Intraoral scanner, intraoral scanning system, intraoral scanning method and computer program product

Abstract

To provide an intraoral scanner, an intraoral scanning system, an intraoral scanning method and a computer program product. [Solution] An intraoral scanner (3) for intraoral stereo vision is proposed. The intraoral scanner (3) comprises a projector (4) for projecting a dot pattern (6) onto an imaging target, the projector (4) having a radiation light source and an optical system for generating the dot pattern (6) from radiation emitted from the radiation light source, and at least two cameras (7, 8) for imaging the pattern (6) projected onto an imaging target (2), the cameras (7, 8) being arranged to have overlapping fields of view on the imaging target. The dot pattern (6) is a non-periodic dot pattern generated by a two-dimensional microlens array arranged non-periodically on a substrate, the microlens array being the final beam shaping optical element before the imaging target in the radial direction. The intraoral scanner (3) can be used in an intraoral scanning system and a method for generating a three-dimensional model of a dentition (2) based on an intraoral scan.

Description

[Technical field]

[0001] The present invention relates to an intraoral scanner comprising: a projector for projecting a structured pattern onto an imaging target, the projector having a radiation light source and an optical system for generating the structured pattern from radiation emitted from the radiation light source; and at least two cameras for imaging the pattern projected onto the imaging target, the cameras being positioned to have overlapping fields of view on the imaging target.

[0002] The invention further relates to an intraoral scanning system, an intraoral scanning method, and a computer program product. [Background technology]

[0003] US Patent No. 5,399,633 discloses an intraoral scanner that uses a microlens array to generate a pattern, which is imaged onto the dentition by further optics. The intraoral scanning system is configured to generate a three-dimensional model of the dentition by triangulation. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2019 / 236934 Summary of the Invention [Problem to be solved by the invention]

[0005] A drawback of known intraoral scanners is the complexity of the multiple parts that must be assembled to manufacture the intraoral scanner.

[0006] Based on this related art, it is an object of the present invention to provide a reduced complexity intraoral scanner and related intraoral scanning system that is capable of generating a three-dimensional model of the dentition with at least as much accuracy as conventional intraoral scanners, and further to provide a corresponding method and computer program product. [Means for solving the problem]

[0007] This object is achieved by an intraoral scanner, an intraoral scanning system, a method and a computer program product with the features of the independent claims. Advantageous embodiments and refinements are specified in the dependent claims.

[0008] In the intraoral scanner, the pattern is a non-periodic dot pattern generated by a two-dimensional microlens array arranged non-periodically on a substrate, the microlens array being the final beam shaping optical element before the imaging object in the radial direction. Such an irregular arrangement of microlenses is particularly useful for measuring the surface of tissues in the oral cavity, since all of the radiation emitted by the radiation source is used to generate the dot pattern. Therefore, the dots of the pattern are clearly visible in the image captured by the camera. Since the microlenses are arranged at a high density, the dot pattern is also very dense, and the shape of the dentition can be determined with high accuracy. Since the lateral extent of the microlenses is small, the microlens array provides high resolution in the depth direction. Therefore, no additional beam shaping optics are required between the microlens array and the imaging object, and the complexity of the intraoral scanner is significantly reduced compared to conventional intraoral scanners.

[0009] It should be noted that in the context of this application, mirrors or flat windows are not considered beam-shaping optical elements.

[0010] Because the rays of the beam reaching the microlens are generally not collimated, the working distance at which the dot size is smallest is greater than the focal length of the microlens, where working distance and focal length are each measured from the bottom surface of the microlens at the surface of the substrate in the radial direction.

[0011] The sub-beams can be laterally deflected by at least one mirror arranged before and / or after the microlens array, with respect to the direction of radiation, depending on the requirements for the size of the dot pattern on the imaged object. If the mirror is arranged after the microlens array, the dot pattern can be larger than the dot pattern obtained if the mirror is arranged in front of the microlens array, but in the latter case the intraoral scanner becomes more compact.

[0012] The mirror can also deflect the incoming light towards the camera so that the field of view is enlarged according to the enlarged size of the dot pattern.

[0013] In one embodiment, the microlenses are arranged in a honeycomb structure, with the center of each microlens offset from the center of its corresponding honeycomb cell by a randomly selected distance and direction, thereby maximizing the density of microlenses that are approximately circular in outline.

[0014] The radiation source is typically a laser diode, and the divergence of the beam emitted by the laser diode is homogenized by an optical element disposed between the laser diode and the microlens array.

