JP2007322162A - Three-dimensional shape measuring apparatus and three-dimensional shape measuring method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a three-dimensional shape measuring apparatus and a three-dimensional shape measuring method which can measure the shape of a specular surface with high accuracy. <P>SOLUTION: The three-dimensional shape measuring apparatus 100 for measuring the shape of a specular surface S of an object under measurement T comprises a display means 10 for displaying a pattern, having wave-like density distribution, a shift means for shifting the pattern displayed on the display means 10, photographic means 20A, 20B for photographing the surface S reflecting the pattern displayed on the display means 10 from two different directions, and a surface shape analysis means for determining the shape of the range of the surface S through correspondence between the image data obtained by photographing the range of the surface S by the photographic means 20A, 20B that have been duplicated. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a three-dimensional shape measuring apparatus and a three-dimensional shape measuring method for measuring a specularly reflecting surface shape.

  In recent years, the surface shape of an object to be measured is measured in a non-contact manner using a three-dimensional surface measuring instrument (also called a non-contact three-dimensional measuring instrument, scanner, digitizer, etc.), and the presence or absence of scratches or distortion on the surface is automatically detected. Inspections are being conducted. There is a great demand for measuring the surface shape of a measurement object having a specularly reflecting surface such as a car body, silicon wafer, or glass, which is an industrial product, but ordinary three-dimensional surface measuring instruments measure the shape of such a surface. I can't. As a method for measuring the shape of the mirror-reflecting surface, there is a method called a phase shift method.

For example, the phase shift method disclosed in Patent Document 1 uses two types of lattice pattern plates on which sinusoidal shading patterns are formed. The phase is shifted by rotating the pattern plate to project the pattern. The surface of the measured object is photographed by one CCD camera, and the surface shape is measured based on images photographed at a plurality of phases.
JP 11-257930 A

  However, in the above conventional phase shift method, the point on the surface of the measurement object corresponding to the pixel of the CCD camera is the incident light vector at the point on the surface of the measurement object corresponding to the pixel adjacent to the pixel. The surface shape is obtained on the assumption that it is approximately present on a plane having a normal vector that is a bisector vector of the reflected light vector. For this reason, there is a problem that the measurement accuracy of the surface shape is low. In particular, when there is a scratch on the surface of the measurement object, since the points on the surface of the measurement object corresponding to the adjacent pixels of the CCD camera are not close to each other, the measurement accuracy of the surface shape is further lowered. there were.

  The present invention has been made to solve such a problem, and an object of the present invention is to provide a three-dimensional shape measuring apparatus and a three-dimensional shape measuring method capable of measuring the shape of a mirror-reflecting surface with high accuracy. .

  In order to achieve the above object, a three-dimensional shape measuring apparatus according to claim 1 is a three-dimensional shape measuring apparatus for measuring a specularly reflected surface shape of an object to be measured, wherein a pattern having a wavy concentration distribution is obtained. Display means for displaying, shift means for shifting the pattern displayed on the display means, imaging means for photographing the surface of the measurement object on which the pattern displayed on the display means is projected from at least two different directions, The imaging means comprises surface shape analyzing means for obtaining a surface shape of the range by associating image data obtained by imaging the overlapping range of the surface.

  The three-dimensional shape measuring apparatus according to claim 2 is the three-dimensional shape measuring apparatus according to claim 1, wherein the surface shape analyzing means at a point on the surface where the pattern displayed on the display means is projected. The points are overlapped and the image data obtained by the imaging unit are associated with each other so that the normal directions of the surfaces are equal.

  A three-dimensional shape measuring apparatus according to a third aspect is the three-dimensional shape measuring apparatus according to the first or second aspect, wherein the photographing unit is a plurality of CCD cameras for photographing in synchronization with the shift of the pattern. It is characterized by that.

  The three-dimensional shape measuring apparatus according to claim 4 is the three-dimensional shape measurement according to any one of claims 1 to 3, wherein the display means displays a pattern having a sine wave or triangular wave density distribution. It is characterized by doing.

The three-dimensional shape measurement method according to claim 5 is a three-dimensional shape measurement method for measuring a specularly reflected surface shape of a measurement object, and shifts and displays a pattern having a wavy concentration distribution,
The surface of the measurement object on which the pattern is projected is photographed from at least two different directions, and the surface shape of the range is obtained by associating image data obtained by photographing the overlapping range of the surface. .

