JP2008145121A - Three-dimensional shape measuring apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a three-dimensional shape measuring apparatus realizing both high measurement accuracy and a wide measurement range. <P>SOLUTION: A narrow band filter is inserted into an optical system to project a two-dimensional grid pattern 3 onto a device under test 2. Then the grid pattern on the device 2 is acquired as an image through the use of a plurality of light reception lenses 4 and imaging elements 5. The gravity position of longitudinal and horizontal lines of the two-dimensional grid pattern 3 is detected by image processing, and positions of respective grid points on the image are obtained from the gravity position. In doing so, a focal position of a projective lens 6 and a focal position of the light reception lens 4 are determined so that the focus of an imaging optical system aligns with an image focus location of the pattern 3. By taking the same procedures while successively switching over filters, data in a range where focal depths of focal positions of the projective lens and the light reception lens overlap can be acquired since they have different focal positions depending on the wavelength band of the projective lens and the light reception lens. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a three-dimensional shape measuring apparatus that measures the three-dimensional shape of an object using triangulation based on the active stereo method.

  In recent years, there is an increasing need for non-contact measurement of the three-dimensional shape and position of a test object in various fields. There are two types of methods for measuring the three-dimensional shape of an object in a non-contact manner.

  One is a passive method that performs measurement without irradiating a specific electromagnetic wave to assist measurement, and the other is a measurement by irradiating a specific electromagnetic wave to an object for measurement and obtaining information from it. Is an active way to do. A passive method includes a lens focus method, a stereo method, and the like, which are generally versatile and have an advantage that there are few restrictions on the shape and size of an object. On the other hand, active methods include an optical radar method, an active stereo method, a moire topography method, an illuminance difference stereo method, an interference method, and the like, and generally have an advantage of higher accuracy than a passive method.

  For measurements that require high accuracy, the active stereo method is usually used. As a representative method in the active stereo method, there is a slit projection method. This is a method of obtaining three-dimensional information of an object using triangulation by projecting slit light from a projector and detecting the position of a slit image reflected on the surface of a test object on a camera image. By scanning the slit light, the three-dimensional shape of the entire object can be obtained.

In this method, the spatial resolution depends on the scanning step, and it is necessary to make the step finer in order to obtain a high resolution. As a method for improving this, there are a gray code pattern projection method for projecting a specific pattern light and a color pattern light projection method described in Japanese Patent Laid-Open No. 2001-330417 (Patent Document 1). In these methods, a specific pattern is projected to finely divide the test object, and triangulation is performed for each divided part. By devising the pattern to be projected, it is possible to increase the number of divisions of the test object with the number of projections smaller than the number of scans of the slit projection method.
JP 2001-330417 A

  High spatial resolution can be achieved by the above-described method of making the steps fine, the gray code pattern projection method, the color pattern light projection method, and the like, but the measurement accuracy per point cannot be improved. In order to improve measurement accuracy, it is desirable to make the numerical aperture on the specimen side of the condenser lens used in the measurement optical system as large as possible. However, if the numerical aperture is increased, the depth of focus on the specimen side will become shallower. As a result, there is a problem of narrowing the measurement range.

  This invention is made | formed in view of such a situation, and makes it a subject to provide the three-dimensional shape measuring apparatus which makes high measurement accuracy and a wide measurement range compatible.

  The first means for solving the problem includes a projection device that projects a two-dimensional lattice pattern onto a test object, and a plurality of light receiving means that receive the two-dimensional lattice pattern projected onto the test object. A three-dimensional shape measuring apparatus, comprising: a projection optical system that constitutes the projection apparatus; and a light receiving optical system that constitutes the light receiving means is a multifocal lens.

  A second means for solving the above-mentioned problem is the first means, wherein the multifocal lens forms a multifocal point using axial chromatic aberration.

  A third means for solving the above-mentioned problem is the first means, wherein the multifocal lens forms a multifocal point using polarized light.

  A fourth means for solving the problem is any one of the first to third means, and the projected two-dimensional lattice pattern can identify the number of rows and columns of lines. In addition, the line has a specific pattern.

  According to the present invention, it is possible to provide a three-dimensional shape measuring apparatus that achieves both high measurement accuracy and a wide measurement range.