[0015] The cameras usually face the same direction because they can be easily mounted side-by-side on the same circuit board, and depth resolution is maximized when the cameras face the same direction.

[0016] The cameras are positioned next to each other so that their fields of view overlap sufficiently, and the center of the microlens array is positioned away from the baseline between the cameras, typically below the cameras.

[0017] The microlens array can be tilted at an angle α towards an axis perpendicular to the beam axis of the beam to equalize the distance between the microlenses and the surface of the imaged object.

[0018] The angle α is typically greater than half the angle β between the beam axis and the axis of the camera's field of view when viewed along the front surface of the substrate that forms the bottom surface of the microlens.

[0019] An intraoral scanning system for generating a three-dimensional model of the dentition using active stereo vision comprises an intraoral scanner as described above, where each camera generates a sequence of image patterns that are projected onto the dentition, and further comprises a data processing unit connected to the intraoral scanner for generating the three-dimensional model of the dentition based on a disparity map generated using the synchronized sequence of images generated by each camera.

[0020] The data processing unit of the intraoral scanning system is configured in particular to perform the following operations: - acquiring a sequence of synchronized images using at least two cameras; - generating a disparity map based on synchronized image pairs, each image being taken from a different camera; generating a depth map based on the disparity map; - Registering the depth map, - Generating a mesh of points defining the surface of the imaged dentition.

[0021] A corresponding method for generating a three-dimensional model of a dentition based on active stereo vision may include the following operations: - acquiring a series of images of the dentition from an intraoral scanner as described above; using a data processing unit connected to the intraoral scanner for generating a three-dimensional model of the dentition based on a disparity map generated using the sequence of images generated by each camera of the intraoral scanner.

[0022] The computer program product contains program code for implementing the method when the computer program product is run on a data processing apparatus.

[0023] Further advantages and characteristics of the invention will become apparent from the following description in which exemplary embodiments of the invention are explained in detail with reference to the drawings. [Brief description of the drawings]

[0024] [Figure 1] FIG. 1 illustrates an intraoral scanning system. [Diagram 2] FIG. 2 is a perspective view of some internal components of the intraoral scanner of FIG. [Diagram 3] FIG. 2 is a front perspective view of the projector and camera of the intraoral scanner. [Figure 4] FIG. 2 is a perspective view of the beam path of a radiated light beam emitted by a laser diode of a projector. [Diagram 5] FIG. 2 is a diagram showing the beam paths of light rays incident on a beam-forming microlens array, the beam paths being shown in the plane of the emitted light emitting area of ​​a laser diode. [Figure 6] FIG. 2 is a diagram showing the beam paths of light rays incident on a beam-forming microlens array, the beam paths being shown in a plane perpendicular to the plane of the emitted light emitting area of ​​the laser diode. [Figure 7] 1A and 1B are diagrams showing the outer shape and spatial arrangement of beam-forming microlenses. [Figure 8] FIG. 2 is a top view of the camera system of the intraoral scanner. [Figure 9] FIG. 2 is a side view of the projector and camera system of the intraoral scanner. [Figure 10] FIG. 1 is a diagram illustrating the basic principle of stereo vision. [Figure 11] FIG. 1 shows stereo vision at a working distance of 8 mm. [Figure 12] FIG. 1 shows stereo vision at a working distance of 10 mm. [Figure 13] FIG. 1 shows stereo vision at a working distance of 12 mm. [Figure 14] FIG. 1 shows stereo vision at a working distance of 14 mm. [Figure 15] FIG. 1 shows stereo vision at a working distance of 16 mm. [Figure 16] FIG. 1 shows stereo vision at a working distance of 18 mm. [Figure 17] 4 shows a flow diagram of the data processing performed by the data processing unit of the intraoral scanner. [Figure 18] FIG. [Figure 19] FIG. 2 is a diagram showing details of core data processing. [Figure 20] FIG. 13 shows details of the model generation performed during data processing. [Figure 21] FIG. 13 demonstrates the surface resolution obtained by the intraoral scanner. [Figure 22] FIG. 13 is a perspective view of a tip region of a modified embodiment. [Figure 23] FIG. 13 is a perspective view of a tip region of a further variation of the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] FIG. 1 shows an intraoral scanning system 1 that can be used to generate a three-dimensional model of the dentition 2. The intraoral scanning system 1 comprises an intraoral scanner 3 extending along a longitudinal axis L. A projector 4 arranged in the scanner 3 emits radiation 5 and projects a dot pattern 6 onto the dentition 2. The dot pattern 6 is synchronously imaged by a first camera 7 and a second camera 8. The operation of the projector 4, the first camera 7 and the second camera 8 may be controlled by a control unit 9 that may be arranged in a handle 10 of the scanner 3. The digital images generated by the first camera 7 and the second camera 8 are transferred to a data processing unit 11. The data processing unit 11 comprises a data bus 12, a processor 13 and storage means 14 for storing data such as digital images and computer programs that may be used for the processing of the digital images as described herein. The data processing device 11 may be provided with conventional input means 15 such as a keyboard and a mouse and conventional output means such as a display 16.

