JP5369564B2 - Shape measuring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To measure the profile of a testing object without moving the testing object along the light axis of an objective lens. <P>SOLUTION: An MLA 12 is placed in a pupil plane or the imaging plane of the objective lens 11, and an imaging element 13 is placed behind the MLA 12, and images the testing object. A zoom lens driving apparatus 51 expands or reduces the size of the image of the object by moving a zoom lens ZL, and a focusing lens driving apparatus 52 brings the testing object in focus by moving the focusing lens FoL. A calculation process circuit 34 generates a plurality of measuring images from the output of the image taking element 13 and measures the profile of the testing object by using a reference image obtained from the pixels in the position crossing the principal ray of the objective lens 11, among the pixels composing the imaging area of the imaging element 13 corresponding to each of a plurality of MLs arranged on the MLA 12, as the correction information. The present invention is applicable to the profile measuring apparatus for measuring the 3-dimensional profile of the testing object. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a shape measuring apparatus.

  While moving the test object along the optical axis of the objective lens, obtain multiple images with different focal positions, and select the image with the maximum focus measure for each point on the test object from the multiple images. It is known that the focus position in the image is the height of the point and the shape of the test object is measured from these pieces of information.

This is generally a technique such as SFF (Shape From Focus).
Japanese Patent No. 2960684

  However, in such a conventional technique, it is necessary to acquire an image while moving the test object along the optical axis of the objective lens, and the measurement takes time.

  The present invention has been made in view of such a situation, and makes it possible to measure the shape of a test object without moving the test object along the optical axis of the objective lens.

The shape measuring apparatus according to one aspect of the present invention includes an objective lens, an optical element having a plurality of lenses arrayed two-dimensionally behind the objective lens, and a two-dimensional imaging element disposed behind the optical element. A focal length changing unit that changes a focal length of the objective lens, an image forming unit that generates a plurality of measurement images having different focal planes from the output of the imaging element, and the imaging corresponding to each of the plurality of lenses obtained from a given pixel in accordance with the optical arrangement of the objective lens among the pixels constituting the imaging area of the element and the optical element, and the reference image in which all pixels are in focus, the correlation between the plurality of measurement images using the position of the optical axis of the objective lens of the object sought, characterized in that it comprises a three-dimensional shape measurement means for measuring the three-dimensional shape of the test object.

  The apparatus further includes a focus position varying unit that changes a focus position of the objective lens.

The three-dimensional shape measuring means obtains distance information to the test object at the approximate center of the field of view of the objective lens, and the focus position varying means is the distance information measured by the three-dimensional shape measuring means. The focus position is controlled based on the above.

The three-dimensional shape measurement means, and the distance to the test object, taking into account the magnification of the objective lens, and measuring the three-dimensional shape of the test object.

In one aspect of the present invention, the focal length of the objective lens is changed, and a plurality of measurement images having different focal planes are generated from the output of the imaging device, thereby forming an imaging region of the imaging device corresponding to each of the plurality of lenses. The light of the objective lens of the test object obtained from the correlation between a reference image obtained by focusing on all the pixels and a plurality of measurement images, obtained from a predetermined pixel corresponding to the optical arrangement of the objective lens and the optical element among the pixels. The three-dimensional shape of the test object is measured using the position in the axial direction .

  According to the present invention, the shape of the test object can be measured without moving the test object along the optical axis of the objective lens.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

In the present invention, the amount of blur (including directionality) (hereinafter also referred to as focus information) from a plurality of virtual object planes (Z ₁ , Z ₂ ,...) Is calculated without moving the lens. In addition, an omnifocal image without blur can be obtained at the same time. Thereby, in this invention, it becomes possible to perform the distance measurement using those information.

  Specifically, from the distance a between the lens and the imaging surface (hereinafter also referred to as the lens position a) and the focal length f of the lens, the measurement point of the object to be measured is calculated according to the Gauss formula of the following equation (1). The distance b can be known.

  1 / b = 1 / f−1 / a (1)

In the measurement according to the present invention, in addition to the lens position a determined mechanically, the pixel signal on the image sensor that has passed through a microlens array (hereinafter referred to as MLA) is recombined and added to virtually add the lens. The imaging plane intervals a ₁ , a ₂ , a ₃ ,..., _An are constructed. A value string of values b obtained by substituting these values into the value a in the equation (1) is the position of a plurality of object planes (Z ₁ , Z ₂ , Z ₃ ,... Z _n ). That is, in such measurement, it is possible to reconstruct an image at that time and obtain focus information on the assumption that light beams emitted from different object surfaces have reached the imaging surface by such reconstruction.

  In this way, the same information as in the conventional SFF method can be obtained without moving the test object along the optical axis of the lens.

  In addition, since the position in such focus detection can be executed by the number of MLAs, how finely the test object can be measured (spatial resolution) is determined by the arrangement of the MLA. In other words, the larger the number of MLA, the higher the spatial resolution of the measurement.

  Next, details of the shape measuring apparatus for measuring the shape of the test object using the above principle will be described for each component.

[1. Optical system]
First, an outline of an optical system of a shape measuring apparatus to which the present invention is applied will be described with reference to FIG.

  As shown in FIG. 1, the optical system is configured to include an objective lens 11, an MLA 12, and an image sensor 13.

  FIG. 1 shows the configuration of the two optical systems shown in FIGS. 1a and 1b. In the configuration shown in FIG. 1a, the focal point of the objective lens 11 that captures an image of the test object (object plane (P plane)) is shown. The MLA 12 is placed at a position separated by a distance, and the image sensor 13 is placed at a position separated by the focal length of the MLA 12. In FIG. 1a, the imaging surface of the imaging device 13 is conjugate with the object plane (P plane), and an image of the test object is formed.

  The MLA 12 is an optical element in which a plurality of microlenses (hereinafter referred to as ML), such as 7 × 7, are arranged in a two-dimensional manner, similar to the above-described MLA (in FIG. 1, seven in the vertical direction are arranged). Shown). A region where an image formed by each ML can be formed, that is, an image signal obtained from the ML region 13ML is an image of a different viewpoint of the test object, that is, an image having a parallax according to the position of each ML region. Yes.

  On the other hand, in the configuration of FIG. 1b, the MLA 12 is placed on the surface where the object plane P at a predetermined position forms an image with the objective lens 11, and the image sensor 13 is placed at a position separated by the focal length of ML. In this case, the surface of the image sensor 13 is conjugate with the pupil plane of the objective lens 11, and the image signal in each ML region is an image obtained by dividing the pupil. In other words, in FIG. 1b, the imaged rays are distributed into rays at different angles as seen from the lens pupil.

