JP2004340728A - Measuring method and device using stereo optical system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a processing time and extend a measurable range in height in comparison with a conventional automatic focus method when measuring a height of an object. <P>SOLUTION: An object to be measured is imaged from the front and slanting directions (S5). The slanting image data of the object is converted such that a pattern in the image data becomes similar to a pattern in the front image data of the object (S6). Then, difference between a specific point on the front image data and a specific point on the converted slanting image data is detected (S7). The detected difference in position has a value corresponding to the object height. Based on the difference in position, the height of the object is detected (S8). Therefore, in comparison with a conventional automatic focus method which requires the fluctuation of distance between an optical system and an object, this system can reduce a processing time and extend a measurable range in height by deepening a depth of field of the optical system. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a measuring method and a measuring apparatus, and more particularly to a measuring method and a measuring apparatus using a stereo optical system, and a semiconductor manufacturing apparatus and a transport line using these methods and apparatuses.
[0002]
[Prior art]
Conventionally, autofocus has been used not only in cameras but also in various devices such as semiconductor manufacturing devices (for example, see Patent Document 1).
[0003]
Autofocus is a method in which an object is imaged at different distances while gradually changing the distance between the object and the optical system, and a distance at which the contrast of image data is highest is obtained. Such autofocus has been used not only for focusing but also for various purposes such as when keeping the height of various objects such as wafers constant or when measuring the thickness.
[0004]
[Patent Document 1]
JP 2003-84189 A
[0005]
[Problems to be solved by the invention]
However, when the autofocus is used, there are the following problems.
(1) In autofocus, it takes a long time to image the object a plurality of times while gradually changing the distance between the object and the optical system. In the example shown in FIG. 49, when searching for the distance at which the contrast is the highest, the object is imaged at each distance indicated by the cross, and the object is imaged a total of seven times. Therefore, it takes seven imaging times.
(2) In autofocus, an optical system having a shallow depth of field is used due to its principle, so that the measurable range is generally narrowed by the shallow depth of field. Specifically, when measuring the height of the object, the measurable range of the height is narrowed.
(3) In the autofocus, when the surface of the object has irregularities, the measurement is generally not performed correctly. For example, due to the unevenness of the surface, a curve having only one peak as shown in FIG. 49 cannot be obtained, and the curve becomes a curve having two or more peaks, and the measurement is not performed correctly.
[0006]
Further, there is a need for a technique for measuring the slack of a sheet (object) having a slack as shown in FIG. However, the sheet shown in FIG. 25 has a slack of 1 mm or more, and it is difficult to measure the slack by an autofocus technique using an optical system having a shallow depth of field (for example, about 50 μm).
[0007]
In an XYZθ mechanism (a mechanism that enables movement in the Z-axis direction and rotation direction in addition to movement in the X-axis direction and Y-axis direction), adjustment is generally performed by a laser length measuring device or an image processing device. There is such a problem. When a laser length measuring device is used, generally, the laser length measuring device must be arranged three-dimensionally, and adjustment is performed in units of one axis, which is complicated and requires a long working time. When an image processing apparatus is used, the straightness and the orthogonality of the X-axis and the Y-axis can be adjusted in a short time based on image data obtained by using a calibration sheet from above. On the other hand, as shown in FIG. 50, the inclination angle φ of the Z axis (or θ axis) appears as a difference in height, and therefore cannot be easily adjusted. Therefore, there is a need for an adjustment method that can easily and quickly adjust the XYZθ mechanism.
[0008]
The present invention has been made in view of such circumstances, and a measuring method and a measuring device for reducing a processing time and expanding a measurable range and these methods, as compared with a case where a conventional autofocus is used. An object of the present invention is to provide a semiconductor manufacturing apparatus and a transport line using the apparatus and the apparatus.
[0009]
Another object of the present invention is to enable measurement of various objects such as irregularities on the surface, various measurements such as measurement of sheet slackness, and easy adjustment of complicated mechanisms such as the XYZθ mechanism. It is to be.
[0010]
[Means for Solving the Problems]
The present invention relates to a measuring method for measuring the height of an object by using an oblique imaging unit that images an object located at an arbitrary height in an oblique direction and a facing imaging unit that images an object from a direction facing the object. An apparatus, a semiconductor manufacturing apparatus and a transport line using these methods and apparatuses.
[0011]
Here, the height of the target object may be obtained as an absolute distance between the imaging optical system of the directly-facing imaging means and the target object, but in general, a specific height serving as a reference is used as a reference. Is calculated as a relative distance.
[0012]
First, an image conversion parameter indicating a relationship between pixels forming the oblique image data and pixels forming image data obtained by converting the oblique image data (hereinafter referred to as “converted image data”), An image conversion parameter for converting the oblique image data so that the shape of the arbitrary pattern and the shape of the pattern in the converted image data are similar to each other is obtained. Next, directly-facing image data obtained by imaging the object located at an arbitrary height by the directly-facing imaging means and oblique image data obtained by imaging by the oblique imaging means are acquired. Next, the oblique image data of the target object is converted into converted image data based on the image conversion parameters. Next, a difference between a position on the directly-facing image data and a position on the converted image data is detected for a specific point on the object. The difference between the detected positions has a value corresponding to the height of the target object. Next, the height of the target object is detected based on the difference between the positions.
[0013]
In carrying out the present invention, it is preferable to obtain the image conversion parameters by the following procedure.
(A) A test pattern having a plurality of characteristic points arranged at a specific height serving as a reference is imaged by a directly-facing imaging unit and an oblique imaging unit.
(B) Obtain the coordinates of a plurality of pixels in the diagonal image data respectively corresponding to a plurality of characteristic points in the test pattern (hereinafter referred to as “diagonal plane coordinates”), and respectively correspond to the characteristic points. The coordinates of a plurality of pixels in the directly facing image data (hereinafter referred to as “facing plane coordinates”) are acquired.
(C) Based on the oblique plane coordinates and the facing plane coordinates of a plurality of characteristic points in the test pattern, the relationship between the coordinates of the pixels constituting the oblique image data and the coordinates of the pixels constituting the facing image data is calculated. The coordinate conversion parameter shown is calculated.
(D) Calculate image conversion parameters based on coordinate conversion parameters at a specific height serving as a reference.
[0014]
The image conversion parameter once calculated in this manner may be stored in the storage device, and may be obtained from the storage device next time.
[0015]
In implementing the present invention, in image data conversion, it is preferable to obtain the light intensity Img2 (I, J) of the pixel [I, J] constituting the converted image data according to (Equation 1).
Img2 (I, J) = Σ [W (k, m) × Img1 (k, m)] / ΣW (k, m) (Equation 1)
Here, assuming that I, J, k and m are integers, Img1 (k, m) is the light intensity of the pixel [k, m] at the k-th row and the m-th column of the oblique image data. W (k, m) is an image conversion parameter, which is a pixel at the I-th row and the J-th column on the light receiving plane of the directly-facing imaging means (hereinafter referred to as “facing light receiving plane”) at a specific reference height. When [I, J] is projected on the light receiving plane of the oblique imaging means (hereinafter, referred to as “oblique light receiving plane”), it represents the area projected on the pixel [k, m] of the oblique light receiving plane. Further, the range for calculating the sum of the right side of the equation (1) is such that the pixel [I, J] on the directly-facing light receiving plane projected on the oblique light receiving plane is positioned above the pixel [k, m] on the oblique light receiving plane. If any part is projected, it is assumed that it covers all the projected pixels [k, m].
[0016]
In practicing the present invention, it is preferable that the detection of the height of the object is performed based on the height conversion information acquired by the following procedure.
(E) A test pattern having a plurality of characteristic points is moved stepwise in the height direction.
(F) At each height, the test pattern is imaged by the directly-facing imaging means and the oblique imaging means.
(G) Oblique image data of the test pattern imaged at each height is converted based on the coordinate conversion parameters and the image conversion parameters.
(H) For a specific point on the test pattern, a difference between the position of the test pattern on the directly facing image data and the position of the test pattern on the converted image data is detected.
(I) Generate height conversion information indicating the relationship between each height of the test pattern and the difference between each position.
[0017]
The height conversion information once generated as described above may be stored in the storage device, and may be acquired from the storage device next time.
[0018]
In practicing the present invention, it is preferable that the illuminating means is provided on an axis that is line-symmetric with the optical axis of the oblique imaging means, with the optical axis of the directly facing imaging means as the axis of symmetry.
[0019]
Further, it is preferable to include a height control means for controlling the height of the object so that the height of the object measured by the measuring device of the present invention is constant.
[0020]
Further, in a semiconductor manufacturing apparatus provided with a cassette for stocking a plurality of semiconductor substrates and transport means for removing and transporting the semiconductor substrates one by one from the cassette, it is preferable to include the measuring device or the height control device of the present invention. is there. Here, the object is a semiconductor substrate taken out of the cassette. A typical semiconductor substrate is a wafer.
[0021]
Further, it is preferable that the measuring device or the height control device of the present invention be provided in a transfer line including a transfer unit that connects a plurality of processes and one or more inspection units based on images. Here, the target object is an object conveyed on the conveyance line.
[0022]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the drawings merely schematically show the shapes, sizes, and arrangements of the components to the extent that the present invention can be understood, and the present invention is not limited to the illustrated examples.
[0023]
With reference to FIG. 1, a case where an object is imaged from obliquely above and a case where an object is imaged from a direction facing directly, for example, directly above, will be described.
[0024]
In FIG. 1, a position A (hereinafter, referred to as an “oblique imaging position”) at which the measurement target 1 is imaged in an oblique direction and a position B (hereinafter, “positive”) at which the measurement target 1 is imaged from a directly facing direction (eg, directly above). The imaging devices 10 and 12 are arranged at “the position relative to the imaging position”, respectively. These imaging devices 10 and 12 include imaging optical systems 10a and 12a and light receiving planes 10b and 12b, respectively. Hereinafter, the imaging device 10 disposed at the oblique imaging position A is referred to as an “oblique imaging device”, its imaging optical system 10a is referred to as an “oblique optical system”, and its light receiving plane 10b is referred to as an “oblique light receiving plane”. Further, the imaging device 12 disposed at the directly-facing imaging position B is referred to as a “facing imaging device”, the imaging optical system 12a is referred to as a “facing optical system”, and the light receiving plane 12b is referred to as a “facing light receiving plane”. .