[0026] 2 shows a perspective view of some internal components of the intra-oral scanner 3. At the tip of the intra-oral scanner 3 a projector 4 is arranged. The projector 4 is provided with a radiation source consisting of a laser diode 17 which emits radiation 5 longitudinally towards the tip of the intra-oral scanner 3. The radiation 5 is deflected laterally by a mirror 18. The deflected radiation is incident on a microlens array 19 which may form a window through which the radiation 5 for the dot pattern 6 passes and exits the intra-oral scanner 3.

[0027] The laser diode 17 may typically operate in the power range of 20-100 mW and emit radiation in the wavelength range of 400-500 nm. It is known that the shorter the wavelength, the better the spatial precision. The laser diode 17 may emit radiation 5 at a wavelength in the range of 450 nm ± 5 nm or 405 nm ± 5 nm, or preferably at a wavelength of about 450 nm or 405 nm. Considering the sensitivity of the cameras 7 and 8, it may be advantageous to use a wavelength of 450 nm ± 5 nm, especially 450 nm. This is because the cameras 7 and 8 tend to be more sensitive at these wavelengths and the signal-to-noise ratio per spatial precision is greatest at this wavelength.

[0028] The projector 4 may further comprise a secondary radiation source for irradiating the dentition 2 with additional radiation, which is not shown in Figures 1 and 2. The secondary radiation source may be used for taking color photographs of the dentition 2 or for detecting caries, in particular as described in US 2006 / 0227216 A1, the xenon lamp being replaced by an LED emitting radiation in the near ultraviolet range.

[0029] The intraoral scanner 3 further comprises a base plate 20, on which the first camera 7 and the second camera 8 are mounted. The base plate may also comprise a control unit 9. The base plate 20 comprises a connector 21 for connecting the intraoral scanner 3 to a data processing unit 11.

[0030] 3 is a front perspective view of the projector 4 and two cameras 7 and 8. The projector 4 includes a laser diode 17. The projector 4 also includes a rod lens 22 that is used as a beam shaping optical system, as described below. A mirror 18 and a final microlens array 19 are provided behind the rod lens 22 along the path of the emitted light 5.

[0031] 4 shows the path of the emitted light 5 from the laser diode 17 to the surface 23 of the imaging target 24 onto which the dot pattern 6 is projected. The emitted light 5 leaves the laser diode 17 as an emitted light beam 25 having an elliptical cross section. The rod lens 22 widens the beam divergence of the emitted light beam 25 in one direction so that the beam divergence becomes more uniform in all directions. The emitted light beam 25, shaped by the rod lens 22, is incident on the mirror 18, which deflects the emitted light beam 25 laterally towards the microlens array 19.

[0032] The microlens array 19 is positioned parallel to the longitudinal axis L of the intraoral scanner 3. The dot pattern 6 is therefore projected laterally with respect to the longitudinal axis L of the scanner 3 such that it lies within a scanning region 26 defined by an overlapping region of a field of view 27 of the first camera 7 and a field of view 28 of the second camera 8, the fields of view 27 and 28 being also positioned laterally such that the scanning region 26 covers at least a portion of the dot pattern 6.

[0033] In Figure 4, the emitted light beam 25 is shown traveling along a beam axis 29. Figure 5 is a cross-sectional view of the emitted light beam 25 along the beam axis 29 in the plane of a emitted light emitting region 30 of the laser diode 17. This emitted light emitting region 30 may be the pn junction surface of the laser diode 17.

[0034] For simplicity, mirror 18 is omitted from FIG.

[0035] As can be seen from FIG. 5, the emitted light beam 25 comprises a number of sub-beams 31 formed by the microlenses 32 of the microlens array 19. The microlenses 32 are arranged on a substrate 33 that is transparent to the emitted light 5 emitted by the laser diode 17. The microlenses 32 are formed on a front surface 34 of the substrate 33, the front surface 34 being the surface facing the imaging target 24. The microlenses 32 of the microlens array 19 are plano-convex lenses, with their flat bottom surface arranged on the front surface 34 of the substrate 33 and their curved surface facing the imaging target 24. The radius of curvature is typically in the range of 2-3 mm, which corresponds to a focal length of 4-6 mm. The radius of curvature may for example be 2.5 mm, which corresponds to a focal length of 5 mm. The microlens array 19 has a total area of ​​20-70 mm. 2 The density of the microlenses 32 may be 1 cm 2 The number of particles per unit area may range from 1200 to 6000.

