Shape measurement method and shape measurement device

Abstract

A shape measurement method according to the present invention comprises: a step for generating a Z-position lookup table (ZLT) in which the Z-position and a setting phase (SP) of a lattice pattern (LP) of a setting image (SI) are associated with each other at each X-Y position; a step for generating a calibration point cloud table (CPT) defining the relationship between the respective Z positions and the calibration position for at least one of the X position and the Y position at each X-Y position; a step for acquiring a target phase (TP) of a lattice pattern (LP) of a target image (TI) at each X-Y position; a step for calculating the Z-position of a measurement object (20) at each X-Y position on the basis of the Z-position lookup table (ZLT) and the target phase (TP); and a step for identifying the three-dimensional shape of the measurement object (20) by calculating, on the basis of the Z-positions of the measurement object (20) and the calibration point cloud table (CPT), the post-calibration X-Y position for the respective Z positions of the measurement object (20) at each of the X-Y positions.

Description

Shape measurement method and shape measurement device

[0001] The present invention relates to a shape measurement method and a shape measurement device.

[0002] A known technology for non-contact shape measurement is described in, for example, Patent Document 1. In the technology disclosed in Patent Document 1, a reference plate including a reference surface is translated in a normal direction by minute increments, and a two-dimensional pattern is projected onto the reference plate at each translation position, and the projected two-dimensional pattern is imaged. This creates a table in which the amount of translation in the normal direction, each coordinate value in a three-dimensional coordinate system consisting of the coordinates of the two-dimensional pattern, and the phase of the two-dimensional pattern are associated one-to-one. When determining spatial coordinates corresponding to the phase between the reference surfaces, spatial coordinates interpolated between the reference surfaces are calculated using elements of the table having values ​​close to the phase between the reference surfaces.

[0003] Patent No. 4873485

[0004] According to the above-mentioned technique, since it is necessary to store a table in which the phase values ​​of the entire space to be measured and the coordinate values ​​of the three-dimensional coordinate axes are associated one-to-one, the table size becomes very large. Furthermore, since it takes time to search the elements stored in the table, there is a risk that the overall processing speed will decrease.

[0005] The present invention has been made in view of the above-mentioned problems, and aims to provide a shape measurement method and a shape measurement device that reduce the processing time for shape measurement.

[0006] One aspect of the present disclosure is a shape measurement method for measuring the shape of a measurement object using a grid projection method, comprising: a first step of translating a reference plate, which is parallel to an XY coordinate system and onto which a grid pattern is projected, in a Z-axis direction that is a normal direction of the reference plate by a predetermined amount, and capturing an image of the grid pattern projected onto the reference plate at each Z position to obtain a plurality of setting images; a second step of generating a Z-position lookup table that associates, at each XY position, a setting phase of the grid pattern in the setting image with a Z position; a third step of translating, at each XY position, the reference plate in the Z-axis direction by a predetermined amount, and capturing an image of a calibration pattern formed on the reference plate at each Z position to obtain a plurality of calibration images; a fourth step of generating calibration definition information that defines, at each XY position, a relationship between each Z position and at least one of an X position and a Y position as a calibration position; and a fifth step of projecting the grid pattern onto the measurement object and capturing an image of the grid pattern projected onto the measurement object to obtain a target image. This shape measurement method includes the steps of: a sixth step of acquiring a target phase of the grid pattern of the target image at each X-Y position; a seventh step of calculating a Z position of the measurement object at each X-Y position based on the Z position lookup table generated in the second step and the target phase acquired in the fifth step; and an eighth step of specifying a three-dimensional shape of the measurement object by calculating a calibrated X-Y position for each Z position of the measurement object at each X-Y position based on the Z position of the measurement object at each X-Y position calculated in the seventh step and the calibration definition information generated in the fourth step.

[0007] Another aspect of the present disclosure includes a reference plate onto which a grid pattern is projected; a movement device that translates the reference plate by a predetermined amount in a normal direction of the reference plate; a projection device that projects the grid pattern onto the reference plate and projects the grid pattern onto a measurement object at each position of the translation; an imaging device that acquires, at each position of the translation, a plurality of setting images obtained by capturing the grid pattern projected onto the reference plate, a plurality of calibration images obtained by capturing images of a calibration pattern formed on the reference plate, and a target image obtained by capturing the grid pattern projected onto the measurement object; and a processing device, wherein the processing device: generates, at each X-Y position, a Z-position lookup table that associates a setting phase of the grid pattern in the setting image with a Z position; generates, at each X-Y position, calibration definition information that defines a relationship between each Z position and at least one of the calibration positions of the X position and the Y position; acquires, at each X-Y position, a target phase of the grid pattern in the target image; The shape measurement device calculates the Z position of the measurement object at each X-Y position based on the Z position lookup table and the object phase, and calculates the calibrated X-Y position for each Z position of the measurement object at each X-Y position based on the Z position of the measurement object at each X-Y position and the calibration definition information, thereby specifying the three-dimensional shape of the measurement object.

[0008] According to one aspect and another aspect of the present disclosure, the Z position of a measurement object is calculated using a Z position lookup table, and the calibrated X-Y position for each Z position of the measurement object can be calculated using this Z position and calibration definition information. This allows the table size to be smaller than a table in which phases and coordinate values ​​on three-dimensional coordinate axes are associated one-to-one. Furthermore, the smaller table size allows for faster processing than when searching for elements stored in a table in which phases and coordinate values ​​on three-dimensional coordinate axes are associated one-to-one.

[0009] Furthermore, calibration definition information is generated using the calibration image, and the post-calibration XY positions for each Z position of the measurement object at each XY position are calculated based on the calibration definition information, thereby improving the accuracy of shape measurement of the measurement object.

[0010] It should be noted that the reference numerals in parentheses in the claims indicate the correspondence with the specific means described in the embodiments to be described later, and do not limit the technical scope of the present invention.