  The three-dimensional shape measurement method according to claim 6 is a three-dimensional shape measurement method for measuring a surface shape of a measurement object, wherein the measurement is performed by shifting a pattern having a triangular wave density distribution and projecting the pattern. The surface of the object is photographed, and the shape of the surface is obtained based on image data obtained by photographing the surface.

  According to the three-dimensional shape measuring apparatus according to claim 1, the overlapping range of the mirror-reflecting surface of the measurement object is correlated with the image data obtained by photographing the photographing unit from at least two different directions. The surface shape analysis means obtains the surface shape. Therefore, image data obtained by photographing the same point on the surface from at least two different directions can be associated with each other, so that the shape of the specularly reflected surface can be measured with high accuracy.

  According to the three-dimensional shape measuring apparatus according to claim 2, the surface shape analyzing means captures the point in an overlapping manner so that the normal direction of the surface at the point on the surface where the pattern is projected is equal. The image data obtained by the photographing means is associated with each other. Therefore, since it is obtained directly without assuming an approximation as in the prior art, the shape of the mirror-reflecting surface can be measured with high accuracy.

  According to the three-dimensional shape measuring apparatus of the third aspect, the photographing means is a plurality of CCD cameras that photograph in synchronization with the shift of the pattern. Therefore, the surface on which the pattern shifted for each pixel of each CCD camera is projected can be sequentially photographed. Therefore, it is possible to measure the shape of the mirror-reflecting surface with higher accuracy and higher speed.

  According to the three-dimensional shape measuring apparatus of the fourth aspect, the display means displays a pattern having a sinusoidal or triangular wave density distribution. Therefore, the process in the surface shape analysis means can be simplified and the association can be reliably performed. Therefore, it is possible to measure the shape of the mirror-reflecting surface with higher accuracy and higher speed.

  According to the three-dimensional shape measuring method according to claim 5, the surface shape of the range is obtained by associating image data obtained by photographing the overlapping range of the mirror-reflecting surface of the measurement object from at least two different directions. Ask. Therefore, image data obtained by photographing the same point on the surface from at least two different directions can be associated with each other, so that the shape of the mirror-reflecting surface can be measured with high accuracy.

  According to the three-dimensional shape measuring method of the sixth aspect, the surface shape is obtained based on the image data obtained by photographing the surface of the measuring object on which the pattern having the shifted triangular wave-like density distribution is projected. Therefore, unlike the prior art, numerical calculation using a trigonometric function with a large amount of calculation is not required, so that the surface shape can be measured at high speed.

  A three-dimensional shape measuring apparatus and a three-dimensional shape measuring method according to an embodiment of the present invention will be described with reference to the drawings. A three-dimensional shape measuring apparatus 100 according to the present embodiment measures a shape of a mirror-reflecting surface S of a measurement object T as shown in a conceptual explanatory diagram in FIG. 1 and displays a pattern ( Display means) 10, two CCD cameras (photographing means) 20A and 20B for photographing the surface S of the measuring object T on which the pattern displayed on the display 10 is projected from different directions, the display 10 and the CCD cameras 20A and 20B. A computer 30 connected to the computer. As shown in FIG. 2, the computer 30 includes a control unit 31, a timing control unit 32, a pattern display control unit (shift means) 33, an imaging control unit 34, an image processing unit 35, an image memory 36, a surface shape. An analysis unit (surface shape analysis means) 37, a ROM (Read Only Memory) 38, and a RAM (Random Access Memory) 39 are provided.

  The display 10 is a display device that displays a pattern having a wavy optical density (brightness) distribution on a flat screen in accordance with a command from the pattern display control unit 33. Specifically, the display 10 is a liquid crystal display (LCD). ) Or CRT (Cathode Ray Tube) display. A high-resolution display having a large number of pixels is preferred.

  The CCD cameras 20A and 20B are imaging devices that capture images in accordance with instructions from the imaging control unit 34 and obtain image data. Specifically, CCD (Charge Coupled Device) pixels are arranged two-dimensionally. And a digital camera (digital still camera) using a CCD sensor and a digital video camera. The CCD cameras 20A and 20B are respectively fixed at predetermined intervals and postures by a pedestal such as a stereo head, and the overlapping ranges of the surface S of the measurement target T on which the pattern displayed on the display 10 is projected from different directions. Take a picture. It is preferable that the display 10 and the pedestal can be installed separately. In the CCD cameras 20A and 20B, the CCD sensor receives reflected light from the measurement target T, and outputs a gray level digital signal corresponding to the intensity of the received light as image data. When a point on the surface S of the measuring object T is photographed by the CCD cameras 20A and 20B, the CCD pixel that photographed the point obtains the light intensity at the point. A high-resolution CCD camera having a large number of pixels (pixels) is preferable.