[0036] The microlenses 32 may be made of polymethylmethacrylate (PMMA) or polycarbonate (PC) since both exhibit suitable moldability. The radius of curvature depends on the focal length required and the type of material used to form the microlenses 32. The radius of curvature ranges from 2-3 mm and also varies depending on the material.

[0037] Each microlens 32 images the exit pupil of the laser diode 17 onto a separate area on the surface 23 of the object 24, thereby generating a dot pattern 6 on the surface 23 of the object 24. The projected image of the dot pattern 6 is focused onto an projection plane 35, defined as the location where each sub-beam 31 converges into a dot 36. This means that the lateral extent of the cross section of each sub-beam 31 is minimized. If all microlenses 32 have the same focal length, then the projection plane 35 is generally planar.

[0038] Each dot 36 of the dot pattern 6 on the projection surface 35 is an image of the exit pupil of the laser diode 17. Because the light rays incident on the microlens 32 in each sub-beam 31 are diverging rather than parallel, the focal length of the microlens 32 is less than the object distance D, where D is the distance along the beam axis 29 from the front surface 34 of the substrate 33 to the projection surface 37.

[0039] The only beam shaping optical element other than the microlens 32 is the rod lens 22 located between the microlens array 19 and the laser diode 17. In the plane of the radiation emission area 30, the rod lens 22 leaves the sub-beam 31 unchanged. Thus, in the plane of the radiation emission area 30, the rod lens 22 only has the effect of a flat window.

[0040] It should be noted that in the context of the present application, mirrors or flat windows are not considered as beam-shaping optical elements. Thus, in a variant of the embodiment, the microlens array 19 may be protected by an exit window located between the microlens array 19 and the imaging target 24. In a further variant of the embodiment, the emitted light beam 25 may be deflected laterally by a mirror located behind the microlens array 19 with respect to the direction of the emitted light 5. This mirror may be an additional mirror different from mirror 18 or may replace mirror 18.

[0041] Due to the small lateral extent of the microlenses 32, the direction of the edge rays 37 defining the shape of the sub-beams 31 on the object side of the microlens array 19 converges at a small angle. The cross section of the sub-beams 31 therefore varies slowly along the beam axis 29. As a result, the dot pattern 6 has a large depth resolution, which takes advantage of the fact that the intraoral scanner 3 is not always held at a certain distance from the dentition and the projection plane 35 generally does not coincide with the surface 23 of the object 24. Furthermore, the surface 23 of the object 24 generally does not constitute a plane.

[0042] The microlens array 19 has a further advantage that the radiation 5 emitted by the laser diode 17 is entirely used to generate the dot pattern 6 since no mask is required to absorb and block the radiation to generate the dot pattern 6 .

[0043] Figure 6 shows the beam paths of the sub-beams 31 associated with individual microlenses 32 in a plane perpendicular to the plane of the radiation emission area. As with Figure 5, the mirror 18 has been omitted for simplicity.

[0044] From FIG. 6 it can be seen that rod lens 30 increases the divergence of radiation beam 25 emitted from laser diode 17 so that beam 25 completely illuminates microlens array 19.

[0045] As mentioned above, the beam 25 of the laser diode 17 has an elliptical cross section because the beam divergence in the plane of the radiation emission region 30 of the laser diode 17 is greater than the beam divergence in a direction perpendicular to the plane of the radiation emission region 30. The rod lens 22 increases the beam divergence in the direction perpendicular to the plane of the radiation emission region 30 and homogenizes the divergence of the beam 25. Thus, the longitudinal axis of the rod lens 22 lies in the plane of the radiation emission region 30 of the laser diode 17.

[0046] It should be noted that the microlens 32 does not necessarily have to be located on the front surface 34 of the substrate 33. Instead, the microlens 32 may be located on the rear surface 38 of the substrate 33, opposite the surface facing the laser diode 17. In this case, the focal length of the microlens 32 must be shorter than the distance between the surface of the substrate 33 and the exit pupil of the laser diode 17, which is imaged onto the projection plane 35.

[0047] 7 is a top view of the microlens array 19. The microlenses 32 are arranged on a hexagonal grid forming a honeycomb structure 39, with hexagonal honeycomb boundaries 40 defining honeycomb cells 41.