[0025]
In FIG. 1, a grid pattern exists on the surface to be measured of the measuring object 1. The intersections of the crosses on the grid pattern are hereinafter referred to as “grid points”. In the lattice pattern, a cross-shaped portion including a lattice point is hereinafter referred to as a “cross mark”. For convenience of explanation, a case where the measurement target 1 has a grid pattern on the surface to be measured will be described. However, the technical content described below is based on whether the characteristic pattern on the measurement target is a grid pattern. It holds regardless of any.
[0026]
The facing optical system 12a is arranged so that its optical axis 12c is orthogonal to the surface to be measured of the measuring object 1. The oblique optical system 10a is disposed such that its optical axis 10c is oblique to the surface to be measured of the measurement object 1. The height measurement of the measuring object 1 is performed by measuring image data (hereinafter referred to as “facing image data”) obtained on the directly-facing light receiving plane 12b via the directly-facing optical system 12a and obliquely receiving light via the oblique optical system 10a. This is performed using image data obtained on the plane 10b (hereinafter, referred to as “oblique image data”).
[0027]
Here, the change of the image data when the height of the measurement object 1 is changed will be described.
[0028]
FIG. 2 is an enlarged cross-sectional view of a portion 1a of the measurement object to be imaged. In FIG. 2, the magnification of both the facing optical system and the oblique optical system is set to 1 for convenience of explanation. FIG. 3A shows the directly-facing image data of the measuring object 1 located at the reference height (height = 0), and FIG. 3B shows the oblique image data. Note that FIGS. 3A and 3B show only cross marks in the lattice pattern. In the directly-facing image data 3a in FIG. 3A, grid points on the surface to be measured of the measurement target 1 are photographed in a square shape as they are. That is, a square is formed by four adjacent grid points. On the other hand, in the oblique image data 3b in FIG. 3B, grid points on the surface to be measured of the measurement target are captured in a vertically long rectangular shape. That is, a vertically long rectangle is formed by four adjacent grid points. More specifically, if an angle between the optical axis 10c of the oblique optical system and the optical axis 12c of the directly-facing optical system (hereinafter referred to as “inclination angle”) is φ, the lattice point of the oblique image data 3b in the X-axis direction is The interval is obtained by multiplying the interval between lattice points in the X-axis direction of the directly-facing image data 3a by cosφ. On the other hand, the interval between the lattice points in the Y-axis direction of the oblique image data 3b is almost the same as the interval between the lattice points in the Y-axis direction of the directly-facing image data 3a.
[0029]
Here, the oblique image data 3b is converted so that the shape of the lattice pattern matches the shape of the lattice pattern in the directly-facing image data 3a. The converted image data (hereinafter, referred to as “converted image data”) is shown in FIG. Here, the conversion may be performed so that the interval between the lattice points of the converted image data 3c is the same as the interval between the lattice points of the directly-facing image data 3a. Further, the positions of the lattice points in the converted image data 3c substantially coincide with the positions of the lattice points in the directly-facing image data 3a. The converted image data 3c has a resolution 1 / cos φ lower than that of the directly-facing image data 3a, but hardly affects the measurement.
[0030]
Next, the directly-facing image data of the measurement object located at a height lower than the reference height is shown in FIG. 4A, and the oblique image data is shown in FIG. 4B. Here, the change in the shooting distance (for example, 0.5 mm) is within the range of the depth of field (for example, 1.5 mm). Further, since the change in the shooting distance (for example, 0.5 mm) is sufficiently smaller than the shooting distance (for example, 100 mm), the change in the scale of the image data due to the change in the shooting distance is ignored. Therefore, the directly-facing image data 4a in FIG. 4A is substantially the same as the directly-facing image data 3a at the reference height shown in FIG. On the other hand, the oblique image data 4b in FIG. 4B is different from the oblique image data 3b at the reference height shown in FIG. Become.
[0031]
Here, the oblique image data 4b is converted so that the shape of the lattice pattern matches the shape of the lattice pattern in the directly-facing image data 4a. FIG. 4C shows the converted image data (converted image data) 4c. Explaining with reference to FIG. 2, the difference d between the positions of the lattice points is d = h × tan φ. Here, h is the height difference, and φ is the angle between the optical axis 12c of the directly facing optical system and the optical axis 10c of the oblique optical system, that is, the inclination angle of the oblique optical system.
[0032]
As described with reference to FIGS. 2 to 4, the position of the grid point is almost constant on the directly-facing image data even if the height of the measurement object changes, but the height change on the converted image data. , The position of the grid point moves. Further, the shape of the pattern captured in the directly facing image data and the shape of the pattern in the converted image data are similar. Under such conditions, the height of the measurement object can be obtained based on the difference between the position of the grid point on the directly-facing image data and the position of the grid point on the converted image data.
[0033]
The principle of measuring the height of the measurement target based on the difference between the positions of the specific points (lattice points in this case) on the measurement target has been described above. However, height measurement under extremely limited conditions has been described as an example. That is, an example has been described in which the relationship between the pixels forming the oblique image data and the pixels forming the directly-facing image data is simply determined by a trigonometric function. However, in actuality, the relationship between the pixels constituting the oblique image data and the pixels constituting the directly-facing image data is more complicated, or the relationship between the pixels constituting the oblique image data and the pixels constituting the directly-facing image data is more complicated. The relationship may differ from device to device.
[0034]
Therefore, actually, as shown in FIG. 5, first, image conversion parameters indicating the relationship between the pixels forming the oblique image data and the pixels forming the converted image data, An image conversion parameter for converting the oblique image data so that the shape of the pattern is similar to the shape of the corresponding pattern in the converted image data is acquired (S1). When an image of an object is captured, the difference between the position of the characteristic point on the object on the directly-facing image data and the position on the converted image data has a value corresponding to the height of the object. Therefore, height conversion information indicating the relationship between the position difference and the height is obtained (S2). This height conversion information may not be obtained depending on the required degree of measurement accuracy. The above steps (S1 and S2) are hereinafter also referred to as “calibration”. Next, a feature pattern is registered (S3). Here, the same pattern as that on the object whose height is to be measured is registered. When a test pattern (for example, a lattice pattern) used in the calibration is on the target, the test pattern may be registered. Then, the object located at an arbitrary height is imaged by the directly-facing imaging device 12 and the oblique imaging device 10 (S4). Next, the oblique image data is converted into converted image data based on the image conversion parameters (S5). Next, for a specific point on the feature pattern of the object, a difference between the position on the directly-facing image data and the position on the converted image data, which has a value corresponding to the height of the object, is detected (S6). . Here, when a pattern different from the test pattern is registered as a feature pattern, a pattern search is performed using the registered feature pattern to determine the position of a specific point. Next, the height of the target object is detected based on the position difference (S7). Here, when the height conversion information has been acquired, the height is detected using the height conversion information.
[0035]
Next, the calibration performed before the measurement will be described separately for the first calibration for generating the image conversion parameter and the second calibration for generating the height conversion information.
[0036]
<First calibration>
For convenience of explanation, a case where calibration is performed using a test pattern composed of a lattice pattern will be described.
[0037]
In the first calibration, first, the test pattern arranged at the reference height (height = 0) is imaged by the directly-facing imaging device 12 and the oblique imaging device 10, respectively. FIG. 3A shows the directly-facing image data 3a of the test pattern imaged at the reference height, and FIG. 3B shows the oblique image data 3b. In FIGS. 3A and 3B, the center point of the test pattern (the grid point at the center of the test pattern) is located at the center of each of the image data 3a and 3b. That is, the center point of the directly-facing image data 3a and the center point of the oblique image data 3b match. In the facing image data 3a and the oblique image data 3b, the horizontal line connecting the grid points of the test pattern is made parallel to the X axis, and the vertical line is made parallel to the Y axis. That is, each vertical line connecting the grid points of the directly facing image data 3a is in the same direction as each vertical line connecting the grid points of the oblique image data 3b, and each horizontal line connecting the grid points of the directly facing image data 3a. Are in the same direction as the horizontal lines connecting the grid points of the oblique image data 3b. Further, the coordinate value of the center point of the directly-facing image data 3a and the coordinate value of the center point of the oblique image data 3b match. The interval between the vertical grid points in the central portion of the directly facing image data 3a is made equal to the interval between the vertical grid points in the central portion of the oblique image data 3b. Such a condition is obtained as a result of adjusting the directly-facing imaging device 12, the oblique imaging device 10, and the like, and may deviate from such a condition. For example, the position of the center point may not completely match between the directly-facing image data 3a and the oblique image data 3b.
[0038]
Next, a plurality of grid points are searched with the cross mark on the directly-facing image data 3a and the oblique image data 3b to obtain the coordinates of each grid point.
[0039]
Then, the coordinates of the grid points on the directly-facing image data 3a are associated with the coordinates of the grid points on the oblique image data 3b, and the coordinate conversion parameters are calculated. The calculated coordinate conversion parameters are stored in the storage device. Further, based on the coordinate conversion parameters, an image conversion parameter indicating a relationship between each pixel of the oblique image data 3b and each pixel of the converted image data 3c is calculated. The calculated image conversion parameters are stored in the storage device.
[0040]
The procedure of the first calibration described above is summarized as follows with reference to the flowchart shown in FIG.
(A) A test pattern having a plurality of characteristic points arranged at a specific height serving as a reference is imaged by a directly-facing imaging device and an oblique imaging device (S11).
(B) Obtain coordinates (diagonal plane coordinates) of a plurality of pixels in the oblique image data respectively corresponding to a plurality of characteristic points in the test pattern, and obtain positive coordinates corresponding to the plurality of characteristic points in the test pattern, respectively. The coordinates of the plurality of pixels in the paired image data (facing plane coordinates) are acquired (S12).