[0011] 1 is a block diagram showing a shape measurement device according to a first embodiment. FIG. 1 is a schematic diagram illustrating a process of generating a Z-position lookup table according to the first embodiment. FIG. 2 is an example of a Z-position lookup table according to the first embodiment. FIG. 2 is a schematic diagram illustrating a process of generating calibration definition information according to the first embodiment. FIG. 3 is a diagram illustrating a process of generating calibration definition information according to the first embodiment, where (a) is a diagram illustrating a calibration pattern and (b) is a diagram illustrating a calibration image. FIG. 4 is a diagram illustrating another example of a calibration pattern according to the first embodiment. FIG. 5 is a schematic diagram illustrating a process of acquiring a target image according to the first embodiment. FIG. 6 is a diagram illustrating a process of calculating a Z-position from a setting phase value based on the Z-position lookup table according to the first embodiment. FIG. 7 is a schematic diagram illustrating a state in which a pixel of a measurement object according to the first embodiment is located between the k plane and the k+1 plane. FIG. 8 is a schematic diagram illustrating a state in which a pixel of a measurement object according to the first embodiment is projected onto the k plane and the k+1 plane. FIG. 9 is a diagram illustrating a point obtained by shifting a point projected onto the k plane within the k plane and a point obtained by shifting a point projected onto the k+1 plane within the k+1 plane for a pixel of a measurement object according to the first embodiment. 1 is a diagram showing a process of calculating a calibrated XY position for each Z position of a measurement object according to embodiment 1. FIG. 2 is a main routine showing the operation of the shape measurement device according to embodiment 1. FIG. 3 is a flowchart showing a setting process according to embodiment 1. FIG. 4 is a flowchart showing a calibration process according to embodiment 1. FIG. 5 is a flowchart showing a shape measurement process according to embodiment 1. FIG. 6 is a block diagram showing a shape measurement device according to embodiment 2. FIG. 7 is a schematic diagram showing a state in which a pixel of a measurement object according to embodiment 2 is located between the k plane and the k+1 plane. FIG. 8 is a diagram showing a point shifted in the k plane and a point shifted in the k+1 plane for a pixel of a measurement object according to embodiment 2. FIG. 9 is a diagram showing a process of calculating a calibrated XY position for each Z position of a measurement object according to embodiment 2. FIG. 10 is a block diagram showing a shape measurement device according to embodiment 3. FIG. 11 is a schematic diagram for explaining a calibration lattice point table according to embodiment 3. FIG. 12 is a schematic diagram showing a state in which a pixel of a measurement object according to embodiment 3 is located inside a unit cube. FIG. 13 is a diagram showing a process of calculating a calibrated XY position of a measurement object according to embodiment 3.FIG. 11 is a schematic diagram for explaining another example of the calibration lattice point table according to the third embodiment.

[0012] (Embodiment 1) 1. Schematic configuration of shape measurement device 10a A shape measurement device 10a according to embodiment 1 will be described with reference to Fig. 1. The shape measurement device 10a according to this embodiment projects a grating pattern LP onto a measurement object 20, captures an image of the projected grating pattern LP, and measures the three-dimensional shape of the measurement object 20 based on an object phase TP of the obtained image. The shape measurement device 10a includes a reference plate 11, a moving device 12, a projection device 13, an image capture device 14, a processing device 15, a storage device 16, and a display device 17.

[0013] The reference plate 11 has a flat reference surface R. A grating pattern LP is projected onto the reference surface R by a projection device 13. The reference plate 11 is configured to be movable in the normal direction of the reference surface R. In this embodiment, the normal direction of the reference surface R is the Z-axis direction, and the reference surface R is the X-Y plane.

[0014] A measurement object 20 may be placed on the reference plate 11. When the measurement object 20 is placed on the reference plate 11, the reference surface R becomes the mounting plane PP. The measurement object 20 may be configured to be placed on a mounting table (not shown) that is located at a different position from the reference plate 11. In this case, the measurement object 20 is placed on a flat mounting plane PP provided on the mounting table.

[0015] The moving device 12 is controlled by the processing device 15 to translate the reference plate 11 by a predetermined amount in the normal direction of the reference plate 11. The moving device 12 is not particularly limited, and known methods such as a servo motor or a ball screw can be used. However, the moving device 12 may be movable along the planar direction of the reference plate 11, or may be rotatable about the X-axis, Y-axis, or Z-axis.

[0016] The projection device 13 is controlled by the processing device 15 to project a grid pattern LP or a calibration image CI onto the reference surface R of the reference plate 11. The projection device 13 is also controlled by the processing device 15 to project the grid pattern LP onto the measurement object 20. Data for generating the grid pattern LP may be stored in the storage device 16, or may be generated by the processing device 15 when capturing an image. Similarly, data for generating the calibration image CI may be stored in the storage device 16, or may be generated by the processing device 15 when capturing an image.

[0017] The imaging device 14 is controlled by the processing device 15 to capture an image of the grid pattern LP projected onto the reference surface R. The captured setting image SI is stored in the storage device 16. The imaging device 14 is also controlled by the processing device 15 to capture an image of the calibration pattern CP formed on the reference surface R. The captured calibration image CI is stored in the storage device 16. The imaging device 14 is also controlled by the processing device 15 to capture an image of the grid pattern LP projected onto the measurement object 20. The captured target image TI is stored in the storage device 16.

[0018] As the imaging device 14, a known imaging device 14 such as a CMOS (Complementary Metal-Oxide-Semiconductor) camera or a CCD (Charge-Coupled Device) camera can be used.

[0019] The processing device 15 controls the moving device 12, the projection device 13, and the imaging device 14. The processing device 15 also displays the three-dimensional shape of the measurement object 20 on the display device 17. Examples of the processing device 15 include a CPU (Central Processing Unit) and a DSP (Digital Signal Processor).

[0020] The storage device 16 stores the grid pattern LP, setting image SI, setting phase SP, Z position lookup table ZLT, calibration pattern CP, calibration image CI, calibration point cloud table CPT (an example of calibration definition information), target image TI, target phase TP, intermediate image II, and processed image PI. The storage device 16 may also be configured to store programs, initial values, and the like in advance. Examples of the storage device 16 include a RAM (Random Access Memory), a ROM (Read Only Memory), a hard disk drive, and a removable storage device.

[0021] The grid pattern LP is a predetermined two-dimensional pattern, and is configured as a grid with equal intervals. However, the intervals between the grids do not have to be equal. The grid is configured as a pattern of light and dark shades.

[0022] The setting image SI is acquired by the imaging device 14 capturing images of the grid patterns LP projected onto the reference surface R of the reference plate 11 at each position where the reference plate 11 moves parallel to the normal direction. The setting image SI is used to generate a Z-position lookup table ZLT.

[0023] The setting phase SP is the phase of the grid pattern LP captured in the setting image SI. The acquired setting phase SP is specified by each coordinate value of the three-dimensional coordinate system that forms the target space TS.

[0024] The Z position lookup table ZLT corresponds the setting phase SP of the grid pattern LP of the setting image SI to the Z position at each XY position in the three-dimensional coordinate system that constitutes the target space TS for measuring the three-dimensional shape of the measurement object 20.

[0025] The Z position lookup table ZLT is not particularly limited and may be, for example, a point cloud table that stores the setting phase SP of the grid pattern LP of the setting image SI and the Z position. The Z position lookup table ZLT may also be a line approximation table that stores parameters of a linear approximation equation of first or higher order that approximates the relationship between the setting phase SP of the grid pattern LP of the setting image SI and the Z position. The parameters of the linear approximation equation include the order of the variable, the coefficient of the variable, and a constant term. For example, if the linear approximation equation is a line, the parameters are the slope and intercept of the line. FIG. 3 shows an example of the Z position lookup table ZLT represented by a graph with the setting phase SP on the vertical axis and the Z position on the horizontal axis.

