Virtual image canceling method and 3D model generation device using the same

Abstract

To provide a virtual image canceling method which does not generate a virtual image(s) caused by a transparent material forming a transparent mounting stand while photographing the whole circumference of a photographic target in a state where the photographic target is mounted on the mounting stand, and a 3D model generation device using the same.SOLUTION: A 3D model generation device comprises: a transparent mounting stand 100 on which a photographic target TG is mounted; a photographing unit 200 for photographing the photographic target TG; a gantry mechanism 300, while rotating the photographing unit 200 with the transparent mounting stand 100 as a center, for moving the same also in a rotation axis direction; a virtual image canceling unit 400 for deleting a virtual image(s) caused by deflection owing to the mounting stand 100 from virtual image-including images VI photographed via the transparent mounting stand among a plurality of images photographed by the photographing unit 200 to obtain a non-virtual image-including image(s) NVI; and a 3D model generation unit 500 for generating a 3D model from the non-virtual image-including image(s) NVI and direct images DI photographed without the transparent mounting stand 100.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention provides a virtual image canceling method that can eliminate virtual images caused by refraction by a transparent mounting base when creating a 3D model of a three-dimensional object to be photographed mounted on the mounting base; It also relates to a 3D model generation device that can use it to generate 3D models without virtual images. [Background technology]

[0002] Photogrammetry is a technology that can create an ultra-high precision 3D model of a photographed object. Specifically, it is a technology that photographs the object from all angles, analyzes and synthesizes the photographic data, and creates a three-dimensional 3D model. This photogrammetry technology is also used to create VR videos and 3DCG virtual spaces, known as the Metaverse. For example, there is a technique described in Japanese Patent Application Laid-Open No. 2023-184317.

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2023-184317

[0004] Photogrammetry technology is used not only in VR videos, but also in the world of cultural properties. For example, some artifacts excavated from ruins are fragile and easily crumbled, and many cannot be directly handled or viewed, so photogrammetry technology is effective for such artifacts. For example, excavated mokkan (long, thin wooden strips used for writing with ink) can be made into 3D models using photogrammetry technology.

[0005] However, if a wooden tablet is placed on a turntable and rotated while photographing the entire circumference, it is not possible to photograph the back side, and there is also a risk that the wooden tablet will collapse due to the rotation of the turntable. Furthermore, even if one tried to turn the tablet over to photograph the back side, there was a risk that the tablet would collapse, making it difficult to actually create a 3D model of the entire surface. This is similar to the difficulty of using photogrammetry to create a 3D model of a standing or seated Buddha statue, as it is not possible to photograph the soles of the feet that are in contact with the base or the underside of the seat, making it difficult to create a 3D model of the entire body.

[0006] Therefore, the inventors of the present invention devised a device that photographs the entire surface of a wooden tablet and creates a 3D model using photogrammetry technology. This device has a transparent mounting base on which the wooden tablet, which is the object to be photographed, is placed, a photographing unit that rotates around the mounting base and moves laterally to photograph the object to be photographed, a moving unit that moves the photographing unit laterally while rotating around the mounting base, and a 3D image generating unit that generates a 3D image of the object to be photographed from multiple images taken by the photographing unit.

[0007] With such a device, it was thought that the surface in contact with the table (the back of the wooden tablet) could also be photographed through the transparent table, making it possible to create a 3D model of the entire surface. DISCLOSURE OF THE INVENTION [Problem to be solved by the invention]

[0008] However, it has been found that the above-mentioned device has the following problems. That is, a false image occurs in the vertical direction of the object to be photographed. Such a virtual image is caused by the fact that the table on which the object to be photographed is placed is made of a transparent material such as glass or acrylic. Because the back side of the object being photographed (the surface in contact with the mounting base) is photographed through the transparent mounting base, refraction occurs due to the transparent material that makes up the mounting base, resulting in an image being captured in an incorrect position (shifted upwards).When a 3D model is created based on this image, a false image is created in the vertical direction. The occurrence of this vertical false image is not limited to cultural assets such as wooden tablets, but is a major drawback of any device that photographs an object placed on a transparent stand. In addition, for parts that are photographed without going through the mounting table, i.e., parts that are not in contact with the mounting table (the top and side surfaces of the object to be photographed), there is no problem of refraction due to the mounting table, so no virtual images are generated.