  Hereinafter, an example of an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a diagram schematically showing an example of an embodiment of the present invention. The two-dimensional lattice pattern 3 is projected from the pattern projection device 1 onto the test object 2, and the image is captured on each image sensor 5 by a plurality of light receiving optical systems 4. By detecting the position of each lattice point of the lattice pattern on the image sensor 5, angles α and β to the lattice are obtained with respect to the base line length L between the light receiving lenses 4, and each lattice point is obtained using the principle of triangulation. Can be determined. Since there are a plurality of lattice points on the test object 2, a plurality of positions on the test object 2 can be determined. Further, the position of the entire surface of the test object 2 can be determined by scanning the lattice by one pitch in the x and y directions.

  The present invention is characterized by the pattern projector 1 and the light receiving optical system 4. Details of the optical system of the pattern projection apparatus 1 are shown in FIG. The pattern projection apparatus 1 includes an illumination optical system 9, a filter turret 8, a two-dimensional grating mask 7, and a projection optical system 6.

  The illumination optical system 9 includes a light source (not shown), a collector lens, and the like, and uniformly illuminates the two-dimensional grating mask 7. The two-dimensional lattice mask 7 has a dark field pattern so that light rays can be transmitted in a lattice shape. The projection optical system 6 and the light receiving optical system 4 are designed so as to deviate on-axis chromatic aberration, and the condensing positions are sequentially shifted depending on the wavelength band. A filter turret 8 is disposed between the illumination optical system 9 and the two-dimensional grating mask 7, and a plurality of narrow band filters 10 are fitted therein. As shown in FIG. 3, consideration is given to the converging positions of the projection optical system 6 and the light receiving optical system 4 being different in the wavelength range selected by each filter. By rotating the filter turret 8, an arbitrary narrow band filter 10 can be inserted into the optical system, and the focal position can be changed accordingly.

  Hereinafter, the measurement method will be described. First, a narrow band filter 10 is inserted into the optical system to project the two-dimensional lattice pattern 3 onto the test object 2. Next, the lattice pattern on the test object 2 is captured as an image using the plurality of light receiving optical systems 4 and the image sensor 5. The center of gravity position of the vertical and horizontal lines of the two-dimensional grid pattern 3 is detected by image processing, and the position of each grid point on the image is obtained therefrom. At this time, the focal position of the projection optical system 6 and the focal position of the light receiving optical system 4 are aligned on the test object 2.

  At this time, since the test object is generally uneven, the two-dimensional grid pattern 3 on the test object 2 is clear at the focal point of the projection lens 6 (the imaging position of the two-dimensional grid pattern 7). , As it goes out of focus. Also, in the light receiving optical system for picking up an image, the lattice pattern at the position where the focal point is in focus clearly forms an image on the image pickup device, but the lattice pattern out of the focus position forms a blurred image on the image pickup device.

  That is, the lattice pattern out of focus by each of the projection optical system and the light receiving optical system is blurred. If the position of the center of gravity is obtained from a blurred lattice pattern, the measurement error increases. Therefore, it is preferable not to use such a lattice pattern for measurement. Therefore, image blur determination is performed at each of the lattice points obtained above, and only those whose blur amount is smaller than the specified reference value are stored as the positions of the lattice points. Thereby, a measurement error can be reduced.

  However, in this state, the positions of all lattice points in the test object cannot be obtained. Therefore, if the filter is sequentially switched and the same operation as described above is performed, the focal position differs depending on the wavelength band of the projection optical system and the light receiving optical system, and therefore it is possible to acquire data in a range in which those focal depths are superimposed. Become.

  This overall depth of focus can be made much deeper than the depth of focus determined by the numerical aperture on the object side of the projection optical system and the light receiving optical system, and the measurement range can be greatly expanded. By adjusting the actual focal depth considering the wavelength band to the assumed specimen size, and switching the wavelength and memorizing the positions of the focused lattice points in sequence, it covers almost the entire specimen surface. That is, the position of the lattice point is obtained.

  Since the position of each lattice point on the image is obtained in the plurality of light receiving optical systems, if the corresponding lattice point is determined, the angles α and β from the baseline length in FIG. 1 can be obtained, and the test object is obtained using triangulation. The coordinates of the grid points inside can be determined.

  Hereinafter, a method for determining a corresponding lattice point will be described. By providing patterns on the vertical and horizontal lines of the two-dimensional grid pattern, the grid points can be easily determined. For example, a pattern as shown in FIG. 4 can be considered. In this example, every 10 vertical and horizontal lines have different patterns. 1-9 are straight lines, 10-19 are short one-dot chain lines, 20-29 are short two-dot chain lines, short chain lines are provided for every 10 lines, and long chain lines are provided for every 100 lines. is doing. As a result, it is possible to identify the number of rows and columns of the target grid points on the image. Therefore, when the position of the lattice point on the image described above is recorded, information such as the number of rows and the number of columns is also recorded.

