objects for the computational determination of digital 3D images.

[0052] In Figure 3 is the 3D scanner 1 of Figure 2in a typical working mode with two dental objects 22. One dental object 22 is located on one object holder 20, the other dental object 22 on the other object holder 20. The dental objects 20 here are, for example, three-dimensional models of an upper and lower jaw, which were formed, for example, using an impression material. It is understood that other dental objects can also be scanned.

[0053] As in Figure 3 As illustrated, the 3D scanner 1 is connected to an external computer 23 for data purposes. For this purpose, the 3D scanner 1 has 18 electrical connections (e.g., USB interface) on the back of the rear panel. The computer 23 contains the Figure 1The illustrated control unit 2 is implemented as a logic module, by which the two scanning units 3 and the stepper motors for the relative positioning of the dental objects 22 can be controlled in order to computationally generate digital 3D representations of the two dental objects 22. This is shown in Figure 3illustrates where a monitor of the computer 23 displays the associated digital 3D representations of the two dental objects 22. The two dental objects 22 can be optically scanned simultaneously, which provides a significant time advantage compared to sequential optical scanning. For this purpose, the two dental objects 22 are preferably moved simultaneously and in the same way, which provides control-related advantages. The external computer 23 is programmatically configured to control the functions of the 3D scanner 1, for which purpose suitable software is executed on the computer 23. It would be conceivable to provide the control unit 2 in the 3D scanner 1 itself, so that an external computer 23 is not required.

[0054] In Figure 4The 3D scanner 1 is shown in another operating mode, in which the object holder 19 is provided with an articulator 24, which is optically scanned by a scanning unit 3. As is known to those skilled in the art, an articulator can simulate the diverse chewing movements of the upper and lower jaws, which is essential for the fabrication of a dental restoration. The articulator 24 contains a jaw model 25 for the upper and lower jaws. Using the articulator 24, individually adjusted articulators can be transferred to the scanning software in an anatomically correct manner. Precise settings of the articulator 24 can be measured and converted into a 3D model. This only needs to be done once per articulator.The upper and lower jaws can then be scanned as dental objects 22 using the calibrated object holders 20. The precise alignment of the dental objects 22 is based on the coordinates of the respective object holder 20 in conjunction with the coordinates of the articulator 24. The upper and lower jaws can then be mathematically brought into occlusion. This is a particular advantage of the 3D scanner 1.

[0055] In the Figures 5A and 5B are details of the object holder 19 and the two object holders 20 of the 3D scanner 1 of the Figures 2 to 4illustrated. Accordingly, the plate-shaped object holder 19 has a mating connector 26 for a connector 27 of the object holder 20 at the object holder position 28 provided for this purpose. When an object holder 20 is placed in an object holder position 28 on the object holder 19, the pins of the connector 27, which protrude from the underside of the object holder 20, come into electrical contact with the pin receptacles of the mating connector 26, thus establishing an electrical plug connection. The mating connectors 26 are connected to the control unit 2 of the 3D scanner 1 for data transmission. The object holders 20 are each provided with a memory chip 29 (e.g. EEPROM) in which data about the object holder 20 is permanently stored. The memory chip 29 is connected to the connector 27 via a data line 30.

[0056] If an electrical connection is established between the plug 27 of an object holder 20 and the corresponding mating plug 26, data stored in the memory module 29 can be transferred to the control unit 2 of the 3D scanner 1, with data transfer advantageously starting automatically as soon as contact is established. This advantageously enables automatic recognition of the respective object holder 20 by storing a code for the object holder 20, with specific calibration data of the object holder 20 preferably also being transferred, which in particular enables a computationally determined occlusion of the upper and lower jaw. This enables the automatic recognition of different object holders 20, along with the associated data of the object holder 20. In particular, this also allows an automatic scanning process to be started. Working with the 3D scanner 1 can thus be considerably simplified.

[0057] Reference is now made to Figure 6 taken, in which an embodiment for optical scanning in the 3D scanner 1 is illustrated using a time-location diagram. In Figure 7 are for the time-location diagram of Figure 6 with exposure and readout times of the image sensor 9. In the upper part of the Figures 6 and 7 From left to right, a temporal sequence of projected images is shown, which are projected by the projection unit 4 onto the surface of an object. The images are used to project various striped light patterns for optical two-dimensional scanning of the surface of the object to determine a digital 3D image of the object. In the lower part of Figure 6 The pixel lines 11 with pixels 10 of the image sensor 9 are illustrated. In the lower part of Figure 7The exposure and readout times of pixel rows 11 are indicated. The representation from left to right corresponds to the temporal sequence.