[0009] The present invention was devised in consideration of the above circumstances, and aims to provide a virtual image canceling method that can photograph the entire circumference of an object to be photographed while the object is placed on a transparent mounting base, without generating virtual images due to the transparent material that makes up the mounting base, and a 3D model generation device that uses the same. [Means for solving the problem]

[0010] The virtual image canceling method of the present invention is designed to form a virtual image-free image from which the portion equivalent to the virtual image has been removed by shifting the virtual image captured through a transparent mounting table by an amount equivalent to the virtual image caused by refraction by the mounting table.

[0011] In addition, the shift is calculated from a non-refracting marking photographed without using a transparent mounting table and a refraction marking photographed through the transparent mounting table, among a plurality of markings formed on the same straight line.

[0012] The 3D model generation device of the present invention includes a transparent mounting base on which an object to be photographed is placed, an imaging unit that photographs the object to be photographed, a gantry mechanism that rotates the imaging unit around the transparent mounting base while also moving it in the direction of the rotation axis, a virtual image canceling unit that removes virtual images caused by refraction by the mounting base from virtual image images captured through the transparent mounting base among multiple images captured by the imaging unit to create virtual image-free images, and a 3D model generation unit that generates a 3D model from the virtual image-free image and a direct image captured without passing through the transparent mounting base.

[0013] The mounting table has at least three markings formed on the upper surface thereof, which are photographable by a photographing unit from the back side of the mounting table, and which are formed in the same straight line. At least one of the markings can be photographed by the photographing unit without going through the mounting table. The virtual image canceling unit calculates the amount of refraction caused by the mounting table from the non-refracted marking photographed without going through the mounting table and the refracted marking photographed through the mounting table, shifts the position on the refracted image of the object photographed through the mounting table according to the calculated amount of refraction, and forms a non-virtual image. The direct image and the non-virtual image are output to a 3D model generation unit.

[0014] The virtual image canceling unit generates a first mask that masks the area that is photographed without going through the mounting table, and a second mask that masks the area that is photographed through the mounting table, overlays the first masking and the second masking to extract only the direct shooting area, which is the area that can be photographed without going through the mounting table, synthesizes the virtual image-free image and the direct shooting area to generate an integrated image, and outputs the integrated image to the 3D model generation unit. [Effects of the Invention]

[0015] The virtual image canceling method of the present invention can remove the virtual image equivalent by shifting the virtual image captured through a transparent mounting table by an amount equivalent to the virtual image caused by refraction by the mounting table, thereby making it possible to generate a 3D model with the object to be photographed placed on the transparent mounting table. This method has the advantage of being particularly useful for generating 3D models of the entire circumference of objects that are very fragile and easily crumbled, such as excavated artifacts such as wooden tablets or standing or wooden Buddha statues.