  In the present invention, it is possible to obtain a plurality of images with different focal planes by reconstructing light rays from different angles from either of the above-described configurations of FIG. 1a or FIG. 1b. Measure the shape of the specimen. Therefore, in the following description, the description will focus on the configuration of FIG. 1a, but the same applies to the configuration of FIG. 1b.

Then, with reference to FIG.2 and FIG.3, the further detailed structure of the optical system of FIG. 1a is demonstrated next. In FIG. 2, three MLs of microlens ML _a , microlens ML _b , and microlens ML _c (hereinafter referred to as ML _a , ML _b , and ML _c , respectively) as a part of MLA 12 composed of a plurality of MLs. For illustrative purposes).

In Figure 2, the light from point P1 on the object plane (P-plane), is refracted after passing through the objective lens 11 as a parallel to the optical axis direction of the light (solid line in the drawing), captured by ML _b element 13 light ray group is focused on a point P1 _b of the ML area on the imaging surface of the (solid line in the figure) and, after passing through the center of the objective lens 11, the ML _c, the ML area on the imaging surface of the imaging device 13 and the like light ray group is focused on the point P1 _c (dotted line in the figure).

In this way, each ML placed on the pupil plane of the objective lens 11 forms an object image independently on the imaging plane. At this time, the ML region point P1 _b and the ML region point P1 _c on the imaging surface receive light rays emitted from the same point P1 on the object surface (P surface) at different angles. .

Similarly, the light beam from point P2 on the object plane (P-plane), passes through the refractive objective lens 11 by injection in the direction indicated by the dashed line in the figure, on the imaging surface of the image sensor 13 by the ML _a as shown in rays that are imaged on a point P2 _a ML-region, with a solid line in the figure, after passing through the objective lens 11 is injected in a direction parallel to the optical axis of the objective lens 11, imaging by ML _b element 13 rays imaged at the point P2 _b of ML areas on the imaging surface, and as shown by a dotted line in FIG refracted after passing through the objective lens 11 is injected in downward direction, the image pickup device by ML _c rays imaged is shown at point P2 _c of ML areas on the imaging surface 13. Further, the light rays from point P3 on the object plane (P-plane), is refracted after passing through the objective lens 11 is injected upward, as shown by a chain line in the figure imaged by ML _a device 13 light is imaged on a point P3 _a ML-region on the imaging surface of, and exit in the direction parallel to the optical axis of the objective lens 11 as shown by the solid line in the figure, the imaging surface of the imaging device 13 by the ML _b The rays that are imaged at point P3 _b in the upper ML region are shown.

Here, if it is assumed that the imaging system includes only the objective lens 11, for example, light emitted from the point P <b> 2 is condensed at one point by the objective lens 11. That is, in the optical system of FIG. 2, the light intensity from all directions emitted from the point P2 such as the points P2 _a , P2 _b , and P2 _c is added, and this becomes a pixel signal. This means that in the optical system of the present invention as shown in FIG. 2, the light emitted from one point on the test object can be sorted and captured by each ML by the ML of the MLA 12. ing.

  FIG. 3 shows that an image viewed from one ML is an image of the P plane composed of light rays emitted from the same direction and emitted from the P plane. This shows that images viewed from different directions by the number of MLs can be obtained. That is, as many as the number of MLs can be used, but this means that light emitted from the P plane enters the corresponding ML and forms an image of the P plane. In other words, parallax images that differ by the number of MLs are obtained.

  Among the light rays shown in FIGS. 3a to 3c, in FIG. 3b, the image by ML in the central portion that collects the light entering perpendicularly to the objective lens 11 is an image obtained by capturing the P plane from the front. 3a and 3c show that the image by ML located toward the end where the light incident obliquely on the objective lens 11 is collected is an image obtained by viewing the P surface from the angle direction of the light emitted from the P surface. Become.

  The pupil plane of the objective lens 11 is divided into the size of each ML of the MLA 12, and each ML forms the same object image. At this time, the image in the ML near the edge is blurred in the vicinity (where the object cannot be seen) and a part of the visual field near the edge is blocked, but basically, the image by each ML is blocked. Is an image obtained by viewing the object plane (P plane) from a different angle at a position conjugate with the object plane (P plane).

  The NA (open angle) of each of the N MLA images is reduced to 1 / N by an opening determined by the aperture of each ML. What should be noted here is that each ML image has a small angle of view from the image forming side and a small amount of light, but the depth of field (depth of focus) becomes very deep. Therefore, regardless of the depth on the object side, the image is almost in focus, that is, a so-called omnifocal image.

Further, an image formed by ML (corresponding to ML _b in FIG. 2) on the optical axis of the objective lens 11 in FIG. 3B (corresponding to the center of the MLA 12 in FIG. 2) has an object position in the optical axis direction. Even if it moves, the image stays in focus and does not move. That is, the image hardly changes even if the position of the object in the optical axis direction changes.

On the other hand, when the ML deviated from the center in FIGS. 3a and 3c (corresponding to ML _a and ML _c in FIG. 2) is viewed, the object position is away from the objective lens 11 as shown by the left arrow in the figure. When moving away, the image to be imaged is shifted in the direction of the upward arrow in FIG. 3a and shifted in the direction of the downward arrow in FIG. 3c. In other words, the pixel position where the light beam strikes shifts in a direction away from the optical axis. Conversely, when the object position is close, the pixel positions of these images are shifted in the direction approaching the optical axis.

  The optical system of the present invention is configured as described above.

[2. Focused image generation procedure]
(2-1) When the MLA 12 is used as the pupil plane of the objective lens 11 Next, referring to FIGS. 4 and 5, the object plane at a plurality of different positions when the optical system of the present invention is employed. A method for detecting the focus information at each point of the object by forming the image or signal will be described. In the present embodiment, as described with reference to FIG. 1, regardless of the configuration of FIG. 1a or FIG. 1b, a plurality of beams having different focal planes are reconstructed by reconstructing rays incident from different angles. It is possible to generate images. Therefore, first, (1) the case where the MLA 12 is placed on the pupil plane of the objective lens 11 will be described, and then (2) the case where the MLA 12 is placed on the imaging plane of the objective lens 11 will also be explained.

  FIG. 4 shows a ray tracing diagram of the objective lens 11 and the MLA 12, and FIG. 5 shows a part of the pixel array on the surface of the image sensor 13 imaged by the MLA 12 and a selection for signal processing for each MLA region. A pixel is shown.

  By the way, the image pickup device 13 disposed on the rear side of the MLA 12 is, for example, a CCD (Charge Coupled Device) sensor, a CMOS (Complementary Metal Oxide Semiconductor) sensor, or the like, and receives a predetermined light that passes through each ML of the MLA 12. The pixel array is arranged in an arrangement pattern corresponding to ML. The number of pixels in the vertical direction and the horizontal direction of this pixel array is appropriately set so that, for example, partial light beams individually transmitted through MLs such as 7 × 7 can be individually received. Here, an example in which a CCD is used as the image sensor 13 will be described.