(C) The coordinates of the pixels constituting the oblique image data and the coordinates of the pixels constituting the facing image data are determined based on the oblique plane coordinates and the facing plane coordinates respectively corresponding to a plurality of characteristic points in the test pattern. A coordinate conversion parameter indicating the relationship is calculated (S13).
(D) An image conversion parameter indicating the relationship between the pixels forming the oblique image data and the pixels forming the converted image data which is the image data after the conversion of the oblique image data, and which is mapped to the directly facing image data. An image conversion parameter for converting the oblique image data is calculated so that the shape of the arbitrary pattern and the shape of the corresponding pattern in the converted image data are similar (S14).
[0041]
<Second calibration>
In the second calibration, the steps shown in FIG. 7 are sequentially performed while moving the test pattern stepwise in the height direction. That is, the test pattern is moved stepwise in the height direction (S21), and the test pattern is imaged at each height by the directly-facing imaging device and the oblique imaging device (S22). The oblique image data captured at each height is converted into converted image data based on the image conversion parameters obtained by the first calibration (S23), and each lattice point is searched on the directly-facing image data, A grid point is searched on the converted image data (S24), and a difference between the positions of the grid points is calculated as shown in Expression 2 (S25).
Position difference = grid point position on converted image data−grid point position on directly-facing image data (Formula 2)
The calculated height conversion information is stored in the storage device in association with the position difference and the height (S26). The above steps (S21 to S26) are performed until the movement of the test pattern within the height measurement range is completed (S27). FIG. 8 shows an example of the height conversion information. In FIG. 8, the difference between the positions of the grid points is calculated for the reference height (height = 0), a plurality of heights higher than the reference height, and a plurality of heights lower than the reference height. In FIG. 8, the difference between the positions at the reference height is “0”, but the difference is not necessarily “0”.
[0042]
Further, in FIG. 8, the relationship between the position difference and the height is linear, but it is not actually linear. This is due to the fact that the scale changes due to a change in the imaging distance of the optical system accompanying a change in the measurement height. In particular, in the vicinity of the upper and lower ends of the depth of field of the optical system, the difference from the scale at the reference height is generally the largest. In this embodiment, the error is corrected by calculating the height conversion information at the time of calibration and storing it in a storage device, and obtaining the height based on the height conversion information at the time of measurement.
[0043]
Further, as described above, there is a slight difference in scale between the directly facing optical system 12a and the oblique optical system 10a according to the measurement height. Therefore, the scale of the image data changes according to the measurement height. However, such a difference in scale of the image data due to the measured height between the directly facing optical system 12a and the oblique optical system 10a is negligible in pattern matching. To improve the matching score, pattern matching may be performed by changing the size of the feature pattern to be compared.
[0044]
<Acquisition of coordinate transformation parameters>
Here, acquisition of the coordinate conversion parameters in the first calibration will be described in detail.
[0045]
First, as a premise, the light receiving planes 10b and 12b in FIG. 9 are provided with a minute light receiving device that receives light and generates a voltage or a current proportional to the intensity of the light (each of the light receiving devices is referred to as a “pixel”). ) Are uniformly arranged in a matrix. A set of sets of light intensities corresponding to pixels arranged in order on the light receiving planes 10b and 12b is referred to as "image data". In order to identify individual pixels in order on the light receiving planes 10b and 12b, rectangular coordinates are provided on the light receiving planes 10b and 12b, and along the rectangular coordinate axes, the i-th axis and the vertical axis are counted from the starting point in the horizontal axis direction. The pixel at the j-th position in the direction will be referred to as a pixel [i, j]. That is, the pixel [i, j] is a pixel on the i-th row and the j-th column.
[0046]
In order to define the positional relationship between the oblique light receiving plane 10b and the directly-facing light receiving plane 12b, as shown in FIGS. 10A and 10B, the two light receiving planes have an (x, y) coordinate system. (Hereinafter, referred to as an “oblique plane coordinate system”) and an (X, Y) coordinate system (hereinafter, “facing plane coordinate system”) are provided. There is a relationship given by equations 3 and 4 between the two coordinate systems. Conversion between the oblique plane coordinate system and the directly-facing plane coordinate system based on the relationship given by Expressions 3 and 4 is hereinafter referred to as “coordinate conversion”.
X = (b ₁ x + b ₂ y + b ₃ ) / (B ₇ x + b ₈ y + 1) (Equation 3)
Y = (b ₄ x + b ₅ y + b ₆ ) / (B ₇ x + b ₈ y + 1) (Equation 4)
Where b _i (Where i = 1, 2, 3, 4, 5, 6, 7, 8) is a coordinate conversion parameter. Since there are a total of eight, if the relationship of four or more points on the directly-facing light receiving plane and the oblique light receiving plane is known, then bi (where i = 1, 2, 3, 4, 5, 6, 7, 8) can be obtained.
[0047]
That is, as shown in FIG. 10C, points corresponding to lattice points p, q, r, and s on the oblique light receiving plane correspond to points P, Q, R, and S on the directly-facing light receiving plane. As the coordinates of these eight points, (x _p , Y _p ), (X _q , Y _q ), (X _r , Y _r ), (X _s , Y _s ), (X _P , Y _P ), (X _Q , Y _Q ), (X _R , Y _R ), (X _S , Y _S ) Is obtained, the parameter b _i (Where i = 1, 2, 3, 4, 5, 6, 7, 8).
[0048]
Parameter b _i (However, i = 1, 2, 3, 4, 5, 6, 7, 8) is stored in the storage device, and the relationship between the oblique light receiving plane coordinates and the directly facing light receiving plane coordinates is unique. Will be decided. That is, by reading the coordinate values and the coordinate conversion parameters stored in the storage device by the central processing unit (CPU) and the above (Equation 3) and (Equation 4) and calculating them, the position on the oblique light receiving plane can be directly opposed. The position can be converted to a position on the light receiving plane to correspond.
[0049]
In the example described above, four points are selected on the directly-facing light receiving plane and four points on the oblique light receiving plane, respectively, and the coordinate conversion parameters are obtained from the corresponding relationships. However, the four points as described above must be necessarily selected. It doesn't have to be.
[0050]
The number of the selected plurality of points is at least four if at least three points are selected so as not to be aligned on a straight line. When three points on a straight line are selected and at least two points are selected outside the straight line, the number of the selected plurality of points is at least five. That is, a plurality of points that satisfy the above-described conditions may be selected. Of course, the more the selected points, the higher the calculation accuracy of the required coordinate conversion parameters.
[0051]
<Acquisition of image conversion parameters>
Here, acquisition of image conversion parameters in the first calibration will be described in detail.
[0052]
First, as a premise, a description will be given of coordinates on the light receiving plane of the imaging apparatus and indices for designating pixels constituting the light receiving plane with reference to FIG. Here, the pixels are arranged, for example, in a matrix, and the shape of each pixel is assumed to be a square for simplicity, and the boundary region between adjacent pixels has only a negligible width. Since each pixel has light receiving sensitivity inside the square and the boundary area between adjacent pixels has only a negligible width, the description will be made assuming that the pixel has light receiving sensitivity everywhere on the light receiving plane.
[0053]
Assuming that the dimension of one side of the pixel is L, the center of the pixel is shown by a black circle in FIG. 11, and the vertices of the four corners of the pixel are shown by crosses. As shown in FIG. 11, an xy coordinate system is used, and coordinate values are indicated as (x, y). On the other hand, an index for designating each pixel is represented as [i, j]. These i and j correspond to the i-th row and the j-th column. For example, the coordinates of the vertices of the four corners of the pixel specified by the index [0,0] are (+ 0.5L, -0.5L), (+ 0.5L, + 0.5L), (-0.5L, +0). .5L) and (-0.5L, -0.5L), and the center coordinates of the pixel are (0,0). Generally, the coordinates of the vertices of the four corners of the pixel specified by the index [i, j] are ((i + 0.5) L, (j−0.5) L), ((i + 0.5) L, (j + 0. 5) L), ((i-0.5) L, (j + 0.5) L), and ((i-0.5) L, (j-0.5) L), and the center coordinates of the pixel Is (iL, jL).
[0054]
In addition, x or y representing coordinate values, and index values i, j or k, m for designating pixels are displayed using lowercase letters as values relating to oblique light receiving planes, and values relating to directly facing light receiving planes. Shall be displayed using uppercase letters. That is, the coordinate value of the oblique light receiving plane is (x, y), and the index for designating the pixels forming the oblique light receiving plane is [i, j] or [k, m]. On the other hand, the coordinate value of the directly-facing light-receiving plane is (X, Y), and the index for designating the pixels constituting the directly-facing light-receiving plane is [I, J]. Here, since the directly-facing position is a position directly above the object, the directly-facing image data is also referred to as “directly above image data”.
[0055]
With reference to FIG. 12, a method of obtaining the light intensity of each pixel constituting the light receiving plane will be described. FIG. 12 shows the index [i _p , I _q (Where p = 1, 2, 3, 4, 5 and q = 1, 2) are enlarged and displayed. A case where a black band-like image having a width is formed on the pixel will be described. In FIG. 12, a hatched portion is a black image portion. It is assumed that light does not reach the black portions shown by hatching, and light of uniform brightness reaches the other white portions. It is assumed that the light intensity received by each pixel is digitized into 256 gradations. That is, it is assumed that the pixel is set to output brightness information as digital data in which 0 indicates no light is received and 255 indicates the brightest light is received. In this case, the index [i _p , I _q (Where p = 1, 2, 3, 4, 5 and q = 1, 2), the digital value captured by the ten pixels as one of the image data is
[I ₁ , J ₂ ] = 0, [i ₂ , J ₂ ] = 10, [i ₃ , J ₂ ] = 40, [i ₄ , J ₂ ] = 96, [i ₅ , J ₂ ] = 128,
[I ₁ , J ₁ ] = 20, [i ₂ , J ₁ ] = 0, [i ₃ , J ₁ ] = 0, [i ₄ , J ₁ ] = 0, [i ₅ , J ₁ ] = 0
And These values are proportional to the area occupied by the bright portions in each pixel.