[0026] The above-mentioned three-dimensional coordinate system is composed of an X-axis, a Y-axis, and a Z-axis. In this embodiment, the XY coordinate system has XY coordinate axes corresponding to the reference plane R and the mounting plane PP. The Z-axis is perpendicular to the reference plane R and the mounting plane PP.

[0027] The calibration image CI is acquired by the imaging device 14 capturing images of the calibration patterns CP formed on the reference surface R of the reference plate 11 at each position as the reference plate 11 moves parallel to the normal direction. The calibration image CI is used to generate the calibration point cloud table CPT. Note that "forming" may involve drawing the calibration pattern CP on the reference plate 11, printing the calibration pattern CP on the reference plate 11, or arranging the calibration pattern CP on the reference plate 11 that has been formed in advance by printing on paper, a plastic plate, or the like, or forming the calibration pattern CP on the reference plate 11 by any method.

[0028] The calibration pattern CP is a predetermined two-dimensional pattern. There are no particular limitations on the calibration pattern CP, and any pattern, such as a checkered pattern (square grid) or a dotted pattern (circle grid), can be appropriately selected. FIG. 5A shows an example of a checkered pattern. The checkered pattern is formed by arranging white rectangles and black rectangles in a staggered pattern. The rectangles may be squares or rectangles. FIG. 6 shows an example of a dotted pattern. The dotted pattern is formed by arranging black circles in a staggered pattern on a white background.

[0029] The target image TI is acquired by the imaging device 14 capturing an image of the grid pattern LP projected onto the measurement object 20. The target image TI is used to identify the three-dimensional shape of the measurement object 20. Furthermore, the target phase TP is the phase of the grid pattern LP captured in the target image TI.

[0030] 2. Process for Generating Z Position Lookup Table ZLT The process for generating the Z position lookup table ZLT will be described with reference to Figures 2 to 5. Note that the process for generating the Z position lookup table ZLT is not limited to the following description.

[0031] In this embodiment, a Z-position lookup table ZLT is generated using a plurality of setting images SI. In this embodiment, X and Y coordinate axes are set corresponding to the reference plane R. Also, the direction perpendicular to the reference plane R is set as the Z coordinate axis. Figure 2 shows the X and Z coordinate axes.

[0032] 2, the reference surface R of the reference plate 11 is placed at a Z coordinate value Z0. The reference surface R placed at the Z coordinate value Z0 is defined as the reference surface R0.

[0033] A grid pattern LP is projected onto a reference plane R. As shown in FIG. 2 , the distribution of brightness values ​​of the grid pattern LP is set to a cosine wave. However, the distribution of brightness values ​​of the grid pattern LP may also be a sine wave. The setting phase SP of the setting image SI acquired by capturing the grid pattern LP monotonically increases between −1 and 1 as the Z coordinate value increases, and the sign is periodically reversed. However, in this embodiment, the phase is set to −1 when it is 0 radians, 0 when it is π radians, and 1 when it is 2π radians.

[0034] The imaging device 14 has an imaging element 14a. The imaging element 14a has pixels (not shown) regularly arranged on a plane. The imaging device 14 captures an image of a grid pattern LP projected onto a reference plane R0 to obtain a setting image SI on the reference plane R0. By matching each pixel of the imaging element 14a with the corresponding pixel, the X and Y coordinate values ​​of each pixel are obtained for the setting phase SP of the acquired setting image SI. Note that known methods such as a phase shift method and a Fourier transform fringe pattern analysis method can be used to obtain the X and Y coordinate values ​​of each pixel.

[0035] Although not shown in detail, in this embodiment, the X coordinate value of the target space TS varies from 0 to L, and the Y coordinate value varies from 0 to M.

[0036] The processing device 15 causes the moving device 12 to translate the reference plate 11 from Z coordinate value 0 to Z coordinate value 1. The reference surface R located at Z coordinate value 1 is set as reference surface R1.

[0037] A grid pattern LP is projected onto the reference surface R1 by the imaging device 14. Thereafter, the X and Y coordinate values ​​are obtained for the setting phase SP of the setting image SI on the reference surface R1 in the same manner as for the reference surface R0.

[0038] Similarly to the above, by further translating the reference plane R from Z coordinate value 2 to Z coordinate value N, the X and Y coordinate values ​​are obtained for the setting phase SP of the setting image SI from reference plane R2 to reference plane RN.

[0039] As the reference plate 11 moves from the position of the reference plane R0 to the position of the reference plane RN, the space defined by the trajectory of movement of the reference plate 11 becomes a target space TS in which the shape of the measurement object 20 can be measured. In this embodiment, the X coordinate value of the target space TS varies from 0 to L, the Y coordinate value varies from 0 to M, and the Z coordinate value varies from 0 to N.

[0040] 2, when one pixel of the imaging element 14a is focused on, the pixel sequentially captures points on a straight line 21. This pixel captures points P0, P1, P2, ..., PN according to the reference planes R0, R1, R2, ..., RN, respectively. The setting phases SP at the respective points P0, P1, P2, ..., PN are set to θ0, θ1, θ2, ..., θN. For the setting phase values ​​θ0, θ1, θ2, ..., θN at the reference planes R0, R1, R2, ..., RN, the X-coordinate value, the Y-coordinate value, and the Z-coordinate value are determined.

[0041] The setting phase SP of the lattice pattern LP of the setting image SI at the Z coordinate value of the three-dimensional coordinate system that constitutes the target space TS is stored as a Z position lookup table ZLT in the storage device 16. The Z position lookup table ZLT according to this embodiment is a point cloud table in which the setting phase SP of the lattice pattern LP of the setting image SI is associated with a Z position, which is a Z coordinate value, at each XY position. However, as described above, the Z position lookup table ZLT is not limited to a point cloud table.

[0042] 3. Process for Generating Calibration Point Cloud Table CPT The process for generating the calibration point cloud table CPT will be described with reference to Figures 2 and 4 to 5. However, the process for generating the calibration point cloud table CPT is not limited to the following description.

[0043] In this embodiment, one type of calibration pattern CP is used, and multiple calibration images CI are obtained by capturing images of the calibration pattern CP at each Z position, and a calibration point cloud table CPT is generated using the multiple calibration images CI. In this embodiment, an example will be described in which a checkered pattern is used as the calibration pattern CP.

[0044] As shown in FIG. 4 , the reference surface R of the reference plate 11 is placed at Z coordinate value Z0. The reference surface R placed at Z coordinate value Z0 is defined as reference surface R0. In this embodiment, a calibration pattern CP formed in advance on paper, a plastic plate, or the like by printing or the like is placed on reference surface R. As a result, the calibration pattern CP is formed on reference surface R of the reference plate 11. The calibration pattern CP formed on reference surface R0 is imaged by the imaging device 14, and a calibration image CI on reference surface R0 is obtained. The same operation is repeated for reference surfaces R0, R1, R2, ..., RN. As a result, multiple calibration images CI are obtained.

[0045] 5A, the calibration pattern CP is made up of white and black squares arranged in a staggered pattern. We will focus on four lattice points a, b, c, and d formed by the vertices of one of the squares.