In FIG. 5, 21 × 21 squares indicate pixels arranged on the imaging surface when the imaging device 13 is viewed from the optical axis direction of the objective lens 11, and 7 × surrounded by a thick line among them. Each of the seven square areas represents one ML area. In FIG. 5, 3 × 3 ML regions are representatively shown in the 7 × 7 ML regions. Therefore, the center area among the ML areas surrounded by the thick line in the figure is the ML area formed by the ML at the center of the objective lens 11 (corresponding to ML _b in FIG. 2).

Figure 4a is a ray diagram in the case where the object plane P0 (the same as the focal plane Z ₀ are focused) is in the focus position. Here, the object plane P0 is uniquely determined by the focal length f of the optical system and the distance a between the objective lens 11 and the imaging surface of the imaging element 13, and specifically, the Gaussian equation of the optical system described above. This is the position of the distance b determined by (1).

  The light emitted from the position that coincides with the optical axis of the objective lens 11 on the object plane P0 is the ML of each MLA 12, and the center of each ML area on the CCD (imaging device 13) surface (coincides with the optical axis of each ML). Condensed to Therefore, as shown in FIG. 5 (as already described, only the pixel area corresponding to each ML region around the ML at the position corresponding to the optical axis of the objective lens 11 is shown), Since the light emitted from the point P02 at the center of the object plane P0 (which coincides with the optical axis of the objective lens 11) is focused on the central pixel (black square in FIG. 5) in each ML region, each pixel area The total luminance of the point P02 on the object plane P0 can be obtained by selecting and adding the outputs of the center pixel.

Light from a position deviating from the point P02 on the object plane P0 is incident on the MLA 12 in parallel at an angle corresponding to the amount of deviation from the point P02, so that these lights have the same amount from the center and in the same direction in each ML region. It will shift. Therefore, selecting the pixel of the same position in each ML regions, by adding, it is possible to reproduce an image of the object plane P0 in the focal plane Z _0.

4b shows an object plane P1 closer to the objective lens 11 than the object plane P0 (actually, as indicated by the dotted object plane P0 in FIG. 4b, the object plane P0 and the imaging plane of the image sensor 13 as in FIG. 4a). Is a light ray diagram from Light emitted from a position that coincides with the optical axis of the objective lens 11 on the object plane P1 is incident on each ML of the MLA 12 at an angle slightly shifted outward from the optical axis of each ML toward the end of the lens. Is done. In this case, the light emitted from the point P12 at the center of the object plane P1 (which coincides with the optical axis of the objective lens 11) is imaged by the center ML region (ML _{b in} FIG. 2) as shown in FIG. In the other ML region, the pixels in the other ML regions are left as they are, and the pixels on the object plane P0 shown in FIG. The total luminance of the point P12 on the object plane P1 can be obtained by selecting and adding the pixels.

  Since light from a position deviating from the point P12 on the object plane P1 enters the MLA 12 in parallel at an angle corresponding to the amount of deviation from the point P12, these lights are the same amount from the center and in the same direction in each ML region. It will shift. Therefore, by selecting and adding pixels at the same position in each ML region, the image of the object plane P1 can be reproduced.

4c shows an object plane P2 that is farther from the objective lens 11 than the object plane P0 (actually, as indicated by the dotted object plane P0 in FIG. 4c, the object plane P0 and the imaging plane of the image sensor 13 are similar to FIG. 4a. Is a light ray diagram from Light emitted from a position that coincides with the optical axis of the objective lens 11 on the object plane P2 is incident on each ML of the MLA 12 at an angle that slightly shifts inward from the optical axis of each ML toward the end of the lens. Is done. In this case, the light emitted from the point P22 at the center of the object plane P2 (which coincides with the optical axis of the objective lens 11) is imaged by the center ML region (ML _{b in} FIG. 2) as shown in FIG. The pixels in the other ML regions are left as they are, and in the other ML regions, the pixels on the inner side of the pixel position on the object plane P0 shown in FIG. Since the light is condensed into a square, the total luminance of the point P22 on the object plane P2 can be obtained by selecting and adding the pixels.

  Since light from a position deviating from the point P22 on the object plane P2 enters the MLA 12 in parallel at an angle corresponding to the amount of deviation from the point P22, these lights are the same amount from the center and in the same direction in each ML region. It will shift. Therefore, by selecting and adding pixels at the same position in each ML region, an image of the object plane P2 can be reproduced.

In FIG. 4, the rectangle 40 on the right side of the image sensor 13 schematically represents 13ML ₁ to 13ML ₇ that are part of the ML region, and the light condensing position in each ML region is black. By filling in, the above description in FIG. 5 is supplemented and conceptually described. That is, in the present embodiment, since an example of 7 × 7 ML areas is taken up, each of the ML areas 13ML ₁ to 13ML ₇ in the vertical direction is schematically shown, and the light is condensed in each ML area. The state of the (imaged) light is shown. In FIG. 4a, the line from the object plane P02 is focused (imaged) at the center of each ML region on the CCD (image sensor 13). Also, as shown in FIG. 4b, the light from the object plane P1 closer to the objective lens 11 than the object plane P0 is collected in the center in the central ML region (although not completely imaged, the ML aperture) In other ML regions, the light is condensed outward as compared with the case of FIG. 4a. Furthermore, as shown in FIG. 4c, the light from the object plane P2 farther from the objective lens 11 than the object plane P0 is collected in the center in the central ML region (although it is not completely imaged, the ML aperture) In other regions, the light is condensed inward compared to the case of FIG. 4a.

  If further captured, in the present embodiment, the front object plane P1 and the rear object plane P2 are actually the object plane P0 determined by the position of the objective lens 11 (the object plane P0 in FIG. In other words, it is a virtual focusing surface that can be obtained without moving the objective lens 11 other than the light receiving surface.

As described above, as in the object plane P0, P1, P2, even when changing the position of the optical axis direction of the object from the object plane P0, (corresponding to the Figure 2 ML _b) the center of the ML of the objective lens 11 As a result, the position of the image does not shift while the image is focused.

That is, (corresponding to P2 _b in FIG. 2) image formed by the ML of lens center of the optical system configuration (corresponding to ML _b in FIG. 2) shown in Figure 3b, the objective lens 11 It consists of light rays that enter vertically, and it is an image that captures the object from the front. Therefore, even if the object position moves in the optical axis direction, the image does not shift. And since it is the image from the divided | segmented small pupil, F value is large and the depth of field (depth of focus) becomes very deep. That is, even if the position of the test object in the optical axis direction changes, the image hardly changes. Therefore, regardless of the depth on the object side, the image is almost in focus, that is, a so-called omnifocal image.