[0056]
Here, for the sake of simplicity, an image composed of two regions, a pure white portion and a pure black portion, has been described as an example. However, the same applies to an image including a more complicated region of intermediate brightness. The digital value captured by a pixel as one of the image data is a value proportional to the product of the brightness of the pixel projected on the pixel and the area occupied by the brightness portion on the pixel.
[0057]
Next, referring to FIG. 13, a pixel for acquiring directly-facing image data (hereinafter, also simply referred to as “facing pixel”) and a pixel for acquiring oblique image data (hereinafter, simply referred to as “oblique pixel”). Will be described.) FIG. 13 shows the pixels existing in the range from i to i + 2 for the column positions of the oblique pixels, separated by lines. That is, the upper left pixel is a pixel given by the index [i, j], and the lower right pixel is a pixel given by [i + 2, j + 2]. The square ABCD represents an image in which the directly-facing pixel given by the index [I, J] is projected on a pixel that obtains a plurality of oblique pixels. Hereinafter, a pixel given by an index [I, J] or [i, j] or the like is abbreviated as a pixel [I, J] or a pixel [i, j], respectively.
[0058]
The vertices A, B, C, and D of the pixel [I, J] projected on the oblique pixel are A ((I-0.5) L and (J-0.5) L on the oblique light receiving plane, respectively. ), B ((I + 0.5) L, (J-0.5) L), C ((I + 0.5) L, (J + 0.5) L), D ((I-0.5) L, J + 0.5) L). The coordinates ((I-0.5) L, (J-0.5) L), ((I + 0.5) L, (J-0.5) L) of the four vertices of pixel [i, j], ( (I + 0.5) L, (J + 0.5) L) and ((I-0.5) L, (J + 0.5) L) are indicated by A, B, C, and D on the directly facing light receiving plane, respectively. Projected to a location.
[0059]
Next, the mathematical expression (1) that provides image data conversion will be described.
[0060]
Img2 (I, J) = Σ [W (k, m) × Img1 (k, m)] / ΣW (k, m) (Equation 1)
Assume that the pixel [I, J] is projected onto a pixel forming an oblique light receiving plane as shown in FIG. In this case, the range for obtaining the sum indicated by Σ is a range from i to (i + 2) for k, and a range from j to (j + 2) for m.
[0061]
Here, the meaning of W (k, m) will be described. W (k, m) represents the area of the pixel [I, J] projected on the pixel [k, m] forming the oblique light receiving plane. That is, a part of the pixel [I, J] is projected on the pixel [k, m], but W (k, m) is the image of the pixel [I, J] on the pixel [k, m]. Represents the area of the projected and overlapping portion. The area of the overlapping portion is also referred to as “partial projection area”. FIG. 14 is a diagram for explaining the meaning of W (k, m). In FIG. 14, k ranges from i to (i + 2) and m ranges from j to (j + 2). FIG. 9 is a diagram in which a portion where the image of [I, J] is projected and overlapped is indicated by oblique lines and displayed in an easily viewable manner. Each square represents a pixel [k, m]. A hatched portion indicates a polygon having an area given by W (k, m) (a portion of a partial projection area where the image of the pixel [I, J] is projected on the pixel [k, m]).
[0062]
(Equation 1) is a range of pixels (i is a range from i to (i + 2) for a pixel [I, J] projected on a pixel forming an oblique light receiving plane, and a range from j to (j + 2) for m. Of the pixel [I, J] on the directly-facing light-receiving plane (digital value related to brightness).
[0063]
Img1 (k, m) is a digital value (value of pixel [k, m]) relating to the brightness received by pixel [k, m], and W (k, m) is pixel [k, m] Since the image partial projection area of the pixel [I, J] above, W (k, m) × Img1 (k, m) is the value of the pixel [k, m] of W (k, m). The value takes into account the weight. Σ [W (k, m) × Img1 (k, m)] is the sum of k over the range from i to (i + 2) and m over j from (j + 2). ΣW (k, m) represents the total projected area of the image of the pixel [I, J] projected on the oblique light receiving plane. Therefore, the right side of (Equation 1) is defined as a range of pixels projected on pixels where the pixel [I, J] forms a directly-facing light receiving plane (a range from i to (i + 2) for k, and a range for m for m). For the range from j to (j + 2)), a weighted average value with W (k, m) as the weight is given.
[0064]
Next, a procedure for determining W (k, m) will be described. FIG. 15 is a diagram provided to explain a procedure for obtaining W (k, m), and is a diagram in which a peripheral portion of a pixel [i, j] is enlarged. Let ABCCD be the position where the four vertices of the pixel [I, J] are projected on the oblique light receiving plane. The four vertices of the pixel [i, j] on the directly-facing light receiving plane are A ', B', C ', and D'.
[0065]
W (k, m) is determined by the following procedure (process).
(A) It is checked whether the points A, B, C, and D are included in the quadrilateral A'B'C'D '.
A is picked up because A is included.
(B) It is checked whether or not each of the points A ′, B ′, C ′, and D ′ is included in the quadrilateral ABCD.
Since C 'is included, C' is picked up.
(C-1) The intersection of the line segment A′B ′ with the line segment AB, the line segment BC, the line segment CD, and the line segment DA is checked.
There are no intersections.
(C-2) The intersection of the line segment B′C ′ with the line segment AB, the line segment BC, the line segment CD, and the line segment DA is checked.
Since a is an intersection, a (existing as an intersection of line segment B′C ′ and line segment AB) is picked up.
(C-3) The intersection of the line segment C′D ′ with the line segment AB, the line segment BC, the line segment CD, and the line segment DA is examined.
Since b is the intersection, b is picked up (existing as the intersection of line segment C'D 'and line segment DA).
(C-4) The intersection of the line segment D'A 'with the line segment AB, the line segment BC, the line segment CD, and the line segment DA is examined.
There are no intersections.
(D) The upper left point is obtained for the picked-up point (four points are registered in the list in the order of A, C ', a, and b in this embodiment). Specifically, as shown in FIG. 15, each line is drawn one by one so that a straight line inclined at 45 degrees with respect to the X axis passes through each point, and the intersection of the Y axis and the straight line inclined at 45 degrees is drawn. The smallest one of the Y coordinates is selected. The reason for selecting the smallest Y coordinate is that the coordinate set on the light receiving plane is a left-handed coordinate system. Thus, the first point is obtained. Here, it is point A.
(E-1) On the basis of a straight line passing through the point A and having an inclination of 45 degrees, the straight line is rotated clockwise around the point A to determine a point that first comes into contact with the point, and this is a second point. And In this case, point a is obtained as the second point.
(E-2) Next, the straight line Aa is rotated clockwise about the point a to find a point that first comes into contact, and this is set as a third point. In this case, point C ′ is obtained as the third point.
(E-3) Next, the straight line aC 'is rotated clockwise about the point C' to find the first contact point, and this is set as the fourth point. In this case, point b is obtained as the fourth point.
(E-4) The above steps are sequentially performed on the picked-up points. By this process, the picked up points are arranged clockwise in order from the upper left.
(F) If the picked-up points are arranged clockwise in order from the upper left, the coordinates of the four vertices of the quadrilateral have been determined, and the coordinates of the quadrilateral AaC'b The area, that is, W (k, m) can be obtained.
[0066]
Next, a method for checking whether or not the point of interest O is included in the quadrilateral, which is performed in the steps (a) and (b), will be described with reference to FIG. FIG. 17A is a diagram illustrating a case where the point O is inside the quadrilateral EFGH, and FIG. 17B is a diagram illustrating a case where the point O is outside the quadrilateral EFGH. Starting from point E of the quadrilateral EFGH, there is a point that moves on the side in the order of point E, point F, point G, and point H. At this time, the angle θ1 extending from the point E to the point F is measured with the direction from the point E to the point F being a positive direction. Similarly, the angle between the point F and the point G is θ2, the angle between the point G and the point H is θ3, and the angle between the point H and the point E is θ4. The direction toward the point shall be measured in the positive direction.
[0067]
When the point O shown in FIG. 17A is inside the quadrilateral EFGH, θ1 + θ2 + θ3 + θ4 is 2π. On the other hand, when the point O shown in FIG. 17B is outside the quadrilateral EFGH, θ1 + θ2 + θ3 + θ4 becomes zero. Although not shown, when the point O is on the side of the quadrilateral EFGH, θ1 + θ2 + θ3 + θ4 becomes π. As described above, θ1 + θ2 + θ3 + θ4 is obtained, and if it is known whether the value is 2π, 0, or π, it is determined whether the point O of interest is included in the quadrilateral EFGH. You can find out.
[0068]
In the steps (a) and (b) of the above-described steps for obtaining W (k, m), the points to be picked up are located on the sides of the quadrilateral, and those located inside the quadrilateral. Treated as
[0069]
Next, with reference to FIG. 18 and FIG. 19, the intersections of the line segments, which are performed in the above-described steps (c-1), (c-2), (c-3), and (c-4), Explain the rules when requesting. FIG. 18 illustrates a case where the line segments do not overlap, and FIG. 19 illustrates a case where the line segments overlap. If an intersection exists on the extension of the line segment as shown in FIG. 18 (A), it is treated as having no intersection. It is determined that the intersection exists, as shown in FIG. 18B, when the intersection exists within the range where the line segment exists. As shown in FIG. 18C, it is also determined that an intersection exists even when the end point of the line segment is an intersection. When one line segment 1 and another line segment 2 overlap as shown in FIG. 19, two points are picked up as intersections.
[0070]
Further, the following processing is performed as a derivative step.
(A) In the above step (a), if all the points of A, B, C, and D are not included in the quadrilateral A'B'C'D ', W (k, m) = It ends with 0.