[0058] As already mentioned in Figure 1 As shown, the image sensor 9 comprises a plurality of pixels 10 arranged in pixel rows 11. In the schematic representation of Figure 6 five pixels 10 per pixel line 11 and nine pixel lines 11 are shown. As in Figure 7 As stated, the image sensor 9 in reality comprises, for example, 3120 pixel lines, each of which contains a large number of pixels.

[0059] In this embodiment, it is assumed that the projection unit 4 projects consecutive images at a frame rate of 50 fps. Furthermore, it is assumed that the camera can capture images at a (maximum) frame rate of 25 fps. A frame rate of 50 fps corresponds to 50 images per second, meaning that each image is projected by the projection unit onto the surface of the object for a duration of 20 ms (milliseconds). Therefore, the frame rate of the projection unit 4 is twice the frame rate of the camera 5.

[0060] The pixel rows 11 of the image sensor 9 are exposed in a staggered manner in rolling shutter mode and read out line by line, starting with the exposure of the first pixel row 11, followed by a later start of the exposure of the immediately following second pixel row 11 and so on until the exposure of the last pixel row 11 begins. This is shown in Figure 6This is schematically illustrated by the fact that each pixel row 11 located below another pixel row 11 is offset by one pixel 10 to the right, corresponding to a later start of exposure. When the image sensor 9 is viewed vertically, the first pixel row 11 is the uppermost pixel row, i.e., the image sensor 9 is exposed line by line from top to bottom, with a later start of exposure for each pixel row 11 immediately below it. The single exposure of all pixel rows corresponds to one pass.

[0061] In accordance with the frame rate of the projection unit 4, the exposure time of the pixel rows 11 is 20 ms each, i.e. each pixel row 11 is exposed for a duration of 20 ms. This can be controlled in a conventional manner, e.g. via an electronic shutter, which can be assumed to be known in the art. The readout time of a pixel row 11 here is, for example, 10 µm (microseconds). The exposure of each immediately subsequent pixel row 11 starts later, for example, by the duration of the readout time of a pixel row 11, i.e. the start of the exposure of two immediately adjacent pixel rows 11 is separated in time by 10 µs. With a total of 3120 pixel rows 11, it follows that the total readout time of all pixel rows 11 of the image sensor 9 in one pass is approximately 31 ms. With an exposure time of 20 ms per pixel line 11, it follows that the single exposure and readout of all pixel lines 11, ieA single scan through all 11 pixel rows requires a time span of 51 ms. The camera has a maximum frame rate of 25 fps, corresponding to a minimum cycle time per image of 40 ms.

[0062] In Figure 7 The start and duration of the exposure of the respective pixel rows 11 are indicated, with the start of the exposure of an immediately following pixel row 11 being shifted by 10 µs. It is evident that there is a temporal overlap of the exposure of pixel rows 11. The readout of a pixel row 11 occurs immediately after the end of the exposure of pixel row 11. For graphic reasons, the readout time of 10 µs for a pixel row 11 is shown greatly enlarged compared to its exposure time of 20 ms.

[0063] As in the Figures 6 and 7As can be seen, for a single exposure of all pixel rows 11 of the image sensor 9 in rolling shutter mode, i.e., per pass, three (individual) images or frames are projected sequentially from the projection unit 4 onto the surface of the object. In other words, the single exposure of all pixel rows 11 per pass is achieved using a set of three images. Accordingly, a set of images consists of three images that serve to expose all pixel rows 11 of the image sensor 9 line by line during one pass.

[0064] The three images of a same set are projected in a time span of 60 ms, while the exposure and readout of all pixel rows 11 per pass takes a time span of 51 ms. In the illustrated embodiment, the start of the exposure of the first pixel row 11 is not synchronized with the start of the projection of the first image of the three images of a same set, but occurs later, under the proviso that no pixel row 11, in particular not the middle pixel row 11, is exposed with two different stripe light patterns. The end of the exposure of the last pixel row occurs before the end of the projection of the third image of the same set. This is possible because there is a time difference of 9 ms between the projection of the three images of a same set and the exposure and readout of all pixel rows in one pass.