[0048] In order to be able to identify sub-patterns of the projected pattern in the dot pattern 6, the microlenses 32 are arranged aperiodically. The center of each microlens 32 is linearly shifted by a randomly selected amount in a randomly selected direction, the amount of shift being smaller than the distance from the honeycomb center 42 to the honeycomb boundary 40 in the selected direction. The amount of shift is typically in the range of 10-20% of the distance from the honeycomb center 42 to the honeycomb boundary 40. The direction and amount of shift are exemplarily indicated by the displacement arrows 43. Thus, by calculating the correlation function between the digital image data and the sub-pattern of the dot pattern 6 retrieved, the position of the sub-pattern of the dot pattern 6 in the digital image can be determined. After the position of the retrieved sub-pattern has been identified, even the positions of the individual dots 36 can be determined.

[0049] The dimensions of the microlenses 32, in particular their thickness and radius of curvature, are selected so that the microlenses 32 abut one another and no gaps remain between the microlenses 32. Thus, within the honeycomb structure 39, the contours of the microlenses 32 at the outer surface of the microlens array 19 are generally spaced away from the front surface 34 of the substrate 33.

[0050] An advantage of arranging the microlenses 32 in a honeycomb structure 39 is that the honeycomb boundaries 40 of the microlenses 32 are approximately circular. Compared to a microlens array having microlenses arranged on a rectangular grid, lensing errors, which are most noticeable away from the center of the microlens, especially near the corners of the contour, are reduced. Additionally, the honeycomb structure 39 allows the density of the dots 36 in the dot pattern 6 to be maximized.

[0051] 8 shows a top view of cameras 7 and 8 and their respective fields of view 27 and 28. Both cameras 7 and 8 may be mounted on base plate 20 such that axes 44 and 45 of fields of view 27 and 28, respectively, point in the same direction. Both cameras 7 and 8 are provided with camera optics 46 that image an object 24 including a dot pattern 6 onto a sensor 47, which is typically a sensor chip that produces a digital image of the object 24 and the dot pattern 6.

[0052] The fields of view 27 and 28 of both cameras 7 and 8 have a rectangular shape with the long side of both fields of view 27 and 28 extending along a baseline 48 of length b that extends between the centers of the sensors 47 of the two cameras 7 and 8. The length b of the baseline 48 is selected such that the fields of view 27 and 28 of both cameras 7 and 8 overlap. The overlapping area of ​​both fields of view 27 and 28 is the scan area 26, which is typically 9-20 mm in width and 7.5-16 mm in height.

[0053] The cameras 7 and 8 may further be provided with a wavelength filter 49 disposed in the optical path of the emitted light coming from the imaged subject 24 and incident on the sensor 47. The wavelength filter 49 may, for example, be disposed before the camera optics 46. Such a wavelength filter may be useful for detecting dental caries or caries, as described in detail in U.S. Patent Application Publication No. 2006 / 0227216.

[0054] It is noted that the intraoral scanner 3 can be operated in two distinct operating modes. In a first operating mode, in which the above-mentioned secondary radiation source is switched on and the laser diode 17 is switched off, a color image of the dentition 2 can be taken. This operating mode can also be used to detect caries or caries in the dentition 2. In a second operating mode, the laser diode 17 is used to generate a three-dimensional model of the dentition 2.

[0055] The second operating mode for generating a three-dimensional model will now be described.

[0056] The microlens array 19 is positioned below and between the two cameras 7 and 8, as shown in FIG. 9, which shows a side view of the cameras 7 and 8 and the microlens array 19 of the projector 4. The microlens array 19 is positioned such that the projected dot pattern 6 is at least partially within the scanning area 26 of the cameras 7 and 8 over a wide range of working distances WD. The working distance WD is the distance between the entrance pupil of one of the cameras 7 and 8 and the scanning area 26 measured along the axis 44 or 45 of the field of view 27 or 28 of the camera 7 or 8. The microlens array 19 is tilted by an angle α in the range of 20° to 30° toward an axis perpendicular to the beam axis 29 of the beam 25 in order to equalize the distance between the microlenses 32 and the surface 23 of the imaged object 24. The angle α is typically greater than half the angle β between the beam axis 29 and one of the axes of the fields of view 44 and 45 when viewed along the front surface 34 of the substrate 33 that forms the bottom surface of the microlenses 32.