(B) In the step (b) described above, if all points of A ′, B ′, C ′, and D ′ are not included in the quadrilateral ABCD, W (k, m) = 0, finish.
(C) When the above steps (c-1) to (c-4) are completed, the number of picked up points is one to two or three to eight.
Therefore,
(C1) If the number of picked up points is 0 or 2, W (k, m) = 0 and the processing ends.
(C2) When the number of picked-up points is three to eight, processing is performed up to step (f) to obtain W (k, m).
[0071]
FIG. 13 shows the relationship between the pixel for the image directly above and the pixel for the oblique image, which shows the relationship between the quadrilateral (the shape of the pixel for the image directly above) and the square (the shape of the pixel for the image oblique). Things. In such a relationship, when the relationship between the quadrilateral and the square becomes the relationship shown in FIG. 14, the number of points picked up is eight, which is the maximum number. FIG. 20 shows the relationship between pixels for directly above images (quadrilaterals) and pixels for oblique images (squares). From this, the maximum number of points to be picked up is eight.
[0072]
The procedure for obtaining W (k, m) described above is only an example, and there may be other suitable procedures other than the method described here. However, no matter what procedure is used to determine W (k, m), it is clear that the present invention can be applied to the measuring method or measuring apparatus according to the present invention.
[0073]
The procedure for obtaining W (k, m) described above is summarized as follows with reference to the flowchart shown in FIG.
[0074]
The positions where the four vertices of the pixel [I, J] forming the directly-facing light-receiving plane are projected on the oblique light-receiving plane are A, B, C, and D. The four vertices are A ', B', C ', D'
(D) It is checked whether the points A, B, C, and D are included in the quadrilateral A'B'C'D '(S31). If neither is included (N), W (k, m) is set to W (k, m) = 0 (S32), and this procedure ends. On the other hand, if any point is included (Y), the included point is picked up (S33).
(E) It is checked whether or not each of the points A ', B', C ', and D' is included in the quadrilateral ABCD (S34). If neither is included (N), W (k, m) is set to W (k, m) = 0 (S35), and this procedure ends. On the other hand, if any point is included (Y), the included point is picked up (S36).
(F) Next, the intersection of line segment A'B 'and line segment AB, line segment BC, line segment CD, line segment DA, line segment B'C' and line segment AB, line segment BC, line segment CD, Intersection of line segment DA,
Intersection of line segment C'D 'with line segment AB, line segment BC, line segment CD, line segment DA, and
The intersections of the line segment D'A 'and the line segment AB, the line segment BC, the line segment CD, and the line segment DA are examined and picked up (S37).
(G) W (k, m) is obtained from all coordinates picked up in the steps S33, S36 and S37 (step S38).
[0075]
As described above, in the first calibration, W (k, m) is obtained as an image conversion parameter.
[0076]
<Height measurement range>
Next, a range of measurable heights (hereinafter referred to as “height measurement range”) will be considered.
[0077]
In FIG. 22, the depth of field d2 of the oblique optical system is represented by the following Expression 5.
d2 = 1 / cosφ + sl × sinφ (Equation 5)
Here, l is the height measurement range, φ is the tilt angle of the oblique optical system, and sl is the pattern search range in the oblique optical system. For example, when the depth of field d2 of the oblique optical system is d2 = 1.5 mm, the height measurement range l = 0.5 mm, and the inclination angle φ is 30 degrees, the search range sl is obtained as 1.845 mm. Here, if the search range sl is to be increased, the optical axis 10c of the oblique optical system is oblique to the surface of the measurement target, and therefore the depth of field of the oblique optical system must be large. When φ = 30 degrees, it is not realistic to enable the search pattern to be searched in the entire field of view of the oblique optical system. Therefore, in the present embodiment, the search range sl is limited, and a feature pattern is searched within the search range sl.
[0078]
Although the case where the magnification of the facing optical system and the magnification of the oblique optical system are the same has been described as an example, the present invention includes a case where the magnification of the facing optical system and the magnification of the oblique optical system are different. In FIG. 22, the height measurement range 1 is very limited with respect to the depth of field L of the directly facing optical system 12a. When it is desired to increase the height measurement range, the magnification of the oblique optical system in the oblique direction is set smaller than the magnification of the directly-facing optical system. When the magnification of the oblique optical system is set to 0.5, the range of the imaging distance (working distance) is about 1.5 times and the depth of field is about 2.5 times as compared with the case where the magnification is 1. Longer, and the pixel resolution is doubled. When the magnification of the oblique optical system is set to 0.5, it is natural that the accuracy is halved, but the search range and the height measurable range in the XY directions are greatly improved.
[0079]
<Measuring device>
FIG. 23 shows an example of the measuring device. In FIG. 23, the illumination device 14 is disposed on an axis that is line-symmetric with the optical axis of the oblique imaging device 10 with the optical axis of the directly-facing imaging device 12 as the symmetry axis. The XYZθ device 20 is a device that freely moves the measurement target in the Z direction (height direction) and the θ axis direction (rotation direction) on the XY plane.
[0080]
The image obtaining unit 31 obtains directly-facing image data from the directly-facing imaging device 12 and obtains oblique image data from the oblique imaging device 10. The image conversion parameter acquisition unit 41 is an image conversion parameter that indicates the relationship between the pixels that form the oblique image data and the pixels that form the converted image data. Image conversion parameters for converting the oblique image data so that the shapes of the corresponding patterns in the image data are similar are calculated. The height conversion information acquisition unit 42 generates height conversion information indicating the relationship between the position difference and the height. The image data conversion unit 43 converts the oblique image data into converted image data. The image search unit 44 searches the image data for a characteristic pattern or a characteristic point. The position difference detector 45 detects a difference between a position on the directly-facing image data and a position on the converted image data for a characteristic point on the measurement object. The height detection unit 46 converts the difference between the positions into a height based on the height conversion information. The image data output unit 47 outputs image data.
[0081]
The XYZθ control unit 51 controls the XYZθ device 20 to carry out transport and position control of the measurement target. The XYZθ control unit 51 controls the height of the target so that the height of the target is constant. The thickness detector 52 detects the thickness of the plate-like measurement target. Specifically, the height of the plane on which the object to be measured is placed is measured, the height of the surface of the object to be measured is measured, and the difference between these two heights is detected as the thickness. The slack detecting unit 53 detects a slack of the measuring object having slack. The XYZθ adjustment unit 54 adjusts the XYZθ device 20 in the X, Y, Z, and θ directions.
[0082]
The storage device 60 stores various information such as coordinate conversion parameters, image conversion parameters, and height conversion information.
[0083]
The image conversion parameter obtaining unit 41, the height conversion information obtaining unit 42, the image data converting unit 43, the image searching unit 44, the position difference detecting unit 45, the height detecting unit 46, the image data output unit 47, the XYZθ control unit 51 , The thickness detecting section 52, the slack detecting section 53, and the XYZθ adjusting section 54 are constituted by a central processing unit (CPU).
[0084]
First Embodiment A case will be described in which the measuring method of the present invention is applied to height control processing as an alternative to autofocus. This height control process is performed, for example, in a semiconductor manufacturing apparatus having a Z table. A semiconductor substrate (measurement target) such as a wafer is placed on a Z table movable in a Z direction (hereinafter, referred to as a “height direction”), and the Z coordinate (height) of the measurement target is kept constant.
[0085]
First, prior to the actual height setting processing, a characteristic pattern of the surface of the measurement object is registered. A fiducial mark may be registered as a feature pattern. Such a characteristic pattern is imaged by the directly-facing imaging device from a direction directly facing the object to be measured. In addition, the height to be set of the measurement target (hereinafter, referred to as “set height”) is specified in advance. When the thickness of the measurement object is specified, the specified thickness is added to a predetermined specific height to obtain a set height. Also, approximate XY coordinates for searching for a feature pattern are specified.
[0086]
The actual height setting process is performed as shown in FIG. In FIG. 24, first, the object to be measured is placed on the Z table, and the Z table is moved XY so that the object to be measured comes to the approximate XY coordinates specified in advance (S101). If the characteristic pattern on the surface of the measurement object is not within the search range, the measurement object is moved on the XY plane so that the characteristic pattern is directly below the directly facing optical system (S102). Next, the Z coordinate (height) of the measurement target is measured by the above-described measurement method of the present invention (S103). Next, a difference between the measured height and the set height is determined (S104). If there is a difference between the measured height and the set height, the Z table is moved in the height direction by the difference (S105), the height of the object to be measured is measured again (S103), and the measured height is measured. And the difference between the height and the set height are determined (S104). If the repeatability of the Z table is sufficiently good, it is determined that there is no difference in height in almost one Z table movement, and the loop ends.
[0087]
Further, the measuring method of the present invention can be applied to the case where the thickness of a semiconductor substrate having a pattern on a surface represented by a wafer is measured. Specifically, first, the height of the surface of the Z table on which the semiconductor substrate is placed is measured by the measuring method of the present invention. Next, the height of the surface of the semiconductor substrate is measured by the measuring method of the present invention. Then, the thickness of the semiconductor substrate is calculated from the difference between the height of the surface of the semiconductor substrate and the height of the surface of the Z table.
[0088]
(Second embodiment) A case where the measuring method of the present invention is applied to a slack measurement processing of a measuring object having slack will be described.
[0089]
FIG. 25 shows an example of a loose sheet (measurement object). The sheet 200 in FIG. 25 has a slack exceeding 1 mm. Incidentally, in the auto focus, an optical system (for example, 50 μm) having a shallow depth of field is generally used. Therefore, when trying to measure the slack of the sheet 200 of FIG. 25, a state occurs in which the measured surface of the sheet 200 moves beyond the depth of field, and in such a state, an image is not formed correctly (so-called defocusing). ), So it is not possible to measure the slack after all. On the other hand, in the measuring method of the present invention, an optical system having a depth of field (for example, 1.5 mm) corresponding to a slack measurement range (for example, from 0 mm to 1.5 mm) can be used.