[0016] Furthermore, since the 3D model generation device of the present invention uses the refraction canceling method to remove virtual images caused by refraction due to a transparent mounting material, it is equipped with a transparent mounting base on which the object to be photographed is placed, a photographing unit that photographs the object to be photographed, a gantry mechanism that rotates the photographing unit around the transparent mounting base while also moving it in the direction of the rotation axis, a virtual image canceling unit that removes virtual images caused by refraction due to refraction due to the mounting base from virtual image images photographed through the transparent mounting base among multiple images photographed by the photographing unit to create virtual image-free images, and a 3D model generation unit that generates a 3D model from the virtual image-free image and direct images photographed without using the transparent mounting base.This makes it possible to generate 3D models of the entire circumference of objects that are particularly fragile and easily broken, such as excavated artifacts such as wooden tablets and standing or wooden Buddha statues. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a conceptual diagram of a 3D model generation device according to an embodiment of the present invention. [Figure 2] 1A and 1B are diagrams of a 3D model generation device according to an embodiment of the present invention, in which FIG. 1A is a schematic front view, FIG. 1B is a schematic left side view, FIG. 1C is a schematic perspective view from the front and left sides, and FIG. 1D is a schematic perspective view from the front and right sides. [Figure 3] 1 is a schematic perspective view of a mounting table that constitutes a 3D model generation device according to an embodiment of the present invention. [Figure 4] 3 is a conceptual diagram illustrating the relationship between an opening of a mounting table, a fiducial marking, and a reference marking that constitutes a 3D model generation device according to an embodiment of the present invention. FIG. [Figure 5] FIG. 1 shows an explanatory diagram (left column) showing the direction in which an object is photographed in a 3D model generation device according to an embodiment of the present invention, a mounting base viewed from that direction and the object being photographed as seen through it (middle column), and a conceptual diagram (right column) showing the mounting base viewed from that direction and the object being photographed as seen through the mounting base when removed. [Figure 6]FIG. 2 is a diagram for explaining processing by a virtual image canceling unit in the 3D model generating device according to the embodiment of the present invention, and is a conceptual diagram showing the relationship between a fiducial marking and a reference marking. [Figure 7] FIG. 10 is a diagram for explaining processing by a virtual image canceling unit in a 3D model generating device according to an embodiment of the present invention, and is a conceptual diagram showing the relationship between a fiducial marking and a reference marking that has shifted due to refraction by the mounting table. [Figure 8] FIG. 10 is a diagram for explaining processing by a virtual image canceling unit in a 3D model generating device according to an embodiment of the present invention, and is a conceptual diagram showing calculation of the amount of refraction by the mounting table from a fiducial marking and a reference marking that has shifted due to refraction by the mounting table. [Figure 9] FIG. 2 is a conceptual diagram illustrating a first mask, illustrating processing by a virtual image canceling unit in the 3D model generating device according to the embodiment of the present invention. [Figure 10] FIG. 2 is a diagram for explaining processing by a virtual image canceling unit in the 3D model generating device according to the embodiment of the present invention, and is a conceptual diagram showing an image extracted by a first mask. [Figure 11] FIG. 10 is a diagram for explaining processing by a virtual image canceling unit in a 3D model generating device according to an embodiment of the present invention, and is a conceptual diagram showing an image obtained by shifting an image extracted by a first mask by an amount corresponding to the amount of refraction. [Figure 12] FIG. 3 is a conceptual diagram illustrating a second mask, illustrating processing by a virtual image canceling unit in the 3D model generating device according to the embodiment of the present invention. [Figure 13] FIG. 2 is a diagram for explaining processing by a virtual image canceling unit in the 3D model generating device according to the embodiment of the present invention, and is a conceptual diagram showing an image extracted by a second mask. [Figure 14] FIG. 10 is a diagram for explaining processing by a virtual image canceling unit in a 3D model generating device according to an embodiment of the present invention, and is a conceptual diagram showing an integrated image obtained by integrating an image extracted by a first mask and an image extracted by a second mask. [Figure 15] 4 is a flowchart of processing by a virtual image canceling unit in the 3D model generating device according to the embodiment of the present invention. [Figure 16] 1 is a 3D model image of a wooden tablet as a photographed object, from which a virtual image has been removed by a 3D model generation device according to an embodiment of the present invention. [Figure 17] 1 is a 3D model image of a wooden tablet as a photographed object, from which a virtual image has been removed by a 3D model generation device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0018] A 3D model generation device 1000 according to an embodiment of the present invention includes a transparent mounting table 100 on which a photographing object TG is placed, an imaging unit 200 that photographs the photographing object TG, a gantry mechanism 300 that rotates the imaging unit 200 around the transparent mounting table 100 while also moving it in the direction of the rotation axis, a virtual image canceling unit 400 that removes a virtual image caused by refraction by the mounting table 100 from a virtual image VI, which is a virtual image captured through the transparent mounting table among multiple images captured by the imaging unit 200, to generate a virtual image-free image NVI, and a 3D model generation unit 500 that generates a 3D model from the virtual image-free image NVI and a direct image DI captured without passing through the transparent mounting table 100.