  The computer 30 integrates and controls the operation processing of the display 10 and the CCD cameras 20A and 20B, and obtains the shape of the surface S of the measurement object T based on the image data obtained by the CCD cameras 20A and 20B. A device, specifically, a general-purpose personal computer or a dedicated computer. The computer 30 is connected to an input unit for inputting a measurement start instruction, a measurement type, and the like from the outside, a display unit for presenting the shape of the analyzed surface T of the measurement object T, a printer, and the like, and various networks. It is preferable to provide a line connection section for the purpose.

  The control unit 31 includes a timing control unit 32, a pattern display control unit 33, and an imaging control unit 34, and includes a CPU (Central Processing Unit) and the like, and controls each unit constituting the three-dimensional shape measuring apparatus 100.

  The timing control unit 32 controls the timing at which the pattern display control unit 33 gives a command to the display 10 and the timing at which the imaging control unit 34 gives a command to the CCD cameras 20A and 20B.

  The pattern display control unit 33 controls a pattern displayed on the display 10 and outputs a luminance level signal displayed on each pixel of the display 10 to the display 10. This luminance level signal is a signal that instructs the display 10 to display a pattern having a wavy optical density distribution on the screen. The pattern display control unit 33 outputs a luminance level signal for changing the phase of the pattern in response to a command from the timing control unit 32.

The pattern having a wavy optical density distribution is, for example, a pattern having a sinusoidal or triangular wave optical density distribution in one direction. The optical density F (dx) of the sinusoidal pattern can be expressed by Expression (1). Here, dx indicates the coordinate in the dX axis direction of the coordinate system fixed to the display 10 so as to coincide with the direction of the sinusoidal optical density distribution, and α indicates the phase. The amount of change of α is, for example, 2 / π, and the pattern is shifted by changing the phase α to 0, π / 2, π, 3π / 2.
F (dx) = F ₀ {1-sin (dx + α)} (1)

The optical density F (dx) of the triangular wave pattern can be expressed by equation (2). Here, dx indicates the coordinate in the dX-axis direction of the coordinate system fixed to the display 10 so as to coincide with the direction of the triangular optical density distribution, and α indicates the phase. Note that remnant (dx + α, 4) is a remainder obtained by dividing dx + α by 4. F (dx) is a repetition function with a period of 4. The amount of change of α is 1, for example, and the pattern is shifted by changing the phase α to 0, 1, 2, and 3.
F (dx) = F ₀ {1− (dx + α)}: 0 ≦ remnant (dx + α, 4) <2
F (dx) = F ₀ {dx + α−3}: 2 ≦ remnant (dx + α, 4) <4
(2)

  The pattern display control unit 33 changes the phase of a pattern having a sine wave or triangular wave optical density distribution in a predetermined direction (dX axis direction) and shifts the phase, and then a direction different from the direction (dX axis direction), For example, a pattern having a sinusoidal or triangular optical density distribution in the direction orthogonal to the dX-axis direction (dY-axis direction) is similarly shifted and shifted.

  The imaging control unit 34 controls the CCD cameras 20A and 20B, and simultaneously outputs trigger signals for instructing to perform imaging to the CCD cameras 20A and 20B. This luminance level signal is a signal that instructs the display 10 to display a pattern having a wavy optical density distribution on the screen. A trigger signal is output in response to a command from the timing control unit 32. As a result, the CCD cameras 20A and 20B can photograph the surface S on which the phase-shifted pattern is projected. When the trigger signal from the imaging control unit 34 is input, the CCD cameras 20A and 20B perform imaging and automatically output image data obtained by imaging to the image processing unit 35.

  The image processing unit 35 performs predetermined image processing on the image data input from the CCD cameras 20A and 20B and outputs the processed image data. The CCD cameras 20A and 20B have certain characteristics specific to the CCDs 20A and 20B, such as image distortion caused by various geometric distortions such as lens distortion, and an internal structure for calibrating these characteristics. The parameters are stored in the RAM 39. The image processing unit 35 performs correction processing on the image data output from the CCD cameras 20A and 20B based on the internal parameters of the CCD cameras 20A and 20B. The image processing unit 35 also performs correction processing such as shading correction and γ correction.