[0046] FIG. 5B shows an example of a calibration image CI captured by the imaging device 14. The calibration image CI may be distorted compared to the calibration pattern CP due to the lens (not shown) of the imaging device 14. In FIG. 5B, the calibration image CI is composed of a plurality of distorted rectangles. However, depending on the performance of the lens of the imaging device 14, the imaging conditions, etc., the rectangles of the calibration image CI may not be distorted.

[0047] Let us now focus on a distorted rectangle in the calibration image CI in Fig. 5B that corresponds to a particular rectangle in the calibration pattern CP in Fig. 5A. Let the four lattice points of the distorted rectangle be A, B, C, and D, respectively.

[0048] Next, the positional deviations between lattice points a, b, c, and d of the calibration pattern CP and lattice points A, B, C, and D of the calibration image CI are associated. This positional deviation is a vector quantity starting from each lattice point a, b, c, and d of the calibration pattern CP and ending at the corresponding lattice point A, B, C, and D of the calibration image CI, and includes information about direction and magnitude. The above association is performed for all quadrangles constituting the calibration pattern CP and all quadrangles constituting the calibration image CI. This makes it possible to define the relationship between each Z position and at least one of the X and Y calibration positions at each X-Y position for one calibration pattern CP and one calibration image CI corresponding to this calibration pattern CP. Note that the above association is performed for the lattice points of the calibration image CI, and for points other than the lattice points, internal interpolation may be performed between multiple lattice points. This makes it possible to define the relationship between the positions of all points constituting the calibration pattern CP and the positions of all points constituting the calibration image CI.

[0049] The above operation is performed for each of the reference planes R0 to RN, from Z coordinate value Z0 to Z coordinate value ZN. In this way, calibration definition information is obtained for multiple calibration images CI, from 0 to N. The calibration definition information according to this embodiment is a calibration point cloud table CPT in which the positions of the points that make up the calibration pattern CP correspond to the positions of the points that make up the calibration image CI. The calibration point cloud table CPT according to this embodiment stores, for each X-Y position in each of the multiple calibration images CI, the correspondence between each Z position and the shift amount, which is the amount of change between each X-Y position of the calibration pattern CP and each X-Y position of the calibration image CI.

[0050] However, the calibration point cloud table CPT may also be configured to store correspondence between each Z position, each XY position of the calibration pattern CP, and each XY position of the calibration image CI at each XY position in each of multiple calibration images CI.

[0051] 7 and 8, a description will be given of a shape measurement process for measuring the shape of the measurement object 20. However, the shape measurement process is not limited to the following description.

[0052] The measurement object 20 is placed in the target space TS. FIG. 7 shows an example in which the measurement object 20 is placed on the reference surface R0 of the reference plate 11. The measurement object 20 can be placed at any position between the reference surface R0 and the reference surface RN. The measurement object 20 may be placed on any of the reference surfaces R0 to RN. The measurement object 20 may also be placed between adjacent reference surfaces R.

[0053] A grating pattern LP is projected onto the measurement object 20. The grating pattern LP projected onto the measurement object 20 is the same as the grating pattern LP projected onto the reference plane R when generating the Z-position lookup table ZLT. The imaging device 14 captures the grating pattern LP to obtain a target image TI. In this state, the imaging device 14 captures an image of a point Q on the measurement object 20. As a result, a target phase value φQ of the grating pattern LP projected onto point Q is obtained.

[0054] When point Q is located on any of reference planes R0 to RN, the Z coordinate value of point Q can be determined by the setting phase value θ stored in the Z position lookup table ZLT and the Z coordinate value.

[0055] 7, when point Q is located between adjacent reference surfaces R, the Z coordinate value of point Q cannot be identified from the Z position lookup table ZLT. In this way, when point Q is located between adjacent reference surfaces R, the Z coordinate value of point Q can be calculated as follows.

[0056] 8, a group of points having a predetermined Z coordinate value is extracted from the Z position lookup table ZLT and plotted with the setting phase value θ on the vertical axis and the Z coordinate value on the horizontal axis. Next, a search is made for each interval of adjacent setting phase values ​​θ to determine in which interval the target phase value φQ is located.

[0057] As a result of the search, as shown in FIG. 8, it is found that the phase is located between the setting phase PA and the setting phase PB.

[0058] Next, an imaginary line is drawn between adjacent setting phases PA and PB, and this line is divided internally by the target phase value φQ. The Z coordinate value ZQ of this internal division point is set as the Z coordinate value of the target phase value φQ. By performing the above steps, it is possible to calculate the Z position for each pixel that constitutes the target image TI. By calculating the Z position for each pixel that constitutes the target image TI in this manner, it is possible to generate an intermediate image II that has a Z position corresponding to each pixel.

[0059] However, the method for calculating the Z coordinate value of the target phase value φQ is not limited to the above. For example, in Fig. 8, the Z coordinate value of the target phase value φQ may be calculated by linearly approximating or curve-lining the points extracted from the Z position lookup table ZLT and substituting the target phase value φQ into the approximated curve or the formula of the approximated curve.

[0060] 4.2. Image Processing Process Each pixel constituting the intermediate image II generated as described above has a corresponding Z position. In other words, the intermediate image II is three-dimensional data in which each pixel has its own X-Y position and Z position. On the other hand, since the X-Y positions of the intermediate image II are pixel data, the intermediate image II is not point cloud data having the X, Y, and Z positions of the measurement object 20. For this reason, image processing can be easily performed on the intermediate image II. However, it is known that it is relatively difficult to perform image processing on point cloud data having X, Y, and Z positions.

[0061] The image processing performed on the intermediate image II is not particularly limited, and any processing can be appropriately selected. As the image processing, for example, one or more of a spatial filter, an edge enhancement process, a noise reduction process (such as a median filter), a smoothing process (such as a Gaussian filter), an MTF correction in the Z direction (an enhancement filter), etc. can be selected.

[0062] Image processing is performed on the intermediate image II to obtain a processed image PI, which is image data with a Z position corresponding to each pixel.

[0063] 4.3. XY Position Calculation Process Next, a method for calculating the XY position of the target image TI will be described with reference to Fig. 9 to Fig. 12. Fig. 9 shows an arbitrary pixel T0 included in the processed image PI. The X position of pixel T0 is x, the Y position is y, and the Z position is z. x and y are the X and Y positions of one pixel T0 in the processed image PI. Meanwhile, z is the Z position calculated based on the target image TI and the Z position lookup table ZLT.

[0064] A plane designated by the symbol k is an arbitrary k-th plane among the planes corresponding to the multiple calibration images CI stored in the calibration point cloud table CPT (where 0≦k≦N−1). Similarly, a plane designated by the symbol k+1 is an k+1-th plane among the planes corresponding to the multiple calibration images CI stored in the calibration point cloud table CPT. FIG. 9 shows an example in which pixel T0 is located between the k-th plane and the k+1-th plane. However, the position of pixel T0 is arbitrary, and it may be located on any calibration image CI.