[0019] The imaging unit 200, the gantry mechanism 300, the virtual image canceling unit 400, and the 3D model generating unit 500 are controlled by a control unit 900.

[0020] The mounting table 100 is made of a transparent material such as flat glass or an acrylic plate, and is set to a size that allows the photographing target TG to be placed thereon. 3, the mounting table 100 is formed long and narrow, with five protruding pieces 110 formed on a pair of opposing long sides. These protruding pieces 110 are feature points in generating a 3D model, and are formed at equal intervals on the top and back surfaces of the mounting table 100. In addition, openings 120 are formed in the mounting table 100 adjacent to the protruding pieces 110. Of these openings 120, only one is provided in the protruding pieces 110 at both ends, and one is provided on each side of the remaining three protruding pieces 110. These openings 120 are the areas through which the fiducial marking STM, which will be described later, can be seen. The mounting table 110 is set horizontally.

[0021] The above-mentioned mounting table 100 has a fiducial marking STM and a reference marking RFM on its upper surface. 3 is shown as a hole opened in the mounting table 100, but in FIG. 4 the opening 120 is shown as a notch for convenience of drawing. When the back side of the object TG is photographed, that is, through the mounting table 100, the fiducial markings STM and the reference markings RFM are photographed together with the object TG. The fiducial marking STM and the reference marking RFM are formed on the same straight line, and the fiducial marking STM is visible through the opening 120 when the mounting table 100 is viewed from the back side, i.e., without passing through the mounting table 100. Furthermore, the reference marking RFM is visible through the mounting table 100 when the mounting table 100 is viewed from the back side.

[0022] In theory, one fiducial marking STM and one reference marking RFM are sufficient, but having multiple marks has the advantage of being able to accommodate larger objects to be photographed TG and generating more accurate 3D models. Furthermore, the protruding pieces 110 also act as feature points in 3D model generation (photogrammetry), which has the advantage of enabling more accurate 3D model generation.

[0023] The photographing unit 200 uses a digital camera capable of taking digital images. The photographing section 200 is attached to a camera mounting section 310 of a gantry mechanism 300, which will be described later, and is adapted to photograph a photographing target TG placed on a mounting table 100 from all directions.

[0024] Of the images captured by this imaging unit 200, the direct image DI captured without going through the transparent mounting table 100 is output to the 3D model generation unit 500, and the virtual image VI captured through the transparent mounting table is output to the virtual image canceling unit 400. In other words, an image taken from above the mounting table 100 is considered to be a direct image DI taken without going through the mounting table 100, and an image taken from below the mounting table 100 is considered to be a virtual image VI taken through the mounting table 100.

[0025] The gantry mechanism 300 has a camera mounting section 310 to which the imaging section 200 is attached, a rotation mechanism 320 that rotates the camera mounting section 310 around the mounting base 100, and a lateral movement mechanism that moves the camera mounting section 310 along the length direction of the mounting base 100 (the lateral direction in Figure 2(A)). The camera mounting section 310 also has a tilt mechanism for tilting the photographing section 200 .

[0026] The rotation mechanism 320 has two circular ring members 321, a plurality of support poles 322 (six in the drawing) connecting the ring members 321, and a drive mechanism 323 that rotates the ring members 321. 2, the two ring members 321 are set on a support (not shown) so that the line connecting their centers is horizontal and parallel to the mounting table 100. Of course, the ring members 321 are set on the support so that they can rotate. One of the ring members 321 has a drive belt 323A that constitutes a drive mechanism 323 looped around it, and is rotated by a drive motor 323B that constitutes the drive mechanism 323. The mounting table 100 is placed on a straight line connecting the centers of the two ring members 321 . As will be described later, since the imaging unit 200 is moved intermittently, it is desirable to use a stepping motor as the drive motor 323B.