[0090]
First, prior to actual slack measurement, the sheet 200 to be slack-measured is sucked by the suction mechanism 210 that can freely move in each of the XYZ directions, as shown in FIG. In addition, a specific mark (such as a fiducial mark) attached to the height measurement position for the slack measurement is registered in advance as a characteristic pattern. When there are a plurality of height measurement positions, feature patterns for the plurality of measurement positions are registered. In the registration of such a characteristic pattern, the characteristic pattern is imaged by the facing image pickup device. Even if the sheet is slack, the approximate value of the slack is known in advance before the measurement (it can also be obtained by visual measurement at the site), so the approximate value is specified in advance. Further, approximate XY coordinates for searching for a feature pattern are specified in advance.
[0091]
In the actual slack measurement process, the process shown in FIG. 26 is performed after the suction mechanism 210 is moved Z so that the measurement surface of the sheet falls within the height measurement range based on the approximate value of the slack. In FIG. 26, the suction mechanism 210 is moved XY so that the sheet comes to the approximate XY coordinates specified in advance (S201). If the characteristic pattern on the sheet 200 is not within the search range on the XY plane, the sheet 200 is moved on the XY plane so that the characteristic pattern is directly below the directly facing optical system (S202). Next, the height of the sheet 200 is measured at the position where the characteristic pattern is located by the above-described measuring method of the present invention (S203). At the same time, the XY plane coordinates of the position are determined. Next, the difference between the measured height and the set height is determined (S204). If the measured height is within the measurement range, the height measurement at the position is terminated. If there is another characteristic pattern, the height of the sheet is measured at the next position. If the measured height is out of the measurement range or if no characteristic pattern is found on the converted image data, the suction mechanism is moved in the height direction by the measurement range (S205).
[0092]
According to the slack measurement process described above, even if the sheet has a large slack (for example, 1 mm or more), the depth of field (for example, 1.5 mm) corresponding to the slack measurement range (for example, up to 1.5 mm) is obtained. The slack can be measured using the optical system, and the slack can be measured by several Z movements.
[0093]
Third Embodiment A case will be described in which the measurement method of the present invention is applied to θ axis adjustment of an XYθ mechanism in which a θ mechanism is mounted on an XY mechanism. FIG. 27 shows an example of the XYθ mechanism. In FIG. 27, the adjustment is performed so that the θ axis is orthogonal to the X axis and the Y axis.
[0094]
FIG. 28 shows a schematic processing flow of θ axis adjustment. In FIG. 28, first, the θ-axis table arranged to rotate with the θ-axis is adjusted to be horizontal with the XY plane (S310). FIG. 29 is a cross section 301 in the θ-axis tilt direction of FIG. 27, and shows the θ-axis table 310 in a state adjusted horizontally with respect to the XY plane. In the state where the θ-axis table 310 is adjusted horizontally with respect to the XY plane in this manner, the angle formed by the plane 302 perpendicular to the θ-axis 300 and the θ-axis table 310 is equal to the inclination angle φ of the θ-axis 300. As shown in FIGS. 30A and 30B, the θ-axis table 310 has a disk shape and is formed of a lattice pattern. Of the plurality of grid points on such a grid pattern, the grid point 311 arranged at the center of the θ-axis table 310 and positioned at the rotation center position of the θ-axis 300 is hereinafter referred to as “rotation center grid point”. Also, the grid point 312, which is arranged at a position (here in the vicinity of the outer periphery of the θ axis table 310) of the “rotation center grid point” 311 and at which the XYZ coordinates are measured for the θ axis adjustment, is referred to as “θ axis adjustment”. Grid point ".
[0095]
Next, while rotating the θ-axis table 310, the XYZ coordinates of the θ-axis adjustment grid point 312 are measured, and the inclination of the θ-axis is detected (S320). FIG. 31 shows a detailed flow of the θ-axis tilt detection processing (S320).
[0096]
In FIG. 31, first, the rotation center position of the θ axis is detected (S321), and the rotation center position of the θ axis 300 is aligned with the rotation center lattice point 311 of the θ axis table 310 (S322). FIG. 32 is a rotation angle 0 degree image captured without rotating the θ axis, FIG. 33A is a counterclockwise rotation image captured by rotating the θ axis counterclockwise by a predetermined angle θ1, and FIG. B) is a clockwise rotation image captured by rotating the θ axis clockwise by the predetermined angle θ2. FIG. 34 is an image obtained by superimposing the three images shown in FIGS. 32, 33A and 33B, respectively. Specifically, the superimposed image is displayed, and a rough rotation center position is designated using a pointing device such as a mouse. Then, the pattern of the rotation angle of 0 degree in FIG. 35A, the pattern of the counterclockwise rotation of FIG. 35B and the pattern of the clockwise rotation of FIG. 33A and the clockwise rotation image of FIG. 33B, respectively, to obtain coordinate values of a plurality of grid points. Then, the rotation center position (Xc, Yc) is obtained by the least square method using the coordinate values of the corresponding grid points. Here, the rotation is performed in both the counterclockwise direction and the clockwise direction. However, the rotation center position may be obtained only in one of the counterclockwise direction and the clockwise direction.
[0097]
Next, the grid point 312 for θ-axis adjustment is moved directly below the directly facing optical system, and the XYZ coordinates of the grid point 312 for θ-axis adjustment are measured (S323). Here, the XY coordinates of the θ-axis adjustment grid point 312 are measured based on directly-facing image data captured by the directly-facing imaging device. Further, the Z coordinate of the θ-axis adjustment grid point 312 is measured based on the directly-facing image data captured by the directly-facing imaging device and the oblique image data captured by the oblique imaging device according to the measuring method of the present invention.
[0098]
Then, the θ-axis is rotated by a fixed angle, and the grid point 312 for θ-axis adjustment is moved directly below the directly facing optical system to measure the XYZ coordinates of the grid point 312 for θ-axis adjustment (S324). Here, the XYZ coordinate measurement of the θ-axis adjustment grid point 312 (S324) is repeatedly performed at a fixed angle, and it is determined whether the rotation has been performed by 2π (S325). When the θ axis rotates π from the state shown in FIG. 29, the θ axis table becomes the state shown in FIG. When the rotation is performed by 2π in total, data indicating the inclination of the θ axis as shown in FIG. 37 is obtained (S326). Here, the inclination of the θ-axis 300 is represented by an inclination direction 303 and an inclination magnitude 304. The inclination direction 303 is obtained from the θ-axis rotation angle when the θ-axis adjustment grid point 312 is the lowest, and the inclination 304 is the maximum value and the minimum height of the θ-axis adjustment grid point 312. Obtained from the difference between the values.
[0099]
After such θ-axis tilt detection, it is determined whether or not the θ-axis tilt is within an allowable range (S330 in FIG. 28). If the angle is not within the allowable range, the installation of the θ-axis is mechanically adjusted (S340), the θ-axis table is adjusted horizontally again (S310), the inclination of the θ-axis is detected (S320), and the inclination of the θ-axis is allowable. It is determined whether it is within the range (S330). If it is within the allowable range, the θ axis adjustment processing ends.
[0100]
The case where the θ axis is rotated by 2π and the inclination of the θ axis is measured has been described above. However, there are cases where the θ axis cannot be rotated by 2π. If the θ axis cannot be rotated by 2π but can be rotated by π (180 degrees), as shown in FIG. 38, two θ axis adjustment grid points 312 and 313 located at point symmetry with respect to the rotation center grid point 311 are used. Is used. FIG. 39A shows a θ-axis table when the rotation angle of the θ-axis is 0 degrees adjusted to the XY plane, and FIG. 39A shows a θ-axis table when the rotation angle of the θ-axis is 180 degrees. Shown in B). While rotating the θ axis within 180 degrees, the XYZ coordinate values of the two θ axis adjustment lattice points 312 and 313 are measured, and the two θ axis adjustment lattice points 312 and 313 outside the θ axis measurement range are measured. The data shown in FIG. 40 is obtained by estimating the XYZ coordinate values. In FIG. 40, the Z coordinate of the first θ-axis adjustment grid point 312 and the second θ-axis adjustment grid point 313 is such that when the θ-axis rotation angle is 0 degree, the θ-axis table 310 is horizontal on the XY plane. So it matches. When the θ-axis is rotated by 180 degrees, the Z displacement of the first θ-axis adjusting grid point 312 and the Z displacement of the second θ-axis adjusting grid point 313 become maximum values. FIG. 40 shows the Z displacement range 3121 of the first θ-axis adjustment lattice point 312 and the Z displacement range 3131 of the second θ-axis adjustment lattice point 313.
[0101]
The data shown in FIG. 40 is obtained when the straight line connecting the two θ-axis adjustment grid points 312 and 313 matches the inclination direction of the θ-axis, and the two θ-axis adjustment grid points 312, If the straight line connecting 313 does not match the tilt angle of the rotation θ axis, the data is as shown in FIG. In FIG. 41, the Z displacement range 3122 of the first θ-axis adjustment grid point 312 and the Z displacement range 3132 of the second θ-axis adjustment grid point 313 are different from the first θ-axis adjustment grid point 312 and the second θ-axis adjustment grid point 312. When the line connecting the θ-axis adjustment lattice points 313 coincides with the inclination direction of the θ-axis (at an angle corresponding to the vertical broken line in the figure), the maximum value is obtained.
[0102]
Next, a case where the rotatable range of the θ axis is smaller than 180 degrees will be described. In such a case, as shown in FIG. 42, four θ-axis adjustment grid points 312, 313, 314, and 315 are selected, and the θ-axis is rotated within the θ-axis measurement range (306 in FIG. 43). When the XYZ coordinate values of the four θ-axis adjustment grid points 312, 313, 314, and 315 are measured, data indicated by solid lines in FIG. 43 is obtained. In FIG. 43, the Z displacement amount of the first θ-axis adjustment lattice point 312 is 3120, the Z displacement amount of the second θ-axis adjustment lattice point 313 is 3130, and the Z displacement amount of the third θ-axis adjustment lattice point 314. The displacement amount is indicated by 3140, and the Z displacement amount of the fourth θ-axis adjustment grid point 315 is indicated by 3150. In FIG. 43, for convenience of description, the line connecting the grid points for the θ-axis adjustment (the line connecting the grid point 312 for the first θ-axis adjustment and the grid point 313 for the second θ-axis) is the θ-axis. The case where the inclination direction coincides with the inclination direction is shown.