[0065] 10, a point T1 is calculated by projecting pixel T0 onto the k-th plane in the Z-axis direction, and a point T2 is calculated by projecting pixel T0 onto the k+1-th plane in the Z-axis direction. The X position of point T1 is x, the Y position is y, and the Z position is z, which is the imaging position of the construction image corresponding to the k-th plane. ｋ The X position of the point T2 is x, the Y position is y, and the Z position is z, which is the imaging position of the calibration image CI corresponding to the (k+1)th plane. ｋ＋１ is.

[0066] The points included in each plane constituting the calibration point cloud table CPT have, at each XY position, a shift amount that is the relationship between each Z position and at least one of the calibration positions of the X position and the Y position. This will be explained in detail below.

[0067] Point T1 in Figure 11 is a point on plane k. Point T1 is at X position x and Y position y, and at Z position z ｋ and a shift amount x' corresponding to the X position and / or the y position of the calibration position.ｋ and the shift amount y' corresponding to the y position ｋ where the shift amount is x' ｋ Only y' is fine, or ｋ Only this is fine.

[0068] Point T1 of the processed image PI is shifted to point T3 after calibration. The X position of point T3 is calculated by multiplying the X position of point T1 by the shift amount x'. ｋ and x + x' ｋ The Y position of point T3 is the Y position of point T1 and the shift amount y'. ｋ and y+y' ｋ is.

[0069] 11 is a point on the plane k+1. The point T2 is located at the Z position z at the X position x and the Y position y. ｋ＋１ and a shift amount x' corresponding to the X position and / or the y position of the calibration point. ｋ＋１ and the shift amount y' corresponding to the y position ｋ＋１ where the shift amount is x' ｋ＋１ Only y' is fine, or ｋ＋１ Only this is fine.

[0070] Point T2 in the processed image PI is shifted to point T4 after calibration. The X position of point T4 is calculated by multiplying the X position of point T2 by the shift amount x'. ｋ＋１ and x + x' ｋ＋１ The Y position of point T4 is the Y position of point T1 and the shift amount y'. ｋ＋１ and y+y' ｋ＋１ is.

[0071] Next, as shown in Fig. 12, an internal division point T5 is calculated between points T3 and T4. The X position of internal division point T5 is x + Δx, the Y position is y + Δy, and the Z position is z. Here, Δx is the correction amount for the X position, and Δy is the correction amount for the Y position. In other words, internal division point T5(x + Δx, y + Δy, z) is obtained by correcting pixel T0(x, y, z) by the correction amount Δx for the X position and the correction amount Δy for the Y position.

[0072] The correction amount Δx is the shift amount x, ｋ , x' ｋ＋１ , and Z position z ｋ , z ｋ＋１ It is calculated from the following formula (1).

[0073]

[0074] The correction amount Δy is the shift amount y′ ｋ , y′ ｋ＋１ , and Z position z ｋ , z ｋ＋１ It is calculated from the following formula (2).

[0075]

[0076] As described above, the operation of calculating the internal division point T5 (x + Δx, y + Δy, z) from pixel T0 (x, y, z) is performed for all pixels in the processed image PI, thereby identifying the three-dimensional shape of the measurement object 20.

[0077] 5. Operation of the Shape Measuring Apparatus 10a According to the Present Embodiment Next, the operation of the shape measuring apparatus 10a according to the present embodiment will be described with reference to Figures 13 to 15. 5.1. Main Routine Figure 13 shows the main routine of the operation of the shape measuring apparatus 10a according to the present embodiment. When the shape measuring apparatus 10a is started, a setting process S1 is executed. The setting process S1 generates a Z-position lookup table ZLT.

[0078] Next, a calibration process S2 is executed to generate a calibration point cloud table CPT.

[0079] Subsequently, a shape measurement process S3 is executed. The shape measurement process S3 calculates the Z position of the measurement object 20 based on the Z position lookup table ZLT, and identifies the three-dimensional shape of the measurement object 20 based on the Z position and the calibration point cloud table CPT.

[0080] Subsequently, a display process S4 is executed, in which the three-dimensional shape of the measurement object 20 is displayed on the display device 17.

[0081] This completes the operation of the shape measuring device 10a.

[0082] 5.2. Setting Process S1 Fig. 14 shows a flowchart of the setting process S1. As shown in Fig. 14, when the setting process S1 is executed, the projection device 13 projects a grid pattern LP onto the reference surface R0 of the reference plate 11 (S11). The imaging device 14 captures an image of the grid pattern LP projected onto the reference surface R0. As a result, the imaging device 14 acquires a setting image SI (S12).

[0083] The setting phase SP is acquired from the setting image SI, and the setting phase SP and the X, Y and Z coordinate values ​​of the setting phase SP are stored in the storage device 16 (S13).

[0084] Next, the moving device 12 translates the reference plate 11 by a predetermined amount in the normal direction of the reference surface R (S14). As a result, the reference surface R0 moves to the reference surface R1. S11, S12, S13, and S14 correspond to the first step.

[0085] It is determined whether the reference plane R has reached the reference plane RN and the projection and imaging of the grid pattern LP has been completed in all sections of the target space TS (S15).

[0086] If projection and imaging of the grid pattern LP has not been completed for all sections (S15: N), S11 to S14 are repeated.

[0087] When the projection and imaging of the grid pattern LP for all sections is completed (S25: Y), a Z-axis direction lookup table is completed that associates the setting phase of the grid pattern of the setting image SI with the Z position at each XY position (corresponding to the second step).

[0088] 5.3. Calibration Process Fig. 15 shows a flowchart of the calibration process S2. As shown in Fig. 15, when the calibration process is executed, the projection device 13 forms the calibration pattern CP on the reference surface R0 of the reference plate 11 by placing a pre-formed calibration pattern CP on the reference surface R0 of the reference plate 11 (S21). The imaging device 14 captures an image of the calibration pattern CP formed on the reference surface R0. As a result, the imaging device 14 acquires a calibration image CI (S22).

[0089] The positions of the grid points of the acquired calibration image CI are stored in the storage device 16 (S23). S21 and S22 correspond to the third step.

[0090] The positions of the grid points of the calibration pattern CP are compared with the positions of the grid points of the calibration image CI, and the amount of movement of the grid points of the calibration image CI from the grid points of the calibration pattern CP is calculated (S24). The amount of movement is a vector quantity starting from the grid point of the calibration pattern CP and ending at the grid point of the calibration image CI, and includes information about the direction and magnitude.

[0091] The calculated movement amount of the grid points of the calibration image CI is stored in the storage device 16 (S25).

[0092] Next, the moving device 12 translates the reference plate 11 by a predetermined amount in the normal direction of the reference surface R (S26). As a result, the reference surface R0 moves to the reference surface R1.

[0093] It is determined whether the reference plane R has reached the reference plane RN and imaging of the calibration pattern CP has been completed in the entire section of the target space TS (S27).