  Note that the line identification pattern is not limited to that described here, and for example, the line width can be changed or the density can be changed. This makes it possible to associate the positions of all grid points on a plurality of image sensors, and to obtain coordinates at a plurality of points on the test object using triangulation. Further, the coordinates of the entire surface of the test object can be obtained by scanning the lattice position by one pitch in the vertical and horizontal directions.

  Specific examples will be described below. As an image pickup device, a 4 million pixel CCD having an element size of 10 mm × 10 mm and a pixel number of 2000 pixels × 2000 pixels is used. The size of one pixel is 5 μm. The projection optical system and the light receiving optical system are common, and the magnification is 10 times and the numerical aperture NA on the object side is 0.007. When this optical system is used, it is possible to image a range of 100 mm × 100 mm as a test object at a time.

  The projection optical system and the light receiving optical system are designed to increase axial chromatic aberration, and it is assumed that the axial color difference on the object side is 100 mm. A narrow band filter is prepared in the vicinity of each wavelength that divides the on-axis color difference of 100 mm into 10 equal parts. In other words, by switching the filter, the focal position on the object side can be changed by about 10 mm by 10 equal parts. The line width of the two-dimensional lattice mask is 5 μm and the pitch is 20 μm. The diameter of the Airy disk on the object side of the projection optical system and the light receiving optical system is 100 μm at a wavelength of 588 nm, and the depth of focus is ± 6 mm.

  First, consider the case where a certain wavelength is selected for measurement. In this case, the line width of one of the two-dimensional lattice patterns projected on the object plane is multiplied by 10 times, and blurring due to diffraction is added to the line width to be about 150 μm. This corresponds to 3 pixels when converted on the image sensor. Since the pitch of the grid is 20 μm, it corresponds to 4 pixels on the image sensor, and each line of the grid is arranged with an interval of 1 pixel. An accuracy of about 1/20 pixel can be achieved by using a technique such as sub-pixel processing to obtain the position of the center of gravity of the line width having a spread of 3 pixels by image processing. This corresponds to 3 μm in terms of the specimen.

  However, since the depth of focus is only ± 6 mm in this state, the focused lattice point is determined only within the range of about 12 mm in the depth direction of the test object. Therefore, by sequentially switching the filters and performing the measurement, the focal position can be changed by 10 mm, and the previous focal depths can be connected in order. As a result, the depth of focus can be increased up to a maximum of 100 mm, and an object size of about 100 × 100 × 100 mm can be measured with an accuracy of about 3 μm.

  In the above description, the wavelength is switched on the pattern projection side. However, it is obvious that the same effect can be obtained even if the switching is performed on the imaging side. Note that if the wavelength is switched on the pattern imaging side, only the light of that wavelength naturally comes to the imaging side, so that wavelength switching on the imaging side is unnecessary.

  In the above embodiment, an example in which the depth of focus is increased by using chromatic aberration for the projection lens and the light receiving lens is described, but the same effect can be expected by using a bifocal lens using polarized light. At this time, two sets of polarizing plates whose polarization directions are orthogonal to each other correspond to filter switching.

It is a figure which shows the example of embodiment of this invention typically. It is a figure which shows the optical system of a pattern projector in detail. It is a figure which shows the focus position for every wavelength of a projection optical system and a light-receiving optical system. It is a figure which shows the example of a two-dimensional lattice pattern.

Explanation of symbols

1: pattern projection device, 2: test object, 3: two-dimensional grating pattern, 4: light receiving optical system, 5: image sensor, 6: projection optical system, 7: two-dimensional grating mask, 8: filter turret, 9: Illumination optical system, 10: Narrow band filter

Claims (4)

  A projection apparatus for projecting a two-dimensional lattice pattern onto a test object, and a three-dimensional shape measuring apparatus comprising a plurality of light receiving means for receiving the two-dimensional grid pattern projected on the test object, the projection 3. A three-dimensional shape measuring apparatus, wherein the projection optical system constituting the apparatus and the light receiving optical system constituting the light receiving means are multifocal lenses.   The three-dimensional shape measuring apparatus according to claim 1, wherein the multifocal lens forms a multifocal point using axial chromatic aberration.   The three-dimensional shape measuring apparatus according to claim 1, wherein the multifocal lens forms a multifocal point using polarized light.   4. The three-dimensional shape measuring apparatus according to claim 1, wherein the projected two-dimensional lattice pattern is specified to a line so that the number of lines and the number of columns can be identified. A three-dimensional shape measuring apparatus characterized by having a pattern.

Images