[0103]
FIG. 44 shows a detailed flow of θ-axis tilt detection (corresponding to S320 in FIG. 28) using the four θ-axis adjustment grid points 312, 313, 314, and 315. In FIG. 44, first, the rotation center position of the θ axis is detected (S321), and the rotation center position of the θ axis 300 is aligned with the rotation center lattice point 311 of the θ axis table 310 (S322). Then, the θ-axis adjustment grid points 312, 313, 314, and 315 are sequentially moved directly below the directly facing optical system, and the XYZ coordinates of the θ-axis adjustment grid points 312, 313, 314, and 315 are measured (S3231). ). Further, the θ-axis is rotated by a fixed angle, and the θ-axis adjustment grid points 312, 313, 314, and 315 are sequentially moved directly below the directly facing optical system, and the θ-axis adjustment grid points 312, 313, 314, and 315 are moved. The XYZ coordinates are measured (S3241). Here, the XYZ coordinate measurement of the θ-axis adjustment grid points 312, 313, 314, and 315 (S3241) is repeatedly performed at a fixed angle, and it is determined whether or not the θ-axis measurement range 306 is exceeded (S3251). If it exceeds the measurement range 306 of the θ axis, it is determined whether there is a point where the Z-direction displacement amount has the maximum value and the minimum value (S3252). , Rotate the θ-axis adjustment grid point (S3271). For example, the θ-axis adjustment lattice point is rotated by rotating the θ-axis table by hand or the like. Then, steps S321 to S3252 are repeated. If there is a point at which the maximum value and the minimum value are found, the direction and magnitude of the inclination of the θ axis are detected (S3261).
[0104]
The processing flow shown in FIG. 44 can also be applied to the case where the lines connecting the θ-axis adjustment grid points do not coincide with the inclination direction of the θ-axis.
[0105]
In FIGS. 31 and 44, the inclination direction of the θ axis on the XY plane is basically obtained. However, the adjustment can be performed without obtaining the inclination direction of the θ axis on the XY plane. When the θ-axis movable range is very small, such an adjustment method is easier. In the following, a description will be given assuming that the θ-axis movable range is narrow. First, the θ-axis table is adjusted to be horizontal with the XY plane. Next, the θ axis is rotated in the maximum + direction, and the XYZ coordinates of the θ axis adjustment grid points 312, 313, 314, 315 are measured. Subsequently, the θ-axis is rotated in the maximum negative direction, and the XYZ coordinates of the θ-axis adjustment grid points 312, 313, 314, and 315 are measured. When the Z coordinate values for the four θ-axis adjustment grid points 312, 313, 314, and 315 measured three times satisfy the accuracy in the specification of the apparatus, the adjustment is completed. If the accuracy in the specifications is not satisfied, the direction of the inclination on the XY plane of the θ axis and the magnitude of the inclination in that direction are estimated from the three measured values, corrected and corrected again. Perform
[0106]
Fourth Embodiment A case will be described in which the measurement method of the present invention is applied to Z-axis adjustment of an XYZ mechanism in which a Z mechanism is mounted on an XY mechanism. FIG. 45 shows an example of the XYZ mechanism. In FIG. 45, the adjustment is performed so that the Z-axis direction is orthogonal to the X-axis direction and the Y-axis direction. Here, the Z-axis adjustment is an adjustment to make the upper surface of the Z-axis table horizontal to the XY plane of the XY mechanism. It is assumed that pitching and yawing of the Z axis itself are adjusted in advance.
[0107]
The Z-axis adjustment is performed by placing an adjustment pattern 410 shown in FIG. 46 on a Z-axis table. At the four corner positions on the adjustment pattern 410, Z-axis adjustment lattice points 411 to 414 are provided. Specifically, the Z-axis adjustment is performed by moving the Z-axis table in the Z-axis direction, measuring the displacement of each of the Z-axis adjustment lattice points 411 to 414, and determining that the displacements of the four adjustment lattice points are constant. It is achieved by adjusting so that
[0108]
FIG. 47 shows a state before the Z-axis adjustment. The surface of the adjustment pattern 410 placed on the Z-axis table is inclined with respect to the XY plane 402 of the XY mechanism. In FIG. 47, the angle between the Z direction and the optical axis direction is very small. At the point measured geometrically at almost the same position in the optical axis direction, the difference between the length in the Z direction and the length in the optical axis direction is very small. When the XY table is moved as shown in FIG. 47 and measurement is performed at substantially the same position with respect to the optical system, the height can be detected as a difference in the Z direction. FIG. 48 is a flowchart showing a method of measuring the Z-axis coordinates. Here, the movable range in the Z-axis direction is long (for example, 30 mm), and is divided into sections in units of the depth of field (for example, 1 mm) of the optical system.
[0109]
Prior to performing the measurement shown in FIG. 48, an axis (optical system mounting axis) movable in a direction perpendicular to the XY plane is installed separately from the XYZ mechanism, and the optical system is mounted on this optical system mounting axis. The movable range of the optical system mounting shaft is longer than the movable range of the Z axis. Further, the Z-axis table is moved to the lowest point of the movable range of the Z-axis. Then, the measurement shown in FIG. 48 is performed.
[0110]
In FIG. 48, the optical system is moved such that the Z-axis adjustment grid point is located at the lower end of the depth of field of the optical system (S401). Next, attention is paid to a certain Z-axis adjustment grid point (for example, the first Z-axis adjustment grid point 411), and the X, Y, and Z coordinates of the grid point of interest 411 are measured (S402). The measured values here are (X1, Y1, Z1). Next, the Z-axis table is moved upward by the depth of field (1 mm) (S404), and the X, Y, and Z coordinates of the grid point of interest 411 are measured (S405). The measured values here are (X2, Y2, Z2). (X2-X1, Y2-Y1, Z2-Z1) is calculated as the displacement amount of the first section (S406). Next, the optical system is moved upward by the depth of field (1 mm) (S407), and the X, Y, and Z coordinates of the grid point of interest 411 are measured (S408). The measured values here are (X3, Y3, Z3). Again, the Z-axis table is moved upward by the depth of field (1 mm) (S404), and the X, Y, and Z coordinates of the grid point of interest are measured (S405). (X4-X3, Y4-Y3, Z4-Z4) are calculated as the displacement amount of the second section (S406). Steps S404 to S408 are performed until it is determined that the acquisition of the displacement amount in all the sections of the movable range of the Z axis has been completed (S403). Then, by integrating the displacement amounts in each section, a difference between the lowest point and the highest point of the movable range of the Z axis can be obtained. Note that, for convenience of explanation, the measurement is performed by focusing on one grid point. However, the accuracy can be improved by measuring and focusing on a plurality of grid points and performing statistical processing. For example, the X, Y, and Z coordinates of each of the Z-axis adjustment grid points 411, 412, 413, and 414 are measured while involving XY movement to the Z-axis adjustment grid points 411, 412, 413, and 414. The broken line in FIG. 47 indicates the Z-axis table 410 'when the XY movement to the second Z-axis adjustment grid point 412 is performed. The Z-axis adjustment grid point moves at least within the measurement range 420 and measures the coordinate value.
[0111]
The measurement data in the Z-axis movable range measured in this way can be used for adjusting the Z-axis tilt. Also, the measured data can be used as a correction value without adjusting the Z-axis tilt.
[0112]
Although the third embodiment has described the θ-axis adjustment of the XYZ θ mechanism and the fourth embodiment has described the Z-axis adjustment of the XYZ mechanism, these adjustment methods are used when assembling the XYZθ mechanism. The present invention can be similarly applied to mounting of the θ mechanism on the XYθ mechanism and mounting of the Z mechanism on the XYθ mechanism.
[0113]
【The invention's effect】
As described above, according to the present invention, the processing time can be reduced and the measurable range of the height can be expanded as compared with the case where the conventional autofocus is used.
[Brief description of the drawings]
BRIEF DESCRIPTION OF DRAWINGS FIG. 1 is a diagram for explaining a state in which a target object is imaged in a facing direction and an oblique direction in a measuring method of the present invention.
FIG. 2 is a diagram illustrating a cross section of a portion where an image of an object is imaged.
FIGS. 3A, 3B, and 3C are diagrams respectively illustrating examples of directly-facing image data, oblique image data, and converted image data obtained by imaging an object having a reference height. .
FIGS. 4A, 4B, and 4C show examples of facing image data, oblique image data, and converted image data obtained by imaging an object having a height lower than a reference height, respectively. FIG.
FIG. 5 is a flowchart showing an outline of a process in an embodiment of the measuring method of the present invention.
FIG. 6 is a flowchart illustrating an outline of a first calibration process;
FIG. 7 is a flowchart illustrating an outline of a second calibration process;
FIG. 8 is a diagram illustrating an example of height conversion information.
FIG. 9 is a diagram for explaining a state in which a lattice pattern formed on the surface of the object is imaged in a facing direction and an oblique direction.
FIGS. 10A, 10B, and 10C are diagrams showing oblique image data, directly facing image data obtained by imaging a lattice pattern, and a superposition thereof.
FIG. 11 is a diagram provided for describing an index for designating pixels constituting image data.
FIG. 12 is a diagram provided for explanation of pixel values.
FIG. 13 is a diagram illustrating a relationship between pixels forming the directly-facing image data and pixels forming the oblique image data at the reference height.
FIG. 14 is a diagram provided for description of W (k, m).
FIG. 15 is a diagram provided for describing a method of obtaining W (k, m).
FIG. 16 is a diagram for explaining points picked up for obtaining W (k, m).