[0065] In the following, it is assumed that the assignment of the images to the image area of ​​the image sensor always remains unchanged. Corresponding to the image area of ​​the image sensor 9 defined by the array 13, the images projected by the projection unit 4 are each in the form of a rectangle or square. Each projected image has a first direction (x') and a second direction (y') perpendicular thereto, wherein the first direction (x') of the projected image is assigned to the first direction (x) of the image sensor 9 and the second direction (y') of the projected image is assigned to the second direction (x) of the image sensor 9. As already explained at the beginning, each projected image is assigned to the image area of ​​the image sensor 9 uniquely (i.e. 1:1).

[0066] A time-staggered exposure of the pixel lines 11 along the second direction (y) of the image sensor 9 is effected by light reflected from respective line-shaped image areas of a projected image arranged along the second direction (y') of the image.

[0067] The images always have the same image dimension SF along the second direction (y'), whereby the image dimension SF is measured from a first image edge to a second image edge along the second direction (y'). When viewed vertically, the first image edge can be understood as the upper edge and the second image edge as the lower image edge. The image dimension SF is measured in a direction associated with the line-by-line readout of the pixel rows 11. Correspondingly, in a direction associated with the line-by-line readout of the pixel rows 11, image regions of an image have an image region dimension SR and light patterns have a light pattern dimension SP. In the direction perpendicular to this (x'), i.e. in the direction of the pixel rows 11, the dimensions of the light patterns are each selected such that all pixels 10 of the same pixel row 11 are exposed.

[0068] The pixel rows 11 of the image sensor 9 are assigned line-shaped areas of the images 1:1 in an analogous manner. The pixels 10 of the image sensor 9 are exposed to only a single stripe light pattern per pass. This takes advantage of the fact that the start of exposure of the pixel rows 11 is staggered in time. The stripe light pattern is projected into different areas of the three images of a set according to the staggered start of exposure of the pixel rows 11. This is done in conjunction with the Figures 6 and 7 explained in more detail using three different stripe light patterns, labeled "Pattern 1", "Pattern 2" and "Pattern 3".

[0069] Pattern 1 is assigned three images 1A, 1B, 1C, which together form a first set of images. The three images 1A, 1B, 1C are used to expose all pixel rows 11 of the image sensor 9 with Pattern 1 in one pass. Pattern 2 is assigned three images 2A, 2B, 2C, which together form a second set of images. The three images 2A, 2B, 2C are used to expose all pixel rows 11 of the image sensor 9 with Pattern 2 in a further (immediately subsequent) pass. Image 2A is identical to Image 1C, i.e. Image 1C belongs to both the first set and the second set. Pattern 3 is assigned three images 3A, 3B, 3C, which together form a third set of images. The three images 3A, 3B, and 3C are used to expose all pixel rows 11 of the image sensor 9 with pattern 3 in a further (immediately subsequent) pass. Image 2C is identical to image 3A, i.e., image 2C belongs to both the second and third sets.In general, with the exception of the second (middle) images of a set, an image also belongs to the immediately adjacent set, unless the set is temporally marginal, ie the images belong to the first set or last set of images that are projected.

[0070] In the Figures 6 and 7 For simplicity, three different striped light patterns are shown: one temporally inner striped light pattern (Pattern 2) and two temporally peripheral striped light patterns (Pattern 1, Pattern 3). It is possible to project more than three striped light patterns; in this case, the number of temporally inner striped light patterns increases.

[0071] The design of the projected images for exposing the image sensor 9 with Pattern 1, Pattern 2 and Pattern 3 is explained in more detail below, with particular reference being made to the image dimension SF, image area dimension SR and light pattern dimension SP, each measured in a direction associated with the line-by-line readout of the pixel rows 11.