[0057] Figure 10 shows the basic principle of computer stereo vision. A point on the surface 23 of the object 24 is imaged on the sensor 47 of the first camera 7 at position x1 and on the sensor 47 of the second camera 8 at position x2. These points on the images are sometimes called homologous points in both digital images. The centers of both sensors 47 are separated by a baseline b. The centers of the entrance pupils formed by the camera optics 46 are also separated by a distance of length b. If the distance between the entrance pupil and the sensor surface is equal to f, the distance d between the position of the dot 36 and the entrance pupil can be calculated based on the principle of central projection as follows:

number

[0058] The sum of x1+x2 is the so-called disparity. The disparity is therefore inversely proportional to the distance d of the dots 36. Within the scanning area 26, a so-called disparity map can be generated, which shows the disparity of points on each image within the scanning area 26. The graphic representation of the disparity map is usually encoded in greyscale, with the grey level being inversely proportional to the disparity. This means that points on the surface 23 of the imaged object 24 at a smaller working distance WD appear brighter than points on the surface 23 of the object 24 at a larger working distance WD.

[0059] Figure 11 shows the spatial relationship between the projector 4 and the cameras 7 and 8 at a working distance of 8 mm. The top figure includes a left digital image of the dot pattern 6 taken by the first camera 7 and a right digital image taken by the second camera 8.

[0060] The bottom diagram shows the resulting disparity map calculated based on the scan area 26 where the left and right digital images overlap.

[0061] The following Figures 12 to 16 correspond to Figure 11, but for working distances of 10 mm, 12 mm, 14 mm, 16 mm, and 18 mm. The disparity maps show that the disparity becomes darker and the disparity value becomes smaller as the working distance increases. As the working distance increases, the disparity decreases, but is still large enough to determine the distance with sufficient accuracy.

[0062] The digital images captured by the two cameras 7 and 8 are processed according to the principle of active stereo vision (ASV), a method known per se in the art, in which structured light is projected onto the object 24 and the three-dimensional object 24 is reconstructed by determining the disparity within the overlapping scanning area 26 based on pairs of digital images captured by at least one pair of cameras placed at different positions.

[0063] Fig. 17 shows the main steps of a method for reconstructing a three-dimensional image based on digital images taken by a first camera 7 and a second camera 8. The method may be executed by a processor 13 of a data processing unit 11 based on digital images recorded by the cameras 7 and 8. In the following, these digital images are also called frames. The method starts with an image capture 50, which is a method step for obtaining a sequence of high-density true-colour digital images produced by the cameras 7 and 8. The digital images or frames are taken synchronously by the first camera 7 and the second camera 8. The frame rate is typically in the range of 50-80 frames / s.

[0064] In a subsequent pre-processing step 60, the frames are pre-processed. For example, the imaged tissues may be automatically classified. Image regions related to soft tissues are removed from the frames and the margins between the various imaged objects are automatically detected. These tasks may be performed by neural networks or other forms of artificial intelligence.

[0065] In the core method 70, information about the three-dimensional shape of the scanned dentition 2 is extracted from the images. For example, a depth map is generated based on the disparity map, and individual depth maps for different positions of the intraoral scanner 3 are combined to form a mesh that defines the surface of the dentition 2.

[0066] In a further model generation step 80, a three-dimensional model of the dentition 2 is generated.

[0067] In a visualization step 90 , the constructed three-dimensional model can finally be visualized on the display 16 of the data processing unit 11 .

[0068] In an optional post-processing step 100, the 3D model is edited and polished by the operator.

[0069] FIG. 18 shows further details of pre-processing 60.

[0070] Pre-processing begins with frame validation 61. In this step, a frame is analyzed to verify whether it contains the expected information content in order to ensure that it shows the oral cavity.

[0071] Frame validation 61 is followed by frame adjustment 62. In this step the frame is prepared for the next processing step: the frame may be rescaled, cropped, deskewed or equalized.

[0072] After frame alignment 62, a segmentation step 63 segments the tissue types imaged in the frame.

[0073] FIG. 19 further illustrates the core of this method.

[0074] As a first step, a correction 71 is performed. The correction 71 is applied to corresponding frames generated by the first camera 7 and the second camera 8 at the same time. Both frames are remapped so that the homologous points in the two frames are located on a horizon line. In the subsequent depth map generation 72, the homologous points in the two frames are matched and the disparity between them is stored in a disparity map. From the disparity map, a depth map of the scene is obtained according to the principles of stereo vision described above in relation to FIG. 10.

[0075] For each new pair of frames, the newly generated depth map is added to the previous one in registration 73 using a registration algorithm based on matching homologous 3D points of the depth maps to form a global point cloud.