[0094] If imaging of the calibration pattern CP has not been completed for all sections (S27: N), S21 to S26 are repeated.

[0095] When imaging of the calibration pattern CP in all sections is completed (S27: Y), calibration definition information that defines the relationship between each Z position and at least one of the calibration positions of the X position and the Y position is completed for each XY position (corresponding to the fourth step). This completes the calibration process.

[0096] 5.4. Shape Measurement Processing Next, the shape measurement processing S3 will be described with reference to FIG. 16. FIG. 16 shows a flowchart of the shape measurement processing S3. When the shape measurement processing S3 is executed, a grid pattern LP is projected onto the measurement object 20 placed on the reference plate 11 (S31). The imaging device 14 captures the grid pattern LP projected onto the measurement object 20 to obtain a target image TI (S32, corresponding to the fifth step), and obtains a target phase TP of the grid pattern LP (S33, corresponding to the sixth step).

[0097] Next, the Z position lookup table ZLT is obtained from the storage device (S34).

[0098] Based on the Z position lookup table ZLT and the target phase TP, the Z position of the measurement target 20 at each XY position is calculated (S35). As a result, an intermediate image II is generated in which the pixels of the target image TI have Z positions. S35 corresponds to the seventh step.

[0099] Next, predetermined image processing is performed on the intermediate image II (S36), thereby generating a processed image PI.

[0100] Next, the calibration point cloud table CPT is obtained from the storage device (S37).

[0101] Next, based on the Z position of each pixel in the processed image PI and the calibration point cloud table CPT, the shift amount, which is the amount of change between each Z position and each XY position of the calibration pattern CP and each XY position of the calibration image CI, is calculated for each XY position in each of the multiple calibration images CI (S38).

[0102] Next, based on the Z position of the measurement object at each XY position and the calibration definition information, the XY correction amount corresponding to each Z position of the measurement object at each XY position is calculated (S39).

[0103] Next, the three-dimensional shape of the measurement object is identified by calculating the calibrated XY position based on the Z position of the measurement object at each XY position and the XY correction amount (S39). As a result, each coordinate value in the three-dimensional coordinate system of the measurement object 20 can be identified, and the three-dimensional shape of the measurement object 20 can be identified (corresponding to step 8). This completes the shape measurement process.

[0104] 6. Effects of the Present Embodiment Next, the effects of the present embodiment will be described. The shape measurement method according to the present embodiment includes a first step, a second step, a third step, a fourth step, a fifth step, a sixth step, a seventh step, and an eighth step.

[0105] In the first step, a reference plate 11, which is parallel to an X-Y coordinate system and onto which a grid pattern LP is projected, is translated by a predetermined amount in the Z-axis direction, which is the normal direction of the reference plate 11, and the grid pattern LP projected onto the reference plate 11 at each Z position is captured to obtain multiple setting images SI. In the second step, a Z-position lookup table ZLT is generated that associates the setting phase SP of the grid pattern LP of the setting image SI with the Z position at each X-Y position. In the third step, the reference plate 11 is translated by a predetermined amount in the Z-axis direction, and the calibration pattern CP formed on the reference plate 11 at each Z position is captured to obtain multiple calibration images CI. In the fourth step, a calibration point cloud table CPT is generated that defines the relationship between each Z position and at least one of the X and Y calibration positions at each X-Y position. In a fifth step, a grid pattern LP is projected onto the measurement object 20, and the grid pattern LP projected onto the measurement object 20 is captured to obtain a target image TI. In a sixth step, a target phase TP of the grid pattern LP of the target image TI is obtained at each XY position. In a seventh step, a Z position of the measurement object at each XY position is calculated based on the Z position lookup table ZLT generated in the second step and the target phase TP obtained in the fifth step. In an eighth step, a calibrated XY position for each Z position of the measurement object 20 at each XY position is calculated based on the Z position of the measurement object 20 at each XY position calculated in the seventh step and the calibration point cloud table CPT generated in the fourth step, thereby identifying the three-dimensional shape of the measurement object 20.

[0106] The shape measuring device 10 a according to this embodiment also includes a reference plate 11 , a moving device 12 , a projecting device 13 , an imaging device 14 , and a processing device 15 .

[0107] A grid pattern LP is projected onto the reference plate 11. The movement device 12 translates the reference plate 11 by a predetermined amount in the normal direction of the reference plate 11. The projection device 13 projects the grid pattern LP onto the reference plate 11 at each position of the translation, and also projects the grid pattern LP onto the measurement object 20. The imaging device 14 acquires, at each position of the translation, a plurality of setting images SI obtained by capturing the grid pattern LP projected onto the reference plate 11, a plurality of calibration images CI obtained by capturing the calibration pattern CP formed on the reference plate 11, and a target image TI obtained by capturing the grid pattern LP projected onto the measurement object 20.

[0108] The processing device 15 generates a Z-position lookup table ZLT that associates the setting phase SP of the lattice pattern LP of the setting image SI with the Z position at each X-Y position, generates a calibration point cloud table CPT that defines the relationship between each Z position and at least one of the calibration positions of the X position and the Y position at each X-Y position, acquires the target phase TP of the lattice pattern LP of the target image TI at each X-Y position, calculates the Z position of the measurement object 20 at each X-Y position based on the Z-position lookup table ZLT and the target phase TP, and specifies the three-dimensional shape of the measurement object 20 by calculating the calibrated X-Y position for each Z position of the measurement object 20 at each X-Y position based on the Z position of the measurement object 20 at each X-Y position and the calibration point cloud table CPT.

[0109] According to this embodiment, the Z position of the measurement object 20 is calculated using the Z position lookup table ZLT, and then this Z position and the calibration point cloud table CPT are used to calculate the calibrated X-Y position for each Z position of the measurement object 20. This makes it possible to reduce the table size compared to a table in which phases and coordinate values ​​on three-dimensional coordinate axes are associated one-to-one. Furthermore, because the table size is reduced, processing speed can be improved compared to searching for elements stored in a table in which phases and coordinate values ​​on three-dimensional coordinate axes are associated one-to-one.

[0110] Furthermore, in an eighth step according to this embodiment, an XY correction amount corresponding to each Z position of the measurement object 20 at each XY position is calculated based on the Z position of the measurement object 20 at each XY position and the calibration point cloud table CPT, and a post-calibration XY position is calculated based on the Z position of the measurement object 20 at each XY position and the XY correction amount, thereby specifying the three-dimensional shape of the measurement object 20. In this way, since the three-dimensional shape of the measurement object 20 can be specified by calculating the XY correction amount, processing speed can be improved compared to searching for elements stored in a table in which phases correspond one-to-one to each coordinate value on the three-dimensional coordinate axes.