[0027] Furthermore, the support poles 322 not only connect the two ring members 321 but also serve as guide rails for the lateral movement mechanism, and the camera mounting section 310 is movably mounted between adjacent support poles 322.

[0028] The lateral movement mechanism moves the camera mount 310 along the support pole 322 . Unlike the rotation mechanism 320, the lateral movement mechanism is designed to manually move the camera mounting portion 310 along the support pole 322. However, it is also possible to configure it so that it moves automatically using a motor or the like.

[0029] The imaging unit 200 is moved by the gantry mechanism 300 configured as described above as follows. A camera mounting part 310 to which a digital camera is attached is set at an initial position, and the camera is rotated around the table 100 by a rotation mechanism 320, and an object to be photographed TG placed on the table 100 is photographed from all directions. In Figure 5, it is explained that the object to be photographed TG is photographed from eight positions, namely, the 0:00 position (see Figure 5(A)), the 10:30 position (see Figure 5(B)), the 9:00 position (see Figure 5(C)), the 7:30 position (see Figure 5(D)), the 6:00 position (see Figure 5(E)), the 4:30 position (see Figure 5(F)), the 3:00 position (see Figure 5(G)), and the 1:30 position (see Figure 5(H)), as viewed counterclockwise from the ring member 321 side (left side of the 3D model generation device 100). However, there is no problem with photographing from positions other than these as long as the photographs are taken at equal intervals.

[0030] When photographing from the 0:00 position (see Figure 5(A)), the subject TG is photographed from directly above, when photographing from the 10:30 position (see Figure 5(B)), the subject TG is photographed from 45° above and to the left, when photographing from the 9:00 position (see Figure 5(C)), the subject TG is photographed from directly to the left, when photographing from the 7:30 position (see Figure 5(D)), the subject TG is photographed from 45° below and to the left, when photographing from the 6:00 position (see Figure 5(E)), the subject TG is photographed from directly below, when photographing from the 4:30 position (see Figure 5(F)), the subject TG is photographed from 45° below and to the right, when photographing from the 3:00 position (see Figure 5(G)), the subject TG is photographed from directly to the right, and when photographing from the 1:30 position (see Figure 5(H)), the subject TG is photographed from 45° above and to the right.

[0031] That is, when photographing from the 0:00 position (see Figure 5(A)), the 10:30 position (see Figure 5(B)), the 9:00 position (see Figure 5(C)), the 3:00 position (see Figure 5(G)), or the 1:30 position (see Figure 5(H)), a direct image DI of the object TG is captured without passing through the mounting table 100.

[0032] On the other hand, when photographing from the 7:30 position (see Figure 5(D)), the 6:00 position (see Figure 5(E)), or the 4:30 position (see Figure 5(F)), a virtual image VI of the object TG is photographed through the mounting table 100.

[0033] The photographing unit 200 is tilted using the tilt mechanism at each of the eight positions described above to photograph a plurality of images.

[0034] When the imaging of the entire circumference is completed and the imaging unit 200 returns to the initial position of 0:00, the imaging unit 200 is moved a certain distance from the initial position along the support pole 322 using the lateral movement mechanism. The number of horizontal movements is determined by the length of the object GT. That is, a long object TG requires more horizontal movements, and a short object TG requires fewer horizontal movements. Furthermore, this horizontal movement distance should be set small if a more detailed image is desired.

[0035] The direct image DI captured of the object TG without going through the mounting table 100 does not have a virtual image due to refraction by the mounting table 100, and is therefore output directly to the 3D model generation unit 500 without going through the virtual image canceling unit 400.

[0036] On the other hand, an image of the object TG photographed through the mounting table 100 becomes a virtual image VI accompanied by a virtual image due to refraction by the mounting table 100, and is therefore output to the virtual image canceling unit 400 rather than the 3D model generation unit 500.