17A and 17B are diagrams for explaining the relationship between a point O and a quadrilateral EFGH.
FIGS. 18A, 18B, and 18C are diagrams for explaining the determination of the presence or absence of an intersection with a line segment;
FIG. 19 is a diagram for describing a case where a line segment overlaps a line segment;
FIG. 20 is a diagram provided to explain the relationship between pixels forming the directly-facing image data and pixels forming the oblique image data.
FIG. 21 is a flowchart for obtaining W (k, m).
FIG. 22 is a diagram provided for describing a relationship between a depth of field of the optical system and a searchable range.
FIG. 23 is a main part block diagram showing one embodiment of a measuring device of the present invention.
FIG. 24 is a schematic flowchart illustrating a height control method according to the first embodiment.
FIG. 25 is a diagram illustrating a sheet having slack as an example of a measurement object.
FIG. 26 is a schematic flowchart illustrating a slack measurement method according to a second embodiment.
FIG. 27 is a view schematically showing an XYθ mechanism before θ-axis adjustment.
FIG. 28 is a schematic flowchart illustrating an XYθ mechanism adjustment method according to a third embodiment.
FIG. 29 is a cross-sectional view of an example of the θ-axis table, showing a state where the rotation angle is 0 ° when one θ-axis adjustment grid point is selected.
FIGS. 30A and 30B are diagrams illustrating examples of a θ-axis table when one grid point for θ-axis adjustment is selected.
FIG. 31 is a flowchart showing details of θ-axis tilt detection processing in the XYθ mechanism adjustment method of the third embodiment.
FIG. 32 is a diagram showing image data of a grid pattern having a rotation angle of 0 °.
33 (A) and (B) are diagrams showing image data of a grid pattern rotated counterclockwise and image data of a grid pattern rotated clockwise.
FIG. 34 is a diagram showing image data obtained by superimposing a grid pattern having a rotation angle of 0 °, a grid pattern rotated counterclockwise, and a grid pattern rotated clockwise.
FIGS. 35 (A), (B) and (C) are diagrams showing patterns used for alignment of the θ-axis rotation center.
FIG. 36 is a cross-sectional view of the θ-axis table, showing a state rotated by 180 degrees when one grid point for θ-axis adjustment is selected.
FIG. 37 is a diagram illustrating displacement of the height of the θ-axis adjustment grid point when one θ-axis adjustment grid point is selected.
FIG. 38 is a diagram illustrating an example of a θ-axis table when two θ-axis adjustment grid points are selected.
FIGS. 39A and 39B are cross-sectional views of the θ-axis table when two grid points for θ-axis adjustment are selected.
FIG. 40 is a diagram illustrating displacement of the height of the θ-axis adjustment grid point when two θ-axis adjustment grid points are selected.
FIG. 41 is a diagram illustrating a displacement of the height of the θ-axis adjustment grid point when two θ-axis adjustment grid points that do not match the direction of the θ-axis tilt are selected.
FIG. 42 is a diagram illustrating an example of a θ-axis table when four θ-axis adjustment grid points are selected.
FIG. 43 is a diagram illustrating displacement of the height of the θ-axis adjustment grid points when four θ-axis adjustment grid points are selected.
FIG. 44 is a flowchart illustrating details of θ-axis tilt detection processing when four θ-axis adjustment grid points are selected.
FIG. 45 is a view schematically showing an XYZ mechanism before Z-axis adjustment.
FIG. 46 is a diagram showing an example of a Z-axis table when four grid points for Z-axis adjustment are selected.
FIG. 47 is a sectional view showing the Z-axis table in a state where the Z-axis is inclined.
FIG. 48 is a flowchart showing an XYZ mechanism adjusting method according to a fourth embodiment to which the present invention is applied.
FIG. 49 is a diagram provided for explanation of auto focus used in a conventional measurement method.
FIG. 50 is a diagram provided for explanation of adjustment of a conventional XYZθ mechanism.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Oblique imaging device, 10a ... Oblique optical system, 10b ... Oblique light receiving plane, 12 ... Facing imaging device, 12a ... Facing optical system, 12b ... Facing light receiving plane, 14 ... Illumination device, 41 ... Acquisition of image conversion parameters Unit, 42: height conversion information acquisition unit, 43, image data conversion unit, 44, image search unit, 45, position difference detection unit, 46, height detection unit, 51, XYZθ control unit (height control unit), 52: thickness detection unit, 53: slack measurement unit, 54: XYZθ adjustment unit (XYZ adjustment unit and XYθ adjustment unit), 60: storage device

Claims (10)

Using oblique imaging means for imaging an object located at an arbitrary height from an oblique direction and facing image capturing means for imaging the object from a directly facing direction, the oblique image data imaged by the oblique imaging means and A measurement method for measuring the height of the object based on directly-facing image data captured by the paired imaging means,
An image conversion parameter indicating a relationship between pixels constituting the oblique image data and pixels constituting image data obtained by converting the oblique image data (hereinafter referred to as “converted image data”). A coordinate conversion parameter obtaining step of obtaining the image conversion parameter for converting the oblique image data so that the shape of an arbitrary pattern to be formed and the shape of the pattern in the converted image data are similar,
Image data acquisition step of acquiring the directly located image data obtained by imaging the object located at an arbitrary height by the directly-facing imaging means and oblique image data obtained by imaging by the oblique imaging means,
An image data conversion step of converting the oblique image data of the object into the converted image data based on the image conversion parameter,
For a specific point on the object, a position difference detecting step of detecting a difference between a position on the directly-facing image data and a position on the converted image data,
Based on the difference between the positions, a height detection step of detecting the height of the object,
A measuring method characterized by comprising: The image data conversion step,
The measurement method according to claim 1, wherein the light intensity Img2 (I, J) of the pixel [I, J] constituting the converted image data is obtained according to the following equation (1).
Img2 (I, J) = Σ [W (k, m) × Img1 (k, m)] / ΣW (k, m) (1)
Where I, J, k and m are integers,
Img1 (k, m) is the light intensity of the pixel [k, m] on the k-th row and the m-th column that constitutes the oblique image data.
W (k, m) is the image conversion parameter, and is the I-th row and J-th column on the light receiving plane of the directly-facing imaging means (hereinafter referred to as “directly-facing light receiving plane”) at a specific reference height. Represents the area projected on the pixel [k, m] of the oblique light receiving plane when the pixel [I, J] is projected on the light receiving plane (hereinafter referred to as “oblique light receiving plane”) of the oblique imaging means. .
The range for calculating the sum of the right side of the equation (1) is that the pixel [I, J] on the directly-facing light receiving plane projected on the oblique light receiving plane is the pixel [k, m] on the oblique light receiving plane. ], It is assumed that all the projected pixels [k, m] are projected. The image conversion parameter acquisition step,
A step of imaging a test pattern having a plurality of characteristic points arranged at a specific height serving as a reference by the directly-facing imaging unit and the oblique imaging unit,
The coordinates of a plurality of pixels in the diagonal image data respectively corresponding to the plurality of characteristic points in the test pattern (hereinafter, referred to as “diagonal plane coordinates”) are obtained. Acquiring coordinates of a plurality of pixels in the image data (hereinafter referred to as “facing plane coordinates”);
Calculating a coordinate conversion parameter indicating a relationship between the coordinates of the pixels constituting the oblique image data and the coordinates of the pixels constituting the facing image data, based on the oblique plane coordinates and the facing plane coordinates; ,
Calculating the image conversion parameter based on the coordinate conversion parameter at the specific height serving as the reference,
The measurement method according to claim 1, comprising: Moving a test pattern having a plurality of characteristic points stepwise in the height direction;
Imaging the test pattern at each height by the directly-facing imaging means and the oblique imaging means;
Converting oblique image data of the test pattern imaged at each height based on the coordinate conversion parameters and the image conversion parameters,
For a specific point on the test pattern, a step of detecting a difference between a position on the directly-facing image data of the test pattern and a position on the converted image data of the test pattern,
Generating height conversion information by associating each height of the test pattern with a difference between each position,
4. The measuring method according to claim 1, wherein the height detecting step detects a height of the object based on the height conversion information. A height control method, comprising: controlling the height of the object so that the height of the object measured by the measurement method according to claim 1 is constant. Oblique imaging means for imaging an object located at an arbitrary height from an oblique direction, and facing imaging means for imaging the object from a direction facing the object,
An image conversion parameter indicating a relationship between pixels constituting the oblique image data and pixels constituting image data obtained by converting the oblique image data (hereinafter referred to as “converted image data”). Coordinate conversion parameter obtaining means for obtaining the image conversion parameter for converting the oblique image data so that the shape of the arbitrary pattern and the shape of the pattern in the converted image data are similar,
Image data conversion means for converting the oblique image data of the object into the converted image data based on the image conversion parameters,
For a specific point on the feature pattern of the object, a position difference detection unit that detects a difference in position between the directly-facing image data and the converted image data,
Height detection means for detecting the height of the object based on the difference between the positions,
A measuring device comprising: 7. The measuring apparatus according to claim 6, wherein an illuminating unit is provided on an axis that is line-symmetric with the optical axis of the oblique imaging unit, with the optical axis of the facing imaging unit as a symmetric axis. 8. A measuring device according to claim 6, further comprising height control means for controlling the height of the object such that the height of the object measured by the measuring device is constant. And height control device. 9. A semiconductor manufacturing apparatus comprising: a cassette for stocking a plurality of semiconductor substrates; and transport means for removing and transporting semiconductor substrates one by one from the cassette, wherein the measuring device according to claim 6 or 7 or the measuring device according to claim 8 A semiconductor manufacturing apparatus comprising a height control device, wherein the object is the semiconductor substrate taken out of the cassette. A transport line comprising transport means connecting a plurality of processes and one or more inspecting means based on images, comprising: the measuring device according to claim 6 or 7 or the height control device according to claim 8, wherein A transport line, wherein an object is an object transported on the transport line.

Images