(2-2) When the MLA 12 is placed on the imaging surface of the objective lens 11 By the way, in the above (2-1), an example in which the configuration of FIG. The case where it is set on the pupil plane of the lens 11 has been described. The method of acquiring the omnifocal image always in focus is not limited to the method (2-1), and as shown in FIG. 1B, the MLA 12 is placed on the focal (imaging) plane of the objective lens 11, ie, the object The same effect can be obtained even when the position is conjugate to the surface. Therefore, next, as a second example of a method for acquiring an omnifocal image, (2-2) a case where the MLA 12 is set on the image plane of the objective lens 11 will be described. In this case, since the procedure for creating an image on a different object plane is different from that of the method (2-1), here, the procedure for creating the image will be mainly described with reference to FIGS.

  6 to 8, in order to make the explanation easy to understand, each light ray (the center of the corresponding microlens ML is incident on each of the five pixels a, b, c, d, and e aligned on the straight line in the image sensor 13. Only the chief ray passing through). Further, each element in each drawing is given a suffix (1, 2, 3,...) For indicating coordinates in a plane perpendicular to the optical axis.

That is, as shown in FIG. 6, an image of a test object by an objective lens 11 (not shown) is imaged on the MLA 12 (Z = 0) (regions incident on each ML are indicated by broken lines, and the coordinates of the center are X ₁ to X ₇ ), rays r ₁ , r ₂ , r ₃ , r ₄ , and r ₅ from an objective lens 11 (not shown) are image planes (on the Z = 0 plane), that is, ML. Gather at one point on the surface. Each region indicates the center coordinates in X ₁ to X X _7, since subject to the same action, will be described below the central region. The light exiting from the center X ₄ in the central region corresponds to each pixel on the surface of the image sensor 13 by the lens action of the microlens ML ₄ , and a ₄ , b ₄ , c ₄ , d ₄ , e ₄ pixel signals. Therefore, each image signal on the imaging plane (Z = 0) is the sum of these. That is, if the image signal is L, L can be obtained by the following equation (2).

L (i) = (a _i + b _i + c _i + d _i + e _i ) (2)

However, as described above, i represents the coordinate in the X direction (i = 1 to 7) in the plane perpendicular to the optical axis. For example, in the case of the coordinate X ₄ , L (4) = (a ₄ + B ₄ + c ₄ + d ₄ + e ₄ ). Actually, since it is necessary to think two-dimensionally, it is necessary to consider the Y direction in addition to the X direction, but only the X direction will be described here in order to simplify the description.

On the other hand, as shown in FIG. 7, the image signal of the center coordinate X ₄ of the central region at the position of Z = h ₁ where the imaging plane is displaced from the surface of each ML is obtained from the following equation (3). Can do. However, i = 1 to 7 and a value in which the coordinate suffix is outside 1 to 7 is a value that is not used.

L (i) = (a _i−2 + b _i−1 + c _i + d _{i + 1} + e _{i + 2} ) (3)

In the case of the coordinate X ₄ , since i = 4, L (4) = (a ₂ + b ₃ + c ₄ + d ₅ + e ₆ ). As is clear from this, the central pixel signal c _i (c _{4 in} this example) in each region of the imaging surface of the imaging device 13 always exists regardless of the focal position. That is, an image obtained by collecting only these signals is a constant image regardless of the position on the object side, and is an image without defocusing.

  Therefore, as shown in FIG. 8, an image composed of signals having the relationship expressed by the following equation (4) has a constant image plane regardless of the image plane (in other words, the image plane is constant). It will always be an all-in-focus image (wherever the inspection is). However, FIG. 8 is a diagram in which a field lens (hereinafter referred to as FL) is added to FIGS. By disposing the FL, the angle of the light beam incident on the ML disposed at a high image height can be suppressed. Therefore, although not described in the above-described example, it is preferable to provide the FL.

L (i) = c _i (4)

That is, for example, in FIG. 8, an image obtained by collecting central pixel signals c ₁ , c ₂ , c ₃ , c ₄ , c ₅ , c ₆ , and c ₇ in each area of the imaging surface of the image sensor 13 is This results in a focused omnifocal image.

  As described above, by adopting the configuration of FIG. 1a or 1b, the procedure for generating a focus image is different, but a plurality of images having different focal planes are generated by reconstructing rays from different angles. .

[3. Measurement signal processing procedure]
Next, signal processing executed to obtain image data for measurement from the image group at the virtual focal position generated by the above procedure will be described. Such signal processing is characterized in that an omnifocal image that is always in focus is used, and by performing edge processing on each acquired image data while shifting the focal plane little by little in the optical axis direction, focusing is achieved. Select an image.

(3-1) Multi-Focus Sequential Processing Method Although several methods can be considered as the above-described edge processing, first, a case where a multi-focal sequential processing method is used as a method for detecting an edge of an image will be described. To do. Further, a case where a Fourier transform (wavelet transform) method is used as an edge process in the multifocal sequential process will be described as an example.

  Above [2. After a group of images on a plurality of different object planes is formed by the procedure for generating a focus image, the focus information is obtained by the following method at the same pixel area position of each image by the same process as SFF described above. Can be detected.

That is, in such signal processing, as shown in FIG. 9, FFT (Fast Fourier) is performed on image groups of a plurality of object planes (..., Z ₋₁ , Z ₀ , Z ₁ ,. The image signal is subjected to Fourier transform (or wavelet transform) by Transform, and only high-frequency components are extracted in the spatial frequency space (on the Fourier plane) with a high pass filter, and then inverse Fourier transform (or inverse wavelet transform). ) By this processing, an image of only the edge (contour) portion is obtained. Next, paying attention to the position (X _i , Y _i ) of each point having an edge in these images, a plurality of object planes (..., Z ₋₁ , Z ₀ , Z ₁ ,...) An image having the sharpest edge (the highest inclination component) is selected from the image group as the best focus.

  FIG. 10 is a graph of the slope component of the edge of the sampling point in the image group of each focal plane. In FIG. 10, the horizontal axis Z represents the position of the object plane (distance from the test object), and the vertical axis α represents the inclination component of the edge, and the value increases in the upward direction in the figure (the focal point). (The degree of focus is high).

  As shown in FIG. 10, the sampling points in the Z direction are discrete because they are values for each object plane. However, as shown by the curve in the figure, the function approximation is used. By obtaining the peak (dotted line in the figure), a position Z having a value after the decimal point is obtained by interpolation processing, and the accuracy is improved.