[0037] In the virtual image canceling unit 400, the virtual image VI is subjected to the following processing to remove the virtual image equivalent to the virtual image, thereby generating a virtual image-free image NVI, and an integrated image ITI is generated (see FIGS. 6 to 15). As shown in FIG. 6, the virtual image VI includes not only the object TG to be photographed but also the stage 100, the support 800 supporting the stage 100, and the like. First, among the multiple markings included in the virtual image VI, the fiducial marking STM seen through the opening 120 of the mounting table 100 is compared with the two reference markings RFM adjacent to the fiducial marking STM.

[0038] The reference marking STM seen through the opening 120 is not refracted by the mounting table 100, and therefore is a marking without refraction even in the virtual image VI. In contrast, the reference markings RFM on both sides of the fiducial marking STM are refracted by the mounting table 100, and therefore appear as refracted markings in the virtual image VI.

[0039] Here, the two reference markings RFM, which are markings with two refractions, and the fiducial marking STM, which is the basis for one marking without refraction, are originally formed on the same straight line, and the distance between one fiducial marking STM and the two reference markings RFM is set to be the same, so the fiducial marking STM (marking without refraction) is located at the midpoint between the two reference markings RFM (markings with refraction) (see Figure 7).

[0040] However, the marking with two refractions is displaced from its original position due to refraction caused by the mounting table 100. For this reason, the marking with two refractions is connected by a straight line L, and the deviation between the midpoint MP of this line L and the reference marking STM, which is a marking with no refraction, is calculated (see FIG. 8). The deviation indicated by the arrow A is equal to the deviation caused by refraction due to the mounting table 100, and corresponds to the amount by which the virtual image VI is shifted (shift amount) (see S1 in FIG. 15).

[0041] The virtual image canceling unit 400 not only calculates the shift amount for shifting the virtual image VI, but also performs the following operations. This virtual image canceling unit 400 generates a first mask M1 that masks only the area captured without passing through the mounting table 100, i.e., the virtual image VI, from the back side of the mounting table 100 (see S2 in Figures 9 and 15).

[0042] As shown in FIG. 6, this first mask M1 masks the opening 120 and the support columns 800 that support the stage 100. Next, as shown in FIG. 10, this first mask M1 is superimposed on the virtual image VI, and only the portion visible through the mounting table 100 is extracted to form a first mask image (see S3 in FIG. 15). The virtual image VI extracted by this first mask M1 is shifted by an amount corresponding to the shift amount (see S5 in FIGS. 11 and 15). The shift amount is indicated by the arrows A on the right and left ends.

[0043] Furthermore, the virtual image canceling unit 400 generates a second mask M2 that masks the area photographed through the mounting table 100 (see S1 in FIGS. 12 and 15).

[0044] This second mask M2 is equivalent to the inverse of the first mask M1, as shown in FIG. As shown in FIG. 13, this second mask M2 masks the portion where the mounting table 100 is located and which is not visible due to the support 800, etc., to generate a second mask image (see S4 in FIG. 15).

[0045] The virtual image VI extracted by the first mask M1 is shifted by an amount corresponding to the shift amount to obtain a virtual image NVI without a virtual image, and the second mask image masked by the second mask M2 is integrated to synthesize an integrated image ITI as shown in Figure 14 (see S6 in Figure 15). This integrated image ITI is one in which the virtual image caused by refraction by the mounting table 100 has been removed.

[0046] In this way, the integrated image ITI from which the virtual image has been deleted by the virtual image canceling unit 400 is output to the 3D model generation unit 500. This 3D model generation unit 500 has photogrammetry software that performs photogrammetry processing, and is able to generate a 3D model by performing photogrammetry processing on multiple input images.