[0111] In addition, in an eighth step according to this embodiment, a Z position is calculated for each pixel constituting the target image TI acquired in the fifth step based on the Z position lookup table ZLT generated in the second step and the target phase TP acquired in the fifth step, thereby generating an intermediate image II having a Z position corresponding to each pixel, image processing is performed on the intermediate image II to generate a processed image PI, and calibrated X-Y positions for each Z position of the measurement target 20 at each X-Y position are calculated based on the Z positions of the processed image PI and the calibration point cloud table CPT generated in the fourth step, thereby identifying the three-dimensional shape of the measurement target 20. This makes it possible to easily perform image processing before generating point cloud data relating to the three-dimensional shape of the measurement target 20.

[0112] The calibration definition information according to this embodiment is a calibration point cloud table CPT that associates each Z position with the amount of change between each XY position of the calibration pattern CP and each XY position of the calibration image CI at each XY position in each of multiple calibration images CI, thereby reducing the amount of calculation.

[0113] Furthermore, according to this embodiment, a calibration point cloud table CPT is generated using the calibration image CI, and based on this calibration point cloud table CPT, a post-calibration XY position is calculated for each Z position of the measurement object 20. This allows the XY position to be uniquely determined without performing phase unwrapping, thereby improving measurement accuracy.

[0114] Second Embodiment Next, a shape measurement apparatus 10b according to a second embodiment will be described with reference to Fig. 17. The storage device 16 according to this embodiment differs from the first embodiment in that it stores a calibration function CF instead of the calibration point cloud table CPT. The calibration function CF represents, at each XY position in each of a plurality of calibration images CI, each Z position and a shift amount, which is the amount of change between each XY position of the calibration pattern CP and each XY position of the calibration image.

[0115] Next, a method for calculating the X-Y position of the target image TI according to this embodiment will be described with reference to Fig. 18 to Fig. 20. Fig. 18 shows an arbitrary pixel T0 included in the processed image PI. The X position of pixel T0 is x, the Y position is y, and the Z position is z. x and y are the X and Y positions of one pixel T0 in the processed image PI. Meanwhile, z is the Z position calculated based on the target image TI and the Z position lookup table ZLT.

[0116] The planes designated by the reference symbol k and the planes designated by the reference symbol k+1 are the same as those in the first embodiment, and therefore, redundant explanations will be omitted.

[0117] 19, the calibration function CF is associated with each plane and stored in the storage device 16. For the plane marked with the symbol k, the following equation (3) is stored as the calibration function CF.

[0118]

[0119] The x and y on the right side of equation (3) are the X and Y positions of point T0. ｋ is the calibration function CF stored for the plane marked with the symbol k. ｋ and y' ｋ is the shift amount corresponding to the X position (x) of point T0 and the shift amount corresponding to the Y position (y) of point T0.

[0120] 19 is a point on plane k. The X position of point T6 is the X position of point T0 and the shift amount x' ｋ and x + x' ｋThe Y position of point T6 is the Y position of point T0 and the shift amount y'. ｋ and y+y' ｋ However, the shift amount is x' ｋ Only y' is fine, or ｋ Only this is fine.

[0121] For the plane designated by the symbol k+1, the following formula (4) is stored as the calibration function CF.

[0122]

[0123] The x and y on the right side of equation (4) are the X and Y positions of point T0. ｋ＋１ is the calibration function CF stored for the plane labeled k+1. ｋ＋１ and y' ｋ +1 is the shift amount corresponding to the X position (x) of point T0 and the shift amount corresponding to the Y position (y) of point T0.

[0124] 19, point T7 is a point on plane k+1. The X position of point T7 is the X position of point T0 and the shift amount x' ｋ＋１ and x + x' ｋ＋１ The Y position of point T7 is the Y position of point T0 and the shift amount y'. ｋ＋１ and y+y' ｋ＋１ However, the shift amount is x' ｋ＋１ Only y' is fine, or ｋ＋１ Only this is fine.

[0125] Next, as shown in FIG. 20 , an internal division point T8 is calculated between points T6 and T7. The X position of internal division point T8 is x + Δx, the Y position is y + Δy, and the Z position is z. Here, Δx is the correction amount for the X position, and Δy is the correction amount for the Y position. In other words, internal division point T8 (x + Δx, y + Δy, z) is obtained by correcting pixel T0 (x, y, z) by the correction amount Δx for the X position and the correction amount Δy for the Y position. The formulas for calculating the correction amounts Δx and Δy are the same as those in the first embodiment, so a duplicated explanation will be omitted.

[0126] As described above, the operation of calculating the internal division point T8 (x + Δx, y + Δy, z) from pixel T0 (x, y, z) is performed for all pixels in the processed image PI, thereby identifying the three-dimensional shape of the measurement object 20.

[0127] Note that, among the symbols used in the second and subsequent embodiments, the same symbols as those used in the previous embodiments represent the same components, etc. as those in the previous embodiments, unless otherwise specified.

[0128] According to this embodiment, the calibration function CF can be used to calculate, from each X-Y position in the calibration image CI, a shift amount, which is the amount of change between each X-Y position of the calibration pattern CP and each X-Y position of the calibration image CI. This method simplifies calculations because the X-Y positions are uniquely determined without performing phase unwrapping, thereby improving calculation speed and measurement accuracy. Furthermore, if the relationship between the Z position and the shift amount x' and / or the shift amount y' is continuous and highly linear, the calibration function CF can be written using an even simpler approximation formula and a smaller table capacity. This can improve the overall efficiency of shape measurement of the measurement object 20.

[0129] Furthermore, since it is only necessary to store the function formula, the capacity required to store the calibration definition information can be reduced.

[0130] Third Embodiment Next, a shape measurement apparatus 10c according to a third embodiment will be described with reference to Fig. 21. The storage device 16 according to this embodiment differs from the first embodiment in that it stores a calibration lattice point table CLT instead of the calibration point cloud table CPT. The calibration lattice point table CLT associates the X, Y, and Z positions of a grid obtained by dividing the calibration space CS, which is formed by the trajectory of the translation of the reference plate 11 in the third step, with the amount of change between each X-Y position of the calibration pattern CP and each X-Y position of the calibration image CI.

[0131] Next, a method for calculating the XY position of the target image TI according to this embodiment will be described with reference to FIGS.

[0132] 22 , in this embodiment, the calibration space CS is divided into cubes consisting of eight lattice points P0 to P7 along three axes, the X-axis, the Y-axis, and the Z-axis, and each of the lattice points P0 to P7 has, in addition to an X-position, a Y-position, and a Z-position, shift amounts x' and y' that are the amounts of change between each X-Y position of the calibration pattern CP and each X-Y position of the calibration image CI. The storage device 16 stores, as a calibration lattice point table CLT, a table that associates coordinate values ​​with shift amounts for the lattice points of all the cubes that make up the calibration space CS.

[0133] 23 shows a case where pixel T0 is located inside a cube having eight lattice points P0 to P7. The coordinate values ​​of pixel T0 are (x, y, z).