  The image memory 36 stores the image data image-processed by the image processing unit 35. In the image memory 36, a set of image data obtained by simultaneously photographing the CCD cameras 20A and 20B is stored in association with the waveform and phase of the pattern at the time of photographing.

  The surface shape analysis unit 37 performs processing based on various data settings stored in the RAM 39 on the image data stored in the image memory 36 according to the three-dimensional shape measurement program Pr stored in the ROM 38. The shape of the surface S of the object T is analyzed.

  The ROM 38 is a memory that stores various programs for controlling the operation of each unit constituting the three-dimensional shape measuring apparatus 100 by the control unit 31, and stores a three-dimensional shape measuring program Pr, for example.

  The RAM 39 is a memory that stores various data such as setting data and operation data used for the processing operation of the three-dimensional shape measuring apparatus 100 in a readable and writable state. For example, the internal parameters of the CCD cameras 20A and 20B, and wavy Are stored, for example, data relating to a luminance level signal based on a mathematical expression F (dx) indicating a waveform, a change amount of the phase α, and the like.

  The three-dimensional shape measurement method according to the present embodiment measures the shape of the surface S of the measurement target T with such a three-dimensional shape measurement apparatus 100. Hereinafter, a three-dimensional shape measurement method executed using the three-dimensional shape measurement apparatus 100 will be described. This three-dimensional shape measurement method includes an arrangement / calibration step (S1), a photographing step (S2), and a surface shape calculation step (S3), and is based on a three-dimensional shape measurement program Pr stored in the ROM 38. This is performed according to a command issued by the control unit 31.

  First, the display 10 and the two CCD cameras 20A and 20B are arranged with respect to the surface S to be measured of the measuring object T, and an arrangement / calibration step (S1) for performing calibration in the arrangement relationship is performed. First, the display 10 and the two CCD cameras 20A and 20B are arranged at appropriate positions with respect to the measurement target T. The measurer arranges the display 10 so that the pattern displayed on the display 10 is projected on the surface S of the measurement target T to be measured. The measurer arranges two CCD cameras 20A and 20B so that the surface S of the measurement object T to be measured can be photographed in an overlapping manner. The area of the surface S that can be measured by one measurement and the measurement accuracy depend on the size of the display 10 and the arrangement of the CCD cameras 20A and 20B. These positional relationships are fixed. Then, the CCD cameras 20A and 20B are calibrated. The internal parameters of the CCD cameras 20A and 20B are specific to the CCD cameras 20A and 20B, respectively, and are stored in the RAM 39 in advance. On the other hand, the external parameter indicating the relationship between the CCD cameras 20A and 20B is a parameter for calibrating the optical axis direction of each CCD camera 20A and 20B and the mutual positional relationship (three-dimensional geometric relationship) at the time of photographing. Yes, determined when CCD cameras 20A and 20B are fixed. This external parameter is obtained and stored in the RAM 39.

  Then, the positional relationship between the display 10 and the CCD cameras 20A and 20B is calibrated. The positional relationship between the display 10 and the CCD cameras 20A and 20B is calibrated using a calibration mirror. A plane mirror or a spherical mirror whose surface shape is known is used as a calibration mirror. For example, in the case of a plane mirror, the normal directions at all points on the surface are the same, and the unknown parameters are three of three-dimensional positions. On the other hand, there are six unknown parameters of the display 10: a three-dimensional position and a posture. Therefore, each parameter can be easily obtained by setting equations for nine parameters in total and using the least square method. In the case of a spherical mirror, since the radius of the spherical surface is known, there are three unknown parameters, that is, three-dimensional positions. Also, by attaching a plurality of targets to the plane mirror and analyzing the image data obtained by photographing with the CCD cameras 20A and 20B, the three-dimensional coordinates of each target are measured to obtain the three-dimensional position and orientation of the plane mirror. Thereafter, six parameters of the display 10 may be obtained. When the CCD cameras 20A and 20B are in a positional relationship in which the display 10 can be directly photographed, calibration can be performed without using a calibration mirror. That is, in the images obtained by directly photographing the display 10 with the CCD cameras 20A and 20B, the points on the display 10 having the same phase (for example, horizontal and vertical) are associated with each other, so that the three-dimensional of the point is obtained. The position can be obtained, and the positional relationship between the camera coordinate system fixed to the CCD cameras 20A and 20B and the display coordinate system fixed to the display 10 can be calculated.