[0047] The 3D model generation unit 500 outputs a direct image DI obtained by photographing the object to be photographed TG without using the mounting table 100, and an integrated image ITI obtained by processing the virtual image VI obtained by photographing the object to be photographed TG via the mounting table 100 in the virtual image canceling unit 400 to remove the virtual image, and the 3D model generation unit 500 generates and outputs a 3D model of the object to be photographed TG using both images (direct image DI and integrated image ITI) that do not contain virtual images. Therefore, the image output from the 3D model generating unit 500 is a 3D model that does not have any virtual images caused by refraction of the mounting table 100 when viewed from all directions. This 3D model is output to a display and can be viewed from any direction.

[0048] FIG. 16 shows an example of a 3D model generated by a 3D model generation device according to an embodiment of the present invention using a wooden tablet as the object TG to be photographed. When the 3D model shown in Figure 16 is compared with a 3D model generated by a conventional device shown in Figure 17 (a 3D model generation device without a virtual image canceling unit 400), it can be seen that the one shown in Figure 17 has a thick virtual image reflected on the underside of the wooden tablet (the surface in contact with the mounting base 100), while the one shown in Figure 16 is a 3D image of the wooden tablet with the part corresponding to the virtual image removed.

[0049] In other words, in a conventional 3D model generating device that does not have the above-mentioned virtual image canceling unit 400, the virtual image VI is output as is to the 3D model generating unit 500, and a 3D model is generated in which a virtual image is attached to the surface that is in contact with the mounting table 100, i.e., the back side of the object to be photographed TG, as shown in Figure 17.

[0050] In contrast, in the case of a 3D model generation device according to this embodiment, an integrated image ITI generated from a virtual image-free image NVI in which the virtual image canceling unit 400 has removed the virtual image from the virtual image VI is output to the 3D model generation unit 500, and a 3D model in which no virtual image is generated on the surface in contact with the mounting table 100, i.e., the back side of the object to be photographed TG, as shown in Figure 16, is generated. [Explanation of symbols]

[0051] 100 Mounting table 200 Photography Department 300 Gantry mechanism 400 Virtual image canceling unit 500 3D model generation unit

Claims

1. A virtual image canceling method characterized by forming a virtual image-free image from which the portion equivalent to the virtual image has been removed by shifting a virtual image captured through a transparent mounting table by an amount equivalent to the virtual image caused by refraction by the mounting table.

2. 2. The virtual image canceling method according to claim 1, wherein the shift is calculated from a non-refractive marking photographed without a transparent mounting base and a refraction marking photographed through a transparent mounting base among a plurality of markings formed on the same straight line.

3. A 3D model generation device comprising: a transparent mounting base on which an object to be photographed is placed; an imaging unit that photographs the object to be photographed; a gantry mechanism that rotates the imaging unit around the transparent mounting base while also moving it in the direction of the rotation axis; a virtual image canceling unit that removes virtual images caused by refraction by the mounting base from virtual image images photographed through the transparent mounting base among multiple images photographed by the imaging unit, thereby generating virtual image-free images; and a 3D model generation unit that generates a 3D model from the virtual image-free image and a direct image photographed without using the transparent mounting base.

4. The 3D model generation device described in claim 3, characterized in that at least three markings are formed on the upper surface of the mounting table and can be photographed by the photographing unit from the back side of the mounting table, on the same straight line, and at least one of the markings can be photographed by the photographing unit without going through the mounting table, and the virtual image canceling unit calculates the amount of refraction due to the mounting table from the non-refracted marking photographed without going through the mounting table and the refracted marking photographed through the mounting table, shifts the position on the refracted image of the photographed object photographed through the mounting table in accordance with the calculated amount of refraction, to form a virtual image-free image, and outputs the direct image and the virtual image-free image to the 3D model generation unit.

5. The 3D model generation device according to claim 3 or 4, characterized in that the virtual image canceling unit generates a first mask that masks the area that is photographed without using the mounting table and a second mask that masks the area that is photographed through the mounting table, superimposes the first masking and the second masking to extract only the direct photography area that can be photographed without using the mounting table, synthesizes the virtual image-free image and the direct photography area to generate an integrated image, and outputs the integrated image to the 3D model generation unit.

Images