[0076] The resulting point cloud is then subjected to progressive meshing 74. Progressive meshing 74 allows the gradually growing 3D scene to be visualized on the display 16 in real time.

[0077] FIG. 20 shows the subsequent model generation 80 in more detail.

[0078] In outlier removal 81, isolated and irrelevant points that do not belong to the surface of the scene are removed from the point cloud in order to reduce noise on the final mesh.

[0079] Segmentation and cluster removal 82 identifies groups of points that represent clusters that do not belong to the surface 32 and removes them from the point cloud to avoid them protruding from the final mesh.

[0080] The number of points in the point cloud continues to be reduced. In removing duplicates 82, vertices in the point cloud that have a Euclidean distance of zero are merged.

[0081] If necessary, further filters are applied: A spatial filter 84 can be applied, which maintains the same topology as the original point cloud, but includes only a small portion of the original points.

[0082] The filtered and greatly reduced point cloud is then used in a 3D mesh reconstruction 85, which based on the point positions and surface normals results in a so-called watertight mesh surface.

[0083] Mesh cropping 86 is then performed on the watertight mesh surface to limit the extent of the original point cloud and finally to remove isolated polygons.

[0084] Color mapping 87 may reassign color information contained in a color image of the scene to mesh surfaces throughout the mesh.

[0085] Finally, mesh smoothing 88 may be used to normalize the surface and reduce polygon noise.

[0086] 21 shows the results of a test performed with the intraoral scanning system described herein. The intraoral scanning system was tested by scanning a tooth surface at a working distance of 10 mm using a microlens array 19 having a total of 1800 microlenses 32.

[0087] The grey shading indicates the accuracy of the reconstructed 3D models. The average accuracy is in the range of 14 μm. The accuracy of the entire 3D model showed different values ​​according to a distribution with variance σ in the range of 13 μm.

[0088] The system and method described herein are based on the principles of active stereo vision (ASV). However, it should be noted that the images can also be evaluated based on the principles of structured light stereo (SLS). In this method, the difference between pairs of images is evaluated according to the principles of active stereo vision as described herein. Furthermore, the images of each camera are evaluated based on their deviation from a reference image at a reference distance. From these differences a depth map can be calculated. In the art, this additional evaluation method is also called triangulation. However, one problem with triangulation is the calibration of the reference image. Therefore, a significant amount of effort is required to obtain additional depth information by triangulation.

[0089] FIG. 22 shows a part of a modified embodiment of the intraoral scanner 3. FIG. 22 shows in particular the tip portion 203 of the handle 10, which is provided with a modified projector 204 having the same components as the projector 4, except that the mirror 18 is omitted. The projector 204 thereby projects a dot pattern 6 along the longitudinal axis L. The tip portion 203 is further provided with a first camera 207, a second camera 208, and a third camera 209 that are arranged around the projector 204 and can take color and / or grayscale images. The tip portion 203 may further be provided with an LED 210 for illuminating the dentition 2, for example with UV light. The projector 204, the first camera 207, the second camera 208, and the third camera 209 may use active stereo vision as described above, as well as passive stereo vision or triangulation. The tip portion 203 may be provided with a hollow tip 211. The distal end 212 of the tip 211 is provided with a mirror 218 that laterally deflects the beam 25. The tip 211 may be removable from the handle 10. The tip 211 is further autoclavable and may include a heating element and / or suitable coatings to inhibit fogging of the mirror 218.

[0090] The chip 211 allows the size of the dot pattern 6 to be increased due to the longer optical path and larger lateral spread of the dot pattern 6 perpendicular to the beam axis 29 than in the embodiment of Figures 1-16. The focal length of the microlens 32 used in the projector 204 must be correspondingly longer and the radius of curvature of the microlens 32 must be larger than in the embodiments shown in Figures 1-16. The embodiment of Figure 22 allows the optical path length and working distance WD to be extended to 50 mm.

[0091] Figure 23 shows a further variant of the embodiment of the intraoral scanner 3. In Figure 23 the distal end portion 203 of the handle 10 is again shown. The distal end portion 203 is covered by a sleeve-like tip 219 which can be detached from the distal end portion 203 of the handle 10. Compared to the embodiment shown in Figure 22, the embodiment of Figure 23 allows for a shorter optical path and a working distance WD of more than 20 mm.

[0092] The laser diode 17 and the secondary radiation source may be alternately activated to generate color images of the dentition 2 and corresponding images may be captured by the cameras 7 and 8. The images captured under illumination by the secondary radiation source may then be used for color mapping 87 to generate a three-dimensional model with a colored surface of the dentition 2.