[0072] In general, a set of three images is projected onto the surface of the object per pass, with the light patterns being projected such that the light pattern dimension SP is SP ≥ SF * 1 / 2, i.e. the light pattern dimension SP is at least half the image dimension SF. A light pattern is projected into each image of a respective set such that an image area with the image area dimension SR is covered, where SR = SF * 1 / 2. The image area in each subsequent image of a respective set is offset by SF * 1 / 4 along the direction assigned to the line-by-line readout of the pixel rows. Furthermore, in at least one set, another light pattern belonging to an (immediately) subsequent pass is projected in the third image. The further light pattern is projected between the one light pattern and the first (e.g. top) image edge.

[0073] Each set comprises three images. The first set comprises images 1A, 1B, 1C, the second set comprises images 2A, 2B, 2C, and the third set comprises images 3A, 3B, 3C. For each image, an image area is defined, which is offset in each subsequent image of a respective set by SF * 1 / 4, i.e., one-quarter of the image dimension along the direction associated with the line-by-line readout of the pixel rows. Figure 6 On the left side of the images, the shifting of the image area in a respective set is indicated. The image area dimension SR of the three images in a set is indicated by SR1 for the first image, SR2 for the second image, and SR3 for the third image. Figure 6 On the right side of the images are the image dimension SF and, as an example for image 3C, the light pattern dimension SP. Exposure of pixel rows 11 with pattern 1:

[0074] Pattern 1 is projected onto the first or upper half of Image 1A, corresponding to the image area with the image area dimension SR1. Preferably, no striped light pattern is projected into the second or lower half of Image 1A. The image area dimension SR (here SR1) corresponds to the light area dimension SP. In Image 1B, Pattern 1 is projected over the entire area, i.e. Pattern 1 extends over the entire Image 1B. Therefore, the light pattern dimension SP is larger than the image area dimension SR (here SR2). It is important that Pattern 1 is projected into the central image area of ​​Image 1B, as indicated by the image area dimension SR2. Accordingly, Pattern 1 could alternatively be projected only into this image area, whereby up to a quarter of the area at the upper and / or lower image edge of Image 1B can be free of the projection of Pattern 1.The image sensor 9 is thus exposed to pattern 1 from images 1A and 1B in a temporally staggered manner up to the middle pixel row 11 and read out line by line. The middle pixel row 11 is only exposed to pattern 1 from image 1B. Pattern 2 is projected onto the first or upper half of image 1C, while pattern 1 is projected onto the second or lower half of image 1C. Thus, the lower half of the image sensor 9 is exposed to pattern 1 from images 1B and 1C in a temporally staggered manner and read out line by line. In addition, the upper half of the image sensor 9 is exposed to pattern 2 from image 1C in a temporally staggered manner. Exposure of the pixel lines with Pattern 2:

[0075] Pattern 2 is projected onto the first or upper half of Image 2A, corresponding to the image area with the image area dimension SR1. For Pattern 2, the image area dimension SR (here SR1) corresponds to the light area dimension SP. Pattern 1 is projected into the second or lower half of Image 2A. For Pattern 1, the image area dimension SR (here SR3) corresponds to the light area dimension SP. In Image 2B, Pattern 2 is projected over the entire area, i.e. Pattern 2 extends over the entire Image 2B. The light pattern dimension SP is larger than the image area dimension SR (here SR2). It is important that Pattern 2 is projected into the central image area of ​​Image 2B, as indicated by the image area dimension SR2. Accordingly, Pattern 2 could alternatively be projected only into this image area, whereby up to a quarter of the area at the top and / or bottom edge of Image 2B can be free from the projection of Pattern 2.The image sensor 9 is thus exposed to pattern 2 of images 2A and 2B in a temporally staggered manner up to the middle pixel row 11 and read out line by line. The middle pixel row 11 is only exposed to pattern 2 of image 2B. Pattern 3 is projected onto the first or upper half of image 2C, while pattern 2 is projected onto the second or lower half of image 2C. Thus, the lower half of the image sensor 9 is exposed to pattern 2 of images 2B and 2C in a temporally staggered manner and read out line by line. In addition, the upper half of the image sensor 9 is exposed to pattern 3 of image 2C in a temporally staggered manner. Exposure of the pixel lines with Pattern 3:

