JP2001066123A - Shape measuring device and method of three-dimensional shape - Google Patents

Abstract

PROBLEM TO BE SOLVED: To measure with namometer level accuracy, by operating so as to move relatively high-frequency shape data and to join the data, so that high-frequency shape components corresponding to an overlapped region agree mutually at its maximum. SOLUTION: A white interferometer 6 equipped in its inside with a plane reference face as a partial face measuring means is installed on a Z-axis stage 3. The white interferometer 6 can obtain directly, without information of adjacent