  Next, a photographing step (S2) is performed in which the surface S of the measurement target T is photographed by the two CCD cameras 20A and 20B while shifting the phase of the pattern displayed on the display 10. The pattern display control unit 33 outputs a luminance level signal to the display 10 in response to a command from the timing control unit 32, and the display 10 displays a pattern having a wavy optical density distribution. The photographing control unit 34 outputs trigger signals to the two CCD cameras 20A and 20B in response to a command from the timing control unit 32, and the CCD cameras 20A and 20B photograph the surface S on which the pattern is projected. . The CCD cameras 20A and 20B each output image data obtained by photographing the surface S on which the pattern is projected. The output image data is subjected to image processing by the image processing unit 35 and stored in the image memory 36.

  For each command from the timing control unit 32, the pattern display control unit 33 sequentially shifts the phase in two different directions, for example, the dX axis direction and the dY axis direction, and a pattern having a wavy optical density distribution is displayed on the display 10. A luminance level signal is output so as to be displayed. Then, for each command from the timing control unit 32, the imaging control unit 34 uses the two CCD cameras 20A and 20B to capture the surface S on which a wave-like pattern in which the phase is sequentially shifted and the shifting direction is different is projected. Thus, a trigger signal is output. Thus, in synchronization with the phase shift of the pattern displayed on the display 10, the two CCD cameras 20A and 20B are used to obtain image data. The image memory 36 stores a set of image data that has undergone image processing after the CCD cameras 20A and 20B have taken images for each phase shifted in two different directions.

  Next, a surface shape calculation step (S3) for obtaining the shape of the surface S of the measurement target T using the image data stored in the image memory 36 is performed.

Hereinafter, a case where a sinusoidal pattern is displayed on the display 10 will be described. A sinusoidal pattern expressed by the equation (1) is displayed on the display 10, and when the surface S of the measuring object T on which the pattern is projected is photographed by the CCD camera 20A, a point PS on the surface _S is photographed. The light intensity I received by the point _Pc on the CCD pixel of the CCD camera 20A can be expressed by equation (3). Here, A is the amplitude of the waveform, and B is the background luminance independent of the waveform. φ is the modulation of the phase, representing the irregularities (modulation component of height) at the point P _S on the surface S.
I = Acos (φ + α) + B (3)

The phase α of the sinusoidal pattern displayed on the display 10 is shifted. For example, 0 the phase α of the pattern, [pi / 2, [pi, when is sequentially changed and 3 [pi] / 2, the light intensity I point P _c of the CCD pixels capturing the point P _S on the surface S has received respectively Assuming I0, I1, I2, and I3, these are given by equation (4).
I0 = Acosφ + B
I1 = Asinφ + B
I2 = −Acosφ + B
I3 = −Asinφ + B (4)

From equation (4), φ is given by equation (5).
φ = tan ⁻¹ {(I3−I1) / (I0−I2)} (5)

  Since the sinusoidal pattern expressed by Equation (1) has a sinusoidal optical density distribution in a predetermined direction (dX axis direction), the pattern α is changed in the dX axis direction. Since the phase α does not change in the direction perpendicular to the dX axis only by shifting to λ, the phase modulation amount φ can be obtained. Therefore, by obtaining a pattern having a sinusoidal optical density distribution in a direction different from the dX axis, for example, the dY axis direction, by changing the phase α in the direction, the phase modulation φ at the lattice point on the surface S is obtained. Can do.

  In this example, φ is obtained from the light intensity I obtained by imaging at each of four phases α. However, it is possible to obtain φ by photographing more than this number of phases α. Thereby, the measurement accuracy is improved. Further, since the phase shift is periodic, it is necessary to uniquely specify the phase over a wide range by phase connection in order to distinguish it from the phase of the adjacent period. Therefore, in the case of a large screen display, one phase connection may not be able to extend over the entire screen. In this case, the cycle may be changed a plurality of times.

From the above, it is possible to point _{P c} on the CCD pixels of the CCD camera 20A is to obtain the phase modulation amount φ at a point _{P S} on the surface S of the received light. The light receiving direction at the point _Pc on the CCD pixel by the CCD camera 20A can be determined because the external parameters of the CCD camera 20A are known. When a vector indicating the shooting direction at the point P _c and vector C, as shown in FIG. 3, the point P _S is positioned on a straight line extending from the point P _c in the direction of the vector C.