[0093] The data processing unit 11 may be a computer having at least one physical or logical processor 13. The data processing unit 11 may also be implemented on multiple physical computers or may be an integrated device including a display 16.

[0094] It should also be noted that the method may be implemented by a computer program product comprising code for implementing the methods described herein, if the code is executed by a processor in the data processing unit 11 or some other device. In some embodiments, the computer program product may be code stored on a computer-readable data storage medium, such as a disk, a compact disk, or a digital versatile disk (DVD). In other embodiments, the computer program product may be code stored in a data storage unit on a server or server array. Such a data storage unit may be a hard disk or a hard disk array, etc. In a further embodiment, the data carrier may be an electrical carrier signal that can be used to transfer the code from the server and can be used to download the program code to the client.

[0095] Throughout the description and claims of this specification, the singular encompasses the plural unless the context otherwise requires. In particular, where the indefinite article is used, the specification is to be understood as contemplating the plural as well as the singular, unless the context otherwise requires.

[0096] It is to be understood that any configuration, integer, structure, compound or group described in connection with a particular aspect, embodiment, or example of the invention is applicable to any other aspect, embodiment, or example described herein, unless inconsistent.

Claims

1. A projector for projecting a structured pattern onto an imaging target, the projector including a radiation light source and an optical system for generating the structured pattern from radiation emitted from the radiation light source; at least two cameras for imaging the pattern projected onto the imaging target, the cameras being positioned to have overlapping fields of view on the imaging target; Equipped with the pattern is a non-periodic dot pattern generated by a two-dimensional microlens array non-periodically arranged on a substrate, The microlens array is the final beam shaping optical element before the imaging object in the radial direction. Intraoral scanner.

2. The rays of the beam reaching the microlens are not collimated. The intraoral scanner of claim 1 .

3. the beam includes sub-beams formed by microlenses, the sub-beams converging on the object side of the microlens array; An intraoral scanner according to claim 1 or 2.

4. a working distance at which the dot size is smallest is greater than a focal length of the microlens, wherein the working distance and the focal length are each measured from a bottom surface of the microlens at a surface of a substrate in a radial direction; An intraoral scanner according to any one of claims 1 to 3.

5. the sub-beams are laterally deflected in the radiation direction by at least one mirror arranged in front of and / or behind the microlens array; An intraoral scanner according to any one of claims 1 to 4.

6. The mirror deflects incident light towards the camera.

6. An intraoral scanner according to claim 5.

7. The microlenses are arranged in a honeycomb structure, with the center of each microlens being offset from the center of its corresponding honeycomb cell by a randomly selected distance and direction. An intraoral scanner according to any one of claims 1 to 6.

8. the radiation source is a laser diode, and the divergence of the beam emitted by the laser diode is homogenized by an optical element disposed between the laser diode and the microlens array; An intraoral scanner according to any one of claims 1 to 7.

9. The cameras are facing the same direction; An intraoral scanner according to any one of claims 1 to 8.

10. the center of the microlens array is positioned away from the baseline between the cameras; An intraoral scanner according to any one of claims 1 to 9.

11. the microlens array is tilted at an angle α towards an axis perpendicular to the beam axis of the beam; An intraoral scanner according to any one of claims 1 to 10.

12. The angle α is generally greater than half the angle β between the beam axis and the axis of the field of view of the camera, as viewed along the front surface of the substrate that forms the bottom surface of the microlens.

12. An intraoral scanner according to claim 11.

13. An intraoral scanner according to any one of claims 1 to 12, wherein each of the cameras generates a sequence of images of a pattern projected onto the dentition; a data processing unit connected to the intraoral scanner for generating a three-dimensional model of the dentition based on a disparity map generated using the synchronized series of images generated by each camera; An intraoral scanning system that generates a 3D model of the dentition based on active stereo vision.

14. The data processing unit performs the following operations: acquiring a series of synchronized images using at least two cameras; generating a disparity map based on synchronized image pairs, each image taken from a different camera; generating a depth map based on the disparity map; registering the depth map; generating a mesh of points defining a surface of the imaged dentition; configured to run, 14. An intraoral scanning system according to claim 13.

15. Obtaining a series of images of the dentition from an intraoral scanner according to any one of claims 1 to 12; using a data processing unit connected to said intraoral scanner for generating a three-dimensional model of the dentition based on a disparity map generated using said series of images generated by each camera; A method for generating a three-dimensional model of a dentition based on active stereo vision comprising:

16. 16. A method for implementing the method of claim 15, comprising: Computer program products.

Images