  However, although the position of Z can be obtained by this procedure, the following problems remain.

  That is, when an edge is detected from an image with a different focus, there are places where the focus (focus) is in focus and places where it is not in accordance with the distance of the object. At this time, if edges with extremely different brightness (signal intensity) are adjacent to each other, the bright edge that is not in focus covers the adjacent dark edge that is in focus, and the position of the edge is erroneously recognized. There is a fear.

  In order to solve such a problem, it is only necessary to obtain a differential value for each pixel using an image in which all edges are in focus, and to detect a correct edge position first.

Therefore, in the present embodiment, even when the image of the center (on the axis) of the objective lens 11 formed in the ML at the center of the MLA 12 (corresponding to ML _b in FIG. 2) changes the position of the object plane. It is an image that does not change (even if the direction of Z changes), has a deep depth of field, is almost completely in focus regardless of the distance on the object side, no image shift, and no blur Focusing on the fact that there is no image (omnifocal image), this image is used.

In the signal processing of the present invention, using this omnifocal image, the Fourier transform and high-pass filter processing are performed as described above to obtain the correct edge position (X _i , Y _i ), and at the same time the neighboring image The contrast signal (image density value) is registered. Next, with respect to those edges, the edge position (X _i , Y _i ) of _interest and its edge are selected from among a group of images of a large number of object planes (..., Z ₋₁ , Z ₀ , Z ₁ ,. The Z in the image in which the neighboring image signal has the highest correlation with the registered contrast signal is obtained.

That is, for each edge position (X _i , Y _i ) obtained from the omnifocal image, the edge position of the same coordinate (X _i , Y _i ) is obtained from a plurality of (including blurred) images having different object positions (Z). The object position (Z) having the highest similarity (correlation) with the edge signal of the unfocused omnifocal image is obtained. Thereby, Z at each edge point (X _i , Y _i ) can be detected.

Specifically, first, the target edge position (X _i , Y _i ) that is the target edge position and the signal intensity f _{i, j} of the neighboring pixels are detected. Next, with respect to those edges, the same edge position (X _i , Y _i ) from the image group of a large number of object planes (..., Z ₋₁ , Z ₀ , Z ₁ ,. Z is obtained when the neighborhood signal g _{i, j} (Z) has the highest correlation with f _{i, j} .

As a method of obtaining this correlation, it is preferable to use a well-known normalized correlation method that absorbs the difference in brightness and contrast at each point. Other methods may be used as long as the correlation can be obtained, such as a method of obtaining the sum of squares .

  Then, for example, Z is obtained when the value of the following formula (5) or formula (6) is the smallest.

Thus, the edge coordinates of the object (X _i, Y _i) is Z at position sought.

  Although an example in which image processing is directly performed on the image signal output from the image sensor 13 has been described here, image preprocessing may be performed for the purpose of highlighting edge signals and reducing detection errors. Good. For example, each image is processed by combining a Fourier transform and a high-pass filter, or by image processing such as a digital differential filter called a Sobel filter or a Laplacian filter, thereby obtaining a so-called edge image in which only the contour portion is emphasized. After that, the above correlation process may be performed.

  Further, when calculating the correlation, it is also possible to obtain the correlation using the interpolation processing as shown in FIG.

  As described above, in the multifocal sequential processing method, when an edge position is specified only by the intensity of an image signal (an image that is in focus in the optical axis direction) is to be specified, the original edge to be specified is caused by strong blur light. Since there are cases where it cannot be detected, in order to avoid this, correlation processing using an omnifocal image is performed.

(3-2) Omnifocal Edge Extraction Method Next, a case where the omnifocal edge extraction method is used as a method for detecting an edge of an image will be described.

In the multi-focal sequential processing method (3-1) described above, processing such as Fourier transform is performed after a large number of focused image groups are formed once, so processing time is inevitably required. Also, as an image for detecting an edge, in the image group of the previous object plane (..., Z ₋₁ , Z ₀ , Z ₁ ,...), There is a focal plane somewhere, and from that focal point. The deviated portion is blurred and it becomes difficult to detect all edges. Therefore, in the all-focus edge extraction method, focusing on the fact that the edge position can be specified on the two-dimensional image in advance by the all-focus image formed by the ML at the center of the MLA 12, the edge position in the optical axis direction is determined only by the specified edge position. Search for. That is, the feature (edge) in the omnifocal image is detected first, only the image of the object plane at the position corresponding to the detected edge position is generated, and Z that is the best focus at that position is determined. If detected, the number of processing points can be reduced, and not only the calculation burden is reduced, but also the calculation time can be shortened, contributing to rapid measurement.

Specifically, using an omnifocal image formed in the center ML of the MLA 12, a 3 × 3 matrix (generally called a differential operator) for contour extraction, which is well known in image processing, is used. The edge extraction calculation process is performed. For example, an edge in an image is detected in the X direction or the Y direction using a Sobel filter or the like. This method is particularly effective for line edge detection. Further, a Laplacian filter corresponding to a second derivative may be used for detection of points such as spots and intersections. These are due to a difference in some structure, color, reflectance, etc., such as a step of the test object. In the all-focus edge extraction method, the values processed by these filters at the detected edge and point positions (X _i , Y _i ) are calculated as defocus amounts.

In the omnifocal edge extraction method, paying attention to each point (position) having an edge of the image, as in (3-1), the above-described many object planes (..., Z ₋₁ , Z ₍₀ , Z ₁ ,...) Corresponding to this defocus amount change is calculated. Since the value obtained by this is the same data as in FIG. 10, the peak of the signal, that is, Z that is the best focus is obtained from this.

As described above, each edge coordinate (X _i , Y _i ) position of the object is obtained by edge processing such as the above-described multifocal sequential processing method (3-1) or the all-focus edge extraction method (3-2). Z at is detected. Thereafter, a distance measurement process using the detected value is performed. The process is described in [4. Distance measurement procedure].

[4. Distance measurement procedure]
(4-1) Method for obtaining object position from multiple focus images Distance measurement using Z at each edge coordinate (X _i , Y _i ) position of the object obtained by the above procedure is performed. There are several methods for the procedure, but here we first generate a number of focus images (a number of images with different focal plane positions, that is, a number of virtual focal plane images), and the best focus among them is generated. A method for obtaining the position will be described.

Here, the value of the position (..., Z ₋₁ , Z ₀ , Z ₁ ,...) Of the object plane on the object image side conjugate with the virtual focal plane is based on the expression (1). It is uniquely determined by the distance a between the lens 11 and the imaging surface of the imaging device 13 or the distance (a ₁ , a ₂ , a ₃ ,..., _An ) between the objective lens 11 and the virtual focal plane.