Since the surface S of the measuring object T is specular reflection, the incident angle and the reflection angle sandwiched normal direction of small contact surfaces at a point P _S on the surface S (output angle) is equal relationship. Therefore, the incident direction toward the point P _S from the point P _D on the display 10 pattern displayed was the Projected in terms P _S on the surface S, the point on the display 10 to display the projected pattern at a point P _S in a direction equally dividing the incident direction from P _D to the point P _S, toward the normal direction of the contact surface at the point P _S. Therefore, the normal vector N of the contact surface at the point P _S is the point P _D from the point P _S normalized vector Vector DS to unit length towards and the converted value to normalize the dimensions in the vector C and Z , Given by equation (6).
N = − (DS + Z · C) / 2 (6)

Since Z is an unknown number in equation (6), the normal vector N cannot be obtained only from image data obtained by photographing with one CCD camera 20A. But two CCD cameras 20A, by associating the image data 20B is obtained by photographing, it is possible to obtain the normal vector N, the further three-dimensional coordinates of a point P _S on the surface S having a phase α Can be sought.

Specifically, an epipolar line is obtained by projecting a straight line including the vector C from image data obtained by one CCD camera 20A onto an image comprising image data obtained by the other CCD camera 20B. Along the epipolar lines, the corresponding point on the condition that it has a common normal line direction, i.e., searching for a point P _S on the surface S. Since the number of points satisfying such a condition is generally limited to one point, a corresponding point can be easily obtained.

Hereinafter, a case where a triangular wave pattern is displayed on the display 10 will be described. A triangular wave pattern expressed by Equation (2) is displayed on the display 10, and when the surface S of the measuring object T on which the pattern is projected is photographed by the CCD camera 20A, a point PS on the surface _S is photographed. The light intensity I received by the point _Pc on the CCD pixel of the CCD camera 20A can be expressed by equation (7). Here, A is the amplitude of the waveform, and B is the background luminance independent of the waveform. φ is the modulation of the phase, representing the irregularities (modulation component of height) at the point P _S on the surface S.
I = A {1- (α + φ)} + B: 0 ≦ remnant (α + φ, 4) <2
I = A (α + φ−3) + B: 2 ≦ remnant (α + φ, 4) <4
(7)

The phase α of the triangular wave pattern displayed on the display 10 is shifted. For example, when the phase α of the pattern is sequentially changed and 0,1,2,3, the light intensity I point _{P c} of the CCD pixels capturing the point _{P S} on the surface S has received respectively I0, I1, Assuming I2 and I3, these are given by equation (8).
I0 = A (1-φ) + B
I1 = -A ・ φ + B
I2 = A (φ−1) + B
I3 = A · φ + B (8)

Equation (8) is modified as follows.
C = I0-I2 = 2A (1-φ)
D = I3-I1 = 2A · φ
C + D = 2A (9)

From equation (9), φ is given by equation (10).
φ = D / (C + D) (10)

  C and D are both integers, and division can be omitted by preparing a LUT (Look Up Table). For imaging, φ can be obtained from only the light intensities I0, I1, and I3 obtained by changing the pattern phase α to 0, 1, 2, and 3. However, since there is no redundancy, the influence of noise increases and the measurement accuracy decreases, so the phase α is changed in four ways. It is also possible to take images of four or more phases α and obtain φ, thereby improving the measurement accuracy. In the above, it is assumed that 0 ≦ α <1, but since actually 0 ≦ α <4, the shift is made in the right direction in advance so that I0> I1 and I2 <I3. . For example, if I0, I1, I2, and I3 are in the order of small, small, large, and large, by shifting once in the right direction, the order becomes large, small, small, and large. If the number of shifts is k, the actual phase is k + α.

Thus, it was possible to calculate the phase of the modulation component φ at a point P _S on the light receiving surface S to point P _c on the CCD pixels of the CCD camera 20A is taken. Thereafter, the shape of the surface S of the measuring object T can be obtained in the same manner as when a sinusoidal pattern is displayed on the display 10.

As described above, according to the three-dimensional shape measuring apparatus 100 and the three-dimensional shape measuring method according to the embodiment of the present invention, the overlapping ranges of the mirror-reflected surface S of the measurement target T can be detected from two different directions by the CCD cameras 20A and 20B. There the correspondence of the image data obtained by photographing performed so that the normal direction of the contact surface at the point P _S on the surface S which Projected pattern equal, determining the shape of the surface S of the range. Therefore, the shape of the surface S that is specularly reflected can be measured with high accuracy.

  Further, the CCD cameras 20A and 20B take images in synchronization with the shift of the pattern displayed on the display 10. Therefore, each CCD camera 20A, 20B can sequentially capture the surface S on which the pattern to be shifted is imaged to obtain image data, and measurement can be performed at high speed. Further, the pattern displayed on the display 10 is a pattern having a sine wave or triangular wave density distribution. Therefore, phase analysis is easy, and the shape of the surface S can be measured with higher accuracy and higher speed. In the case of using a pattern having a triangular wave-like density distribution, numerical calculation of a trigonometric function is not required as compared with a pattern having a sinusoidal density distribution, and measurement can be performed at higher speed.

  Note that the three-dimensional shape measuring apparatus 100 and the three-dimensional shape measuring method shown in the present embodiment are only one aspect of the three-dimensional shape measuring apparatus and the three-dimensional shape measuring method according to the present invention, and the gist of the present invention is as follows. Various modifications can be made without departing from the scope. For example, although an example in which a pattern is shifted and displayed on the screen of the display 10 has been shown, a pattern projector that projects a pattern formed on a pattern plate or the like onto the surface S may be used. Specifically, a liquid crystal projector or a slide projector can be used. Alternatively, a plate or cloth having a pattern having a wavy density distribution formed on the surface thereof may be used to imprint the pattern on the surface S. An expensive phase shifter is required to shift the pattern with high precision by moving the pattern plate or the like on which the pattern is formed.

  Moreover, although the example in which each pixel of the display 10 and the CCD cameras 20A and 20B is related is shown, it may be related in units of sub-pixels, thereby enabling measurement with higher accuracy. Further, although an example in which two CCD cameras 20A and 20B are used has been shown, photographing may be performed by moving one CCD camera in two predetermined sections. Three or more CCD cameras may be used. Thereby, a measurement range can be enlarged and measurement accuracy can be improved.

  Further, when the shape of the surface S to be measured of the measurement target T cannot be measured by a single measurement, a target is pasted on the surface S, and the surface S on which the target is pasted by moving the measurement target T is attached. By capturing images in different measurements, the shape of the surface S can be determined in an integrated manner.

  Although an example in which the wavy pattern displayed on the display 10 is shifted in different directions has been shown, the pattern displayed on the display 10 can be shifted only in one direction. Thereby, when using a pattern projector, the number of pattern plates is one, and the shift by motor control becomes easy. Hereinafter, an embodiment in this case will be described.

The pattern is simply a triangular waveform with two different periods. The optical density F (dx) of the pattern can be expressed by equation (11). F _A (dx) is a trigonometric function that repeats at a period of 4, and F _A (dx) is a trigonometric function that repeats at a period of 16n (n is a natural number of 2 or more).
F (dx) = F _A (dx) + F _B (dx)
F _A (dx) = F _A0 {1- (dx + α)}: 0 ≦ remnant (dx + α, 4) <2
F _A (dx) = F _A0 {dx + α−3}: 0 ≦ remnant (dx + α, 4) <4
F _B (dx) = F _B0 {1- (dx + α)}: 0 ≦ remnant (dx + α, 4) <8n
F _B (dx) = F _B0 {dx + α−3}: 8n ≦ remnant (dx + α, 4) <16n
(11)

Pattern represented by the formula (11) is displayed on the display 10, when the surface S of the measuring object T which this pattern was projected photographed by the CCD camera 20A, a CCD camera captured the point P _S on the surface S The light intensity I received by the point _Pc on the 20A CCD pixel can be expressed by equation (12). Here, A and E are the amplitude of the waveform, and B is the background luminance independent of the waveform. φ is the modulation of the phase, representing the irregularities (modulation component of height) at the point P _S on the surface S.
I = I _A + I _B + B
I _A = A {1- (α + φ)}: 0 ≦ remnant (α + φ, 4) <2
I _A = A {α + φ-3}: 2 ≦ remnant (α + φ, 4) <4
I _B = B {1- (α + ψ)}: 0 ≦ remnant (α + ψ, 4) <8n
I _B = B {α + ψ−3}: 8n ≦ remnant (α + ψ, 4) <16n (12)

The phase α of the pattern displayed on the display 10 is shifted in seven ways. For example, when the phase α of the pattern 0,1,2,3,4N, 8n, sequentially changing the 12n, light intensity point _{P c} of the CCD pixels capturing the point _{P S} on the surface S has received I Are I 0, I 1, I 2, I 3, I 4, I 5, I 6, and I 7, respectively, and these are given by equation (13).
I0 = A (1-φ) + E (1-ψ) + B
I1 = −A · φ + E (1−ψ−φ / 16n) + B
I2 = A (φ-1) + E (1-ψ-2φ / 16n) + B
I3 = A · φ + E (1-ψ-3φ / 16n) + B
I4 = A (1−φ + 4n) + E (−ψ) + B = A (1−φ) + E (−ψ) + B
I5 = A (1−φ + 8n) + E (φ−1) + B = A (1−φ) + E (φ−1) + B
I6 = A (1−φ + 12n) + E · ψ + B = A (1−φ) + E · ψ + B (13)

  By using Expression (10), ψ can be obtained from I0, I4, I5, and I6. Further, φ can be obtained by subtracting the portion related to ψ from I 0, I 1, I 2, and I 3 in equation (13) and using equation (10) again.

By obtaining ψ, it is possible to estimate how many periods of F _A (dx) constituting a part of the pattern. As shown in Expression (14), the quotient obtained by dividing ψ × 16n by 4 is the period k of F _A (dx), and the phase after connection is 4k + φ.
k = quotient (ψ × 16n, 4) = ψ × 16n (14)

  If the pattern having a triangular wave-like density distribution is shifted and the shape of the surface is obtained based on image data obtained by photographing the surface of the measurement object on which the pattern is copied, the above embodiment is used. It is not limited to. For example, a pattern having a triangular wave-shaped density distribution is shifted and displayed on the display, and the mirror-reflecting surface of the measurement object on which the pattern is projected is captured based on image data obtained by photographing with a single CCD camera. The surface shape may be obtained by the conventional phase shift method. Also, based on image data obtained by shifting a pattern having a triangular wave-shaped density distribution and projecting it from a projector, and photographing a non-specularly reflected surface of the measurement object on which the pattern is projected with a CCD camera, a well-known The surface shape may be obtained by a phase shift method.

It is a conceptual explanatory view showing a three-dimensional shape measuring apparatus 100 according to an embodiment of the present invention. 1 is a conceptual configuration diagram showing a three-dimensional shape measuring apparatus 100. FIG. Is a diagram illustrating a mirror reflection of light at the point P _S on the surface S.

Explanation of symbols

10 Display (display means, shift means)
20A, 20B CCD camera (photographing means)
30 Computer (shift means, surface shape analysis means)
T Measurement object S Surface

Claims (6)

A three-dimensional shape measuring apparatus for measuring a mirror-reflecting surface shape of a measurement object,
Display means for displaying a pattern having a wavy density distribution;
Shift means for shifting the pattern displayed on the display means;
Photographing means for photographing the surface of the measurement object, which is a pattern displayed on the display means, from at least two different directions;
A three-dimensional shape measuring apparatus, comprising: surface shape analyzing means for obtaining a surface shape of the range by associating image data obtained by photographing the overlapping range of the surface with the photographing means.   The surface shape analyzing means obtains the photographing means by photographing the points in duplicate so that the normal directions of the surface at the points on the surface where the pattern displayed on the display means is projected are equal. The three-dimensional shape measuring apparatus according to claim 1, wherein the image data is associated.   The three-dimensional shape measuring apparatus according to claim 1, wherein the photographing unit is a plurality of CCD cameras that photograph in synchronization with the shift of the pattern.   The three-dimensional shape measuring apparatus according to any one of claims 1 to 3, wherein the display means displays a pattern having a sinusoidal or triangular wave density distribution. A three-dimensional shape measuring method for measuring a specularly reflecting surface shape of a measurement object,
Shift and display a pattern with a wavy concentration distribution,
Photographing the surface of the measurement object on which the pattern is projected from at least two different directions,
A three-dimensional shape measuring method, wherein the surface shape of the range is obtained by associating image data obtained by photographing the overlapping range of the surface. A three-dimensional shape measuring method for measuring a surface shape of a measurement object,
Shift the pattern with triangular wave density distribution,
Photograph the surface of the object to be measured in which the pattern is projected,
A three-dimensional shape measuring method, wherein the shape of the surface is obtained based on image data obtained by photographing the surface.

Images