[0134] The processor 15 calculates the coordinate value of point T9 by dividing pixel T0 located inside a cube having eight lattice points P0 to P7 internally based on the volume ratio of the cube and interpolating. The coordinate value of point T9 is (x + Δx, y + Δy, z). The correction amounts Δx and Δy are expressed based on the shift amounts x' and y' stored at the eight lattice points P0 to P7.

[0135] According to this embodiment, the calculation is performed using data from many grid points, so the interpolation accuracy is high. Also, since each grid point has data, if the shift amount needs to be corrected, the data related to the shift amount can be easily corrected.

[0136] 25 illustrates a modification of the third embodiment. In this embodiment, a cube consisting of eight lattice points P0 to P7 is further divided into a plurality of tetrahedrons, and it is determined to which of the divided tetrahedrons pixel T0 inside the cube belongs. After that, the shift amounts on the four lattice points surrounding the input point are weighted to calculate the correction amounts Δx and Δy.

[0137] According to this embodiment, a tetrahedron is the smallest solid that divides three-dimensional space, and the number of lattice points used at one time is the smallest, at four, so it is possible to reduce memory access and improve calculation speed. However, the method for dividing the unit cube is not particularly limited, and for example, it may be divided into six tetrahedrons of equal volume along the diagonals.

[0138] However, it is also possible to divide the unit cube diagonally to form two triangular prisms, determine which of the divided triangular prisms the pixel T0 located inside the cube belongs to, and then calculate the correction amount by weighting six shift amounts on six lattice points of the triangular prism surrounding the input point.

[0139] The present invention is not limited to the above-described embodiments, and can be applied to various embodiments within the scope of the present invention.

[0140] 10a, 10b, 10c: shape measurement device, 11: reference plate, 12: moving device, 13: projection device, 14: imaging device, 14a: imaging element, 15: processing device, 16: storage device, 17: display device, 20: measurement object, CF: calibration function, CI: calibration image, CLT: calibration lattice point table, CP: calibration pattern, CPT: calibration point cloud table, CS: calibration space, II: intermediate image, LP: lattice pattern, PI: processed image, PP: placement plane, R: reference surface, S1: setting process, S2: calibration process, S3: shape measurement process, S4: display process, SI: setting image, SP: setting phase, TI: target image, TP: target phase, TS: target space, ZLT: Z-position lookup table

Claims

1. A shape measurement method for measuring the shape of a measurement object (20) using a grating projection method, comprising: a first step of translating a reference plate (11) parallel to an XY coordinate system, onto which a grating pattern (LP) is projected, in a Z-axis direction that is a normal direction of the reference plate, by a predetermined amount, and capturing an image of the grating pattern projected onto the reference plate at each Z position to obtain a plurality of setting images (SI); a second step of generating a Z-position lookup table (ZLT) that associates, at each XY position, a setting phase (SP) of the grating pattern of the setting image with a Z position; and a third step of translating the reference plate in the Z-axis direction by a predetermined amount, and capturing an image of a calibration pattern (CP) formed on the reference plate at each Z position to obtain a plurality of calibration images (CI). a fourth step of generating calibration definition information (CPT, CF, CLT) that defines, for each X-Y position, a relationship between each Z position and at least one of the calibration positions of the X and Y positions; a fifth step of projecting the grid pattern onto the measurement object and capturing an image of the grid pattern projected onto the measurement object to acquire an object image (TI); a sixth step of acquiring, for each X-Y position, an object phase (TP) of the grid pattern of the object image; a seventh step of calculating a Z position of the measurement object at each X-Y position based on the Z position lookup table generated in the second step and the object phase acquired in the fifth step; and an eighth step of specifying a three-dimensional shape of the measurement object by calculating a calibrated X-Y position for each Z position of the measurement object at each X-Y position based on the Z positions of the measurement object at each X-Y position calculated in the seventh step and the calibration definition information generated in the fourth step.

2. The shape measurement method according to claim 1, wherein the eighth step calculates an XY correction amount corresponding to each Z position of the measurement object at each XY position based on the Z position of the measurement object at each XY position and the calibration definition information, and specifies the three-dimensional shape of the measurement object by calculating the calibrated XY position based on the Z position of the measurement object at each XY position and the XY correction amount.

3. The shape measurement method according to claim 1, wherein the eighth step calculates a Z position for each pixel constituting the object image acquired in the fifth step based on the Z position lookup table generated in the second step and the object phase acquired in the fifth step, thereby generating an intermediate image (II) having a Z position corresponding to each pixel; performing image processing on the intermediate image to generate a processed image (PI); and calculating calibrated X-Y positions for each Z position of the object to be measured at each X-Y position based on the Z positions of the processed image and the calibration definition information generated in the fourth step, thereby identifying the three-dimensional shape of the object to be measured.

4. The shape measurement method according to claim 3, wherein the image processing includes at least one of noise removal processing, smoothing processing, and MTF correction processing in the Z-axis direction.

5. A shape measurement method according to claim 1, wherein the calibration definition information is a calibration point cloud table (CPT) that associates, at each XY position in each of the plurality of calibration images, each Z position with a shift amount (x', y') that is the amount of change between each XY position of the calibration pattern and each XY position of the calibration image.

6. A shape measurement method according to claim 1, wherein the calibration definition information is a calibration function (CF) that represents, at each XY position in each of the plurality of calibration images, each Z position and a shift amount that is the amount of change between each XY position of the calibration pattern and each XY position of the calibration image.

7. A shape measurement method according to claim 1, wherein the calibration definition information is a calibration lattice point table (CLT) that associates the X, Y and Z positions of a grid obtained by dividing a calibration space (CS) formed by the trajectory of the parallel translation of the reference plate in the third step into a grid, with shift amounts that are amounts of change between each X-Y position of the calibration pattern and each X-Y position of the calibration image.

8. A system comprising: a reference plate onto which a grid pattern is projected; a moving device (12) for translating the reference plate by a predetermined amount in the normal direction of the reference plate; a projection device (13) for projecting the grid pattern onto the reference plate and projecting the grid pattern onto a measurement object at each position of the translation; an imaging device (14) for acquiring, at each position of the translation, a plurality of setting images obtained by imaging the grid pattern formed on the reference plate, a plurality of calibration images obtained by imaging the calibration pattern projected onto the reference plate, and an object image obtained by imaging the grid pattern projected onto the measurement object; and a processing device (15), wherein the processing device generates, at each X-Y position, a Z-position lookup table that associates the setting phase of the grid pattern in the setting image with a Z position, and generates, at each X-Y position, calibration definition information that defines the relationship between each Z position and at least one of the calibration positions of the X position and the Y position, A shape measurement device (10a, 10b, 10c) that acquires a target phase of the grid pattern of the target image at each X-Y position, calculates a Z position of the measurement object at each X-Y position based on the Z position lookup table and the target phase, and calculates a calibrated X-Y position for each Z position of the measurement object at each X-Y position based on the Z position of the measurement object at each X-Y position and the calibration definition information, thereby specifying a three-dimensional shape of the measurement object.

Images