Therefore, if the position of the focal plane (including the case where it is on the imaging plane (that is, a ₁ , a ₂ , a ₃ ,..., _An )) is determined, the test object to be focused on that position. And the distance b to that part can be obtained using the equation (1). As a result, the coordinate (X, Y, Z) of the portion of the test object, ie, the coordinate (X, Y) information of the two-dimensional image output from the image sensor 13 together with the obtained distance b, that is, Three-dimensional spatial coordinates can be determined.

(4-2) Method for Obtaining Object Position from Image Information of Two Different Focal Planes In the above method (4-1), a large number of focus images (images obtained when a number of different focal planes are assumed) are obtained. The distance to each part of the test object can be accurately measured by generating and detecting the best focal position from each focus image. For that purpose, for example, a procedure of detecting peaks from multipoint information as shown in FIG. 10 is required. In FIG. 10, the horizontal axis Z indicates the distance to the test object, and the vertical axis α indicates the degree of focus (degree of focus). From this figure, it can be seen that the distance to the in-focus portion of the test object at a certain coordinate (X, Y) of the two-dimensional image can be obtained as the value of Z that is the maximum value of α. However, there are many processing procedures, and there is a high possibility that processing will take time.

  Therefore, here, as another method, a measurement procedure for speeding up with a small amount of processing will be described. In this method, a large number of images are generated and peaks are not detected from the multipoint information as shown in FIG. 10, but the difference signal between the two is obtained using image information of at least two different focal planes. That is, the position of Z is obtained from the difference in the defocus amount at the two positions. Since this Z corresponds to the distance b from the objective lens 11 to the test object, the coordinate (X, Y) information of the two-dimensional image of the image sensor 13 is combined with this value and the coordinate (X, Y) of the test object. X, Y, Z), that is, a three-dimensional shape can be measured.

  However, there is a high possibility that the difference in local reflection intensity of the test object is affected only by the value calculated by this method. That is, the signal intensity varies depending on the brightness distribution of the original illumination light and the difference in reflectance depending on the position of the test object. In particular, if there is a portion where the difference in brightness is significant in the adjacent region, both signals are mixed in the defocused state, so that there is a possibility that an edge peak is erroneously detected.

Therefore , the calculation value obtained by such calculation processing is divided by each pixel value of the image formed by the ML at the center of the MLA 12 and normalized so that brightness (light intensity) is obtained for each local area. Therefore, it is possible to obtain a standardized defocus blur amount.

  As described above, in the present embodiment, for example, the distance measurement process is performed by the method (4-1) or (4-2), so that the coordinates (X, Y, Z) is determined and the three-dimensional shape is measured.

  Thus, in the present embodiment, since the three-dimensional shape of the test object can be measured using the image signal obtained by one imaging, a plurality of imaging operations are required as in the conventional SFF. Compared to the case, measurement can be performed at high speed.

[5. Shape measuring device]
Next, the configuration of the shape measuring apparatus according to the present embodiment that performs the above-described distance measurement process and measures the three-dimensional shape of the test object will be described with reference to FIG.

  As shown in FIG. 11, the shape measuring apparatus includes an objective lens 11, an MLA 12, an image sensor 13, a control circuit 31, a user interface 32, a drive circuit 33, an arithmetic processing circuit 34, and a memory 35. In FIG. 11, the objective lens 11 to the imaging element 13 correspond to the objective lens 11 to the imaging element 13 of FIG.

  The control circuit 31 controls the drive circuit 33, the arithmetic processing circuit 34, and the memory 35 in accordance with instructions input from the user via the user interface 32. Further, the control circuit 31 controls the operation of each part of the shape measuring apparatus of FIG.

  The drive circuit 33 drives the image sensor 13. Further, the drive circuit 33 outputs the image signal from the image sensor 13 to the arithmetic processing circuit 34.

  The arithmetic processing circuit 34 performs predetermined arithmetic processing based on the image signal from the drive circuit 33. Then, the arithmetic processing circuit 34 stores the calculation result in the memory 35.

  The arithmetic processing circuit 34 measures the three-dimensional shape of the test object from image data obtained by imaging the image formed by each ML of the MLA 12 by the imaging device 13. Details of this processing will be described together with the description of the operation of the arithmetic processing circuit 34 of FIG.

  The shape measuring apparatus is configured as described above.

  Next, the three-dimensional shape measurement process performed by the arithmetic processing circuit 34 of FIG. 11 will be described with reference to the flowchart of FIG.

  The arithmetic processing circuit 34 acquires the image signal from the drive circuit 33 in step S11, and generates a focus image in step S12. Specifically, the above-mentioned [2. As described in [Procedure for generating a focus image], the arithmetic processing circuit 34 uses the method (2-1) in the case of the configuration of FIG. 1a and (2-2) in the case of the configuration of FIG. 1b. By this method, light rays from different angles are reconstructed to generate a plurality of images having different focal planes.

In step S13, the arithmetic processing circuit 34 performs measurement signal processing. Specifically, the above [3. As described in the “Signal processing procedure of measurement”, the arithmetic processing circuit 34 performs the edge processing such as the “multifocal sequential processing method” in (3-1) or the “all-focus edge extraction method” in (3-2). The Z at each edge coordinate (X _i , Y _i ) position of the object is obtained using an omnifocal image that is always in focus.

  Then, the arithmetic processing circuit 34 performs a distance measurement process in step S14, measures the three-dimensional shape of the object in step S15, and the process ends. Specifically, the above-mentioned [4. As described in the “distance measurement procedure”, the arithmetic processing circuit 34 selects the “image information of two different focal planes” in (4-1) “Method for obtaining an object position from a large number of focus images” or (4-2). The method of determining the object position from the “coordinates (X, Y, Z) of the surface of the test object is determined, and the three-dimensional shape is measured.

As described above, in the shape measuring apparatus in FIG. 11, the arithmetic processing circuit 34 [2. Procedure for generating focus image], [3. Measurement signal processing procedure], [4. The distance measurement procedure] is sequentially executed, thereby performing signal processing using an omnifocal image obtained from the image formed by the ML at the center of the MLA 12, and each edge coordinate (X _{i) of the} object obtained thereby. , Y _i ), the coordinates (X, Y, Z) of the surface of the test object, that is, the three-dimensional shape is measured from Z at the position.

That is, as shown in FIG. 1a, when the MLA 12 is arranged so as to be conjugate with the exit pupil of the objective lens 11, the arithmetic processing circuit 34 determines the position of the optical axis of the objective lens 11 among the plurality of MLs. The multifocal image (reference image) obtained from the image formed on the imaging surface of the imaging device 13 by the ML at the center of the MLA 12 and a plurality of images (measurement images) having different focal planes obtained from the output of the imaging device 13 The three-dimensional shape of the test object is measured using the Z position of the test object obtained from the correlation with ( ).