[0076] Pattern 3 is projected onto the first or upper half of Image 3A, corresponding to the image area with the image area dimension SR1. For Pattern 3, the image area dimension SR (here SR1) corresponds to the light area dimension SP. Pattern 2 is projected into the second or lower half of Image 3A. For Pattern 2, the image area dimension SR (here SR3) corresponds to the light area dimension SP. In Image 3B, Pattern 3 is projected over the entire area, i.e. Pattern 3 extends over the entire Image 3B. The light pattern dimension SP is larger than the image area dimension SR (here SR2). It is important that Pattern 3 is projected into the central image area of ​​Image 3B, as indicated by the image area dimension SR2. Accordingly, Pattern 3 could alternatively be projected only into this image area, whereby up to a quarter of the area at the top and / or bottom edge of Image 3B can be free from the projection of Pattern 2.The image sensor 9 is thus temporally staggered and read out line by line up to the middle pixel row 11 with Pattern 3 of images 3A and 3B. The middle pixel row 11 is only exposed with Pattern 3 of image 3B. Pattern 3 is projected onto the second or lower half of image 3C, while no light pattern is projected onto the first or upper half of image 2C. Thus, the lower half of the image sensor 9 is temporally staggered and read out line by line with Pattern 3 of images 3B and 3C.

[0077] If more than three striped light patterns are provided, with the pixel rows always being exposed to the same striped light pattern during a scan, the second set of images (which is temporally inner) is repeated one or more times. The two sets of images at the outermost points in time remain unchanged.

[0078] By projecting two light patterns into one image, the camera can advantageously capture images much faster, making better use of its maximum frame rate. This is a major advantage of the 3D scanner, as it allows digital 3D representations of objects to be generated very quickly.

[0079] In Figure 8A schematic representation of an embodiment of optical scanning in the 3D scanner 1 according to the invention is illustrated using a time-location diagram. The projection unit 4 projects a red image, a green image, and a blue image sequentially to create the impression of a white image, which is illustrated by the three images R, G, and B. This results in the so-called "rainbow effect," since the colors do not arrive on the object at the same time. Since the cameras 5 can only read the individual pixel rows 11 with a temporal offset in rolling shutter mode, a rainbow effect is created because different pixel rows 11 record different parts of the red, green, and blue image.

[0080] To solve this problem, the projection unit 4 and the one or more cameras 5 are synchronized, with the exposure time of the image sensor 4 being set to a precise multiple of the RGB cycle of the projection unit 4. Due to the precise timing, each pixel row 11 receives the same exposure time for the red, green, and blue images, enabling a true-color recording without color distortion.

[0081] From the above it follows that the invention provides an improved 3D scanner with which digital 3D representations of objects can be generated in a time-efficient manner and with high image quality. List of reference symbols

[0082] 13D scanner 2Control unit 3Scan unit 4Projection unit 5Camera 6, 6'Light beam 7Surface 8Object 9Image sensor 10Pixel 11Pixel row 12Pixel row 13Array 14Base housing 15Interior 16Floor section 17Cover section 18Back wall 19Object holder 20Object holder 21Plate 22Dental object 23Computer 24Articulator 25Jaw model 26Mating connector 27Connector 28Object holder position 29Memory module 30Data line

Claims

1. 3D scanner (1) for optically scanning the surface (7) of an object (8), comprising at least one scanning unit (3) which has a projection unit (4) for projecting images containing light patterns onto the object (8) and at least one camera (5) for recording the images, wherein an image frequency of the projection unit (4) corresponds to N times an image frequency of the at least one camera (5), wherein N is a natural number greater than 1, wherein each camera (5) has an image sensor (9) with a plurality of lines of pixels (11) arranged in a row, wherein the pixel lines (11) are exposed in respective passes in sequence in a time-delayed manner and read out line-by-line, wherein, in a direction associated with the line-by-line readout of the pixel lines (11), images have an image dimension SF, image areas have an image area dimension SR, and light patterns have a light pattern dimension SP, and a control unit (2) for controlling the at least one scanning unit (3), characterized in that the control unit (2) is configured such that - per run, a set of N+1 images (1A, 1B, 1C; 2A, 2B, 2C; 3A, 3B, 3C) is projected onto the surface (7) of the object (8), wherein the light patterns are projected such that the following applies to the light pattern dimension SP: SP ≥ SF * 2 / (N+2), - wherein a light pattern is projected into each image (1A, 1B, 1C; 2A, 2B, 2C; 3A, 3B, 3C) of a respective set in such a way that an image area with the image area dimension SR is covered, wherein SR = SF * 2 / (N+2), wherein the image area in each subsequent image (1B, 1C; 2B, 2C; 3B, 3C) of a respective set is offset by SF * 1 / (N+2) along the direction associated with the line-by-line readout of the pixel lines, and - wherein in at least one set (1A, 1B, 1C; 2A, 2B, 2C), in one to a maximum of N-1 successive images (1C, 2C), ending with the (N+1')th image, one or more further light patterns belonging to subsequent runs are respectively projected.

2. 3D scanner (1) according to claim 1, in which the light patterns are projected in the N+1 images (1A, 1B, 1C; 2A, 2B, 2C; 3A, 3B, 3C) of a respective run such that the following applies to the light pattern dimension SP: SP = SF * 2 / (N+2).

3. 3D scanner (1) according to one of claims 1 or 2, wherein in the at least one scanning unit (2), the frame rate of the projection unit (4) corresponds to twice the frame rate of the at least one camera (5), wherein - per run, a set of three images (1A, 1B, 1C; 2A, 2B, 2C; 3A, 3B, 3C) is projected onto the surface (7) of the object (8), wherein the light patterns are projected such that the following applies to the light pattern dimension SP: SP ≥ SF * 1 / 2, - wherein a light pattern is projected into each image (1A, 1B, 1C; 2A, 2B, 2C; 3A, 3B, 3C) of a respective set such that an image area with the image area dimension SR is covered, where SR = SF * 1 / 2, wherein the image area in each subsequent image (1B, 1C; 2B, 2C; 3B, 3C) of a respective set is shifted by SF * 1 / 4 along the direction associated with the line-by-line readout of the pixel lines, and - wherein in at least one set (1A, 1B, 1C; 2A, 2B, 2C), a further light pattern belonging to a subsequent run is projected in the third image.

4. 3D scanner (1) according to claim 3, in which one set of three images (1A, 1B, 1C; 2A, 2B, 2C; 3A, 3B, 3C) is projected onto the surface (7) of the object (8) per run, wherein the light patterns are projected such that the following applies to the light pattern dimension SP: SP = SF * 1 / 2 .

5. 3D scanner according to claim 3, wherein in every second image (1B, 2B, 2C) the light pattern dimension SP and the image dimension SF are equal.

6. 3D scanner according to one of claims 3 to 5, wherein, in a first run, a light pattern with a light pattern dimension SP, where SP = SF * 1 / 2, is projected into a first half of the first image (1A), while no light pattern is projected into a second half of the first image (1A), wherein the division into the first image and the second image is related to the direction associated with the line-by-line readout of the pixel lines (11).

7. 3D scanner according to any one of claims 3 to 6, wherein, in a last run, a light pattern with a light pattern dimension SP, where SP = SF * 1 / 2, is projected into a second half of the first image (3C), while no light pattern is projected into a first half of the first image (1A), wherein the division into the first image and the second image is related to the direction associated with the line-by-line readout of the pixel lines (11).

8. 3D scanner according to one of claims 3 to 7, in which, in a respective run, a middle pixel line (11) of the image sensor (9) is exposed only with the second image (1B, 2B, 3B).

9. 3D scanner according to one of claims 1 to 8, which has an object holder (19) with two or more object holders (20), each object holder (20) serving to place an object (8).

10. 3D scanner according to claim 9, wherein the object holders (20) are each rotatable about one or more axes of rotation.

11. 3D scanner according to claim 9 or 10, wherein the control unit (2) is configured such that the surfaces of two or more objects (8) are optically scanned simultaneously.

12. 3D scanner according to any of claims 9 to 11, wherein the object holders (20) each have a memory module (29) with stored data about the object holder (20), wherein, in the case of an object holder (20) arranged on the object support (19), the memory module (2) is connected to the control unit (2) in terms of data technology, wherein the control unit (2) is configured such that the data of the memory module (29) is read out.

13. 3D scanner according to one of claims 1 to 12, wherein the projection unit (4) and the at least one camera (5) are synchronized such that an exposure time of the image sensor (4) corresponds to a multiple of an RGB cycle of the projection unit (4).

14. 3D scanner according to one of claims 1 to 13, which is designed in the form of a desktop scanner for dental technology.