In the present embodiment, the configuration shown in FIG. 1a is mainly described. However, as shown in FIG. 1b, the imaging surface of the imaging device 13 is arranged so as to be conjugate with the exit pupil of the objective lens 11. In this case, the arithmetic processing circuit 34 includes an omnifocal image (reference image) obtained from the pixels constituting the imaging region of the imaging element 13 where the principal rays of the objective lens 11 intersect, and a focal plane obtained from the output of the imaging element 13. The three-dimensional shape of the test object is measured using the Z position of the test object obtained from the correlation with a plurality of different images (measurement images).

[6. Measurement accuracy maintenance procedure]
In the present embodiment, the function of the measurement sensor can be improved by moving the measurement sensor system or using a zoom lens. Hereinafter, this embodiment will be described.

  In FIG. 13, when the horizontal axis is a (lens position) and the vertical axis is b (object distance), the focal length f of the objective lens 11 is 25 mm, and the value of the lens position a is 27 mm to 45 mm. A graph of the value of the object distance b when changed is shown. In FIG. 13, the value of the lens position a increases as the horizontal axis “a” goes to the right in the figure, and the value of the object distance b increases as the vertical axis “b” goes upward in the figure. The lens position a and the object distance b correspond to a and b in the Gauss formula of the above formula (1).

  As is apparent from this graph, the value of the object distance b to the test object varies greatly depending on the value of the lens position a.

  That is, according to FIG. 13, when the value of the lens position a decreases, the object distance b rapidly increases (the distance increases), and the so-called depth of field increases, and the sensitivity of the focus information to changes in the object distance b is increased. It turns out that falls.

  In actual measurement, the working distance between the object to be examined and the objective lens 11 is not sufficient, and the object to be measured may be out of the measurement range due to the large object or may not be close to the object to be examined. . In such a case, measuring the shape of a large test object at once, or measuring with the same accuracy even if the working distance changes leads to an improvement in the function of the apparatus.

  Therefore, in the present embodiment, a zoom lens is used as the objective lens 11. Hereinafter, an example using a zoom lens will be described.

  The zoom lens enlarges a distant object or reduces a nearby object. At this time, since there is an optimum lens position a for the focal length f of the objective lens 11, it is preferable that the value of the lens position a can be changed in order to make the lens position a an optimum value.

  The shape measuring apparatus having such a configuration has a configuration as shown in FIG. 14, for example.

  In FIG. 14, portions corresponding to those in FIG. 11 are denoted by the same reference numerals, and description of portions having the same processing is omitted.

  The objective lens 11 is configured as a zoom lens with a variable focal length. As the configuration of the zoom lens, a known one can be used as a zoom lens for a single-lens reflex camera or the like. With this zoom lens, it is also possible to adjust the focus up to the test object. Therefore, the zoom lens includes a zoom lens ZL and a focusing lens FoL as shown in FIG.

  The zoom lens driving device 51 moves the zoom lens ZL among the plurality of lenses constituting the objective lens 11 in the optical axis direction (left and right direction in the drawing) of the objective lens 11 under the control of the control circuit 31. As a result, the focal length f of the objective lens 11 is changed. That is, the focal length f of the objective lens 11 is changed by the movement of the zoom lens ZL, so that the size of the image of the test object can be enlarged or reduced.

  Further, the focusing lens driving device 52 focuses the object to be examined by moving the focusing lens FoL among the plurality of lenses constituting the objective lens 11 under the control of the control circuit 31.

  As described above, the zoom lens ZL is used to change the focal length f of the objective lens 11 to change the magnification of the test object, and the focus lens FoL is used to focus on the test object, thereby allowing the test object to be focused. Even if the distance (working distance) changes, the measurement can be performed in an optimum state.

  As shown in FIG. 14, for example, if the control circuit 31 and the arithmetic processing circuit 34 are configured by the CPU 41 as shown in FIG. 14, the CPU 41 is used for zooming according to the flowchart shown in FIG. The lens driving device 51 and the focusing lens driving device 52 are controlled.

  In step S <b> 31, the CPU 41 controls the drive circuit 33 in accordance with a measurement start command supplied from the user interface 32 and captures an image signal from the image sensor 13.

  In step S32, the CPU 41 measures the distance from the captured image signal to the test object H. Here, first, the distance to the focusing point P1 of the test object H that coincides with the optical axis of the objective lens 11 is measured.

  In step S33, the CPU 41 outputs a control signal to the focusing lens driving device 52 so as to focus on the focusing point P1. Then, the focusing lens driving device 52 controls the focusing lens FoL of the objective lens 11 for focusing according to a control signal from the CPU 41. Specifically, in the case of the optical system of FIG. 1a, the imaging surface (front surface) of the imaging element 13 and the focusing point P1 are conjugate, and in the case of the optical system of FIG. The control circuit 31 controls the focusing lens driving device 52 so that the surface of each ML and the focusing point P1 are conjugate.

  In step S34, the CPU 41 enables measurement of the test object H in an optimum state. Specifically, when the test object H is in the imaging surface of the image sensor 13 and the value of the lens position a is in the optimum range, the test object H is already in an optimal state. Therefore, the CPU 41 does not perform any particular process, and the process proceeds to step S35.

  Note that the meaning of “when the value of the lens position a is in the optimum range” is as described with reference to FIG. 13, and thus the description thereof is omitted here. However, the curve in FIG. 13 depends on the focal length f of the objective lens 11 as is apparent from the equation (1). Therefore, the larger the value of the focal length f (the more telephoto), the smaller the change in the object distance b relative to the change in the lens position a. That is, the measurement sensitivity is improved particularly when the object distance b is increased.

  On the other hand, when the test object H is not in a state where it can be measured optimally, the CPU 41 determines whether or not the test object H protrudes from the imaging surface of the image sensor 13, and the test object H is removed from the imaging surface. If it is determined that the object is protruding, a control signal is output to the zoom lens driving device 51 so that the test object H enters the imaging surface, the zoom lens ZL of the objective lens 11 is moved, and the focus of the objective lens 11 is moved. The distance f is decreased (wide angle).

  When measuring with the focal length f of the objective lens 11 being small, the relationship between the lens position a and the object distance b is determined according to the set wide-angle focal length f. A control signal is output to the focusing lens driving device 52 to move the focusing lens FoL so that a change in the object distance b with respect to the change in the position a, that is, measurement accuracy can be obtained. In general, the wider the angle, the deeper the focal depth and the more difficult it is to detect the change in the object distance b. Therefore, there are cases where a predetermined measurement accuracy cannot be obtained even if the focusing lens FoL is moved. In that case, in order to change the working distance, it is necessary to measure the test object H away from the test object H or away from the test object H. If this is not possible, it is necessary to divide and measure the test object H by setting the optimum focal length f and lens position a. A technique for joining divided images is already known as an image processing technique, and the CPU 41 can measure and measure the test object H, and then connect the divided and measured images using this technique.

  In addition, when the CPU 41 determines that the test object H does not protrude from the imaging surface of the image sensor 13 and is within the imaging surface, the test object H within the imaging surface further moves to the imaging surface. It is determined whether it is small. That is, when the test object H is small with respect to the imaging surface, the measurement accuracy deteriorates. Therefore, when the CPU 41 determines that the test object H in the imaging surface is small with respect to the imaging surface, A control signal is output to the zoom lens driving device 51 to move the zoom lens ZL so that the focal length f of the objective lens 11 is increased (telephoto) so that the test object H does not deviate from the imaging surface. If the relationship between the lens position a and the object distance b deviates from the optimum range by changing the focal length f of the objective lens 11, the CPU 41 outputs a control signal to the focusing lens driving device 52. Then, the lens position a is set so that it falls within the optimum range.

  In step S35, the CPU 41 obtains the distance (Z) to the measurement point of the test object H that can be measured optimally, and in step S36, the position (X, Y) of the measurement point on the imaging surface. Then, the spatial coordinates (X, Y, Z) of each measurement point of the test object H in a state where it can be measured optimally are measured, and the three-dimensional shape of the test object H is obtained. At this time, it is natural to correct each measurement point in consideration of the focal length f of the objective lens 11 and the distance to each measurement point.

  As described above, the CPU 41 controls the zoom lens driving device 51, the focusing lens driving device 52, and the like. Thereby, the shape of the test object H can be measured in an optimum state without moving the test object H along the optical axis of the objective lens 11.

  By the way, the above description is general. Usually, the test object H is known in advance by visual observation or the like. Therefore, the test object H and the apparatus are arranged so that the working distance is substantially appropriate. Measurement is started after the positional relationship is determined. Therefore, in reality, there is little change in the focal length f of the objective lens 11 or the lens position a.

  In the above description, the change in the focal length f and the change in the lens position a are controlled by the CPU 41. However, the image corresponding to the image signal output from the image sensor 13 is displayed on the monitor 36, and the user The focal length f and lens position a may be adjusted while viewing the image displayed on the monitor 36. Of course, in this case, the various measured values described above are also displayed on the monitor 36.

  In FIG. 14, the CPU 41 has been described as including the control circuit 31 and the arithmetic processing circuit 34, but a part of the configuration illustrated in FIG. 14 includes the CPU 41, the user interface 32, the memory 35, and the monitor 36. It can also be seen as a personal computer. In this case, the personal computer controls the zoom lens driving device 51 and the focusing lens driving device 52 in accordance with an instruction from the operator who performs the operation while viewing the image displayed on the monitor 36.

  The series of processes described above can be executed by hardware, or can be executed by software. When a series of processing is executed by software, a program constituting the software may execute various functions by installing a computer incorporated in dedicated hardware or various programs. For example, it is installed from a recording medium in a general-purpose personal computer or the like.

  This recording medium is not only composed of a computer, but also a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory or the like on which the program is recorded, which is distributed to provide a program to users. It is configured by a hard disk drive or a ROM (Read Only Memory) in which a program is recorded, which is provided to the user in a state of being preinstalled in the computer.

  Furthermore, the program for executing the above-described series of processes is installed in a computer via a wired or wireless communication medium such as a local area network, the Internet, or digital satellite broadcasting via an interface such as a router or a modem as necessary. You may be made to do.

  In the present specification, the step of describing the program stored in the recording medium is not limited to the processing performed in chronological order according to the described order, but is not necessarily performed in chronological order. It also includes processes that are executed individually.

  The embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.

It is a figure explaining the outline | summary of the optical system to which this invention is applied. It is a figure explaining the further detailed structure of an optical system. It is a figure which shows that the image seen from a certain ML is the image seen for every direction comprised with the light ray from the same direction. It is a figure which shows the tracking result of the light ray by an objective lens and MLA. It is a figure which shows the pixel arrangement | sequence in the image pick-up element surface imaged by MLA. It is a figure which shows the example of a structure when MLA is made into the image plane of an objective lens (Z = 0). The MLA is a diagram illustrating an example of a configuration when the image plane of the objective lens (Z = h _1). It is a figure which shows the example of an omnifocal image. It is a figure which shows the flow of the signal processing using the Fourier-transform method of an image group. It is the figure which graphed the inclination component of the edge of a sampling point. It is a figure which shows the structure of one Embodiment of the shape measuring apparatus to which this invention is applied. It is a flowchart explaining a three-dimensional shape measurement process. It is a figure which shows the relationship between a lens position and an object position. It is a figure which shows the further another structure of one Embodiment of the shape measuring apparatus to which this invention is applied. It is a flowchart explaining the three-dimensional shape measurement process which measures the to-be-tested object H in the optimal state.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 11 Objective lens, 12 MLA, 13 Image sensor, 31 Control circuit, 32 User interface, 33 Drive circuit, 34 Arithmetic processing circuit, 35 Memory, 36 Monitor, 41 CPU, 51 Zoom lens drive device, 52 Focus lens drive device , ML micro lens, FL field lens, FoL focusing lens, ZL zoom lens

Claims (4)

An objective lens;
An optical element having a plurality of lenses arranged two-dimensionally behind the objective lens;
A two-dimensional image sensor disposed behind the optical element;
A focal length changing means for changing a focal length of the objective lens;
Image forming means for generating a plurality of measurement images having different focal planes from the output of the image sensor;
A reference image in which all the pixels are focused, obtained from predetermined pixels corresponding to the optical arrangement of the objective lens and the optical element among the pixels constituting the imaging region of the imaging element corresponding to each of the plurality of lenses. comprising the, using the position of the optical axis of the objective lens of the object obtained from the correlation between the plurality of measurement images, and a three-dimensional shape measurement means for measuring the three-dimensional shape of the test object A shape measuring apparatus characterized by that. The shape measuring apparatus according to claim 1, further comprising a focus position varying unit that changes a focus position of the objective lens. The three-dimensional shape measuring means obtains distance information to the test object at the approximate center of the visual field of the objective lens,
The shape measuring apparatus according to claim 2, wherein the focus position varying unit controls the focus position based on the distance information measured by the three-dimensional shape measuring unit. The three-dimensional shape measurement means, and the distance to the test object, taking into account the magnification of the objective lens, the claim 3, characterized in that to measure the three-dimensional shape of the test object Shape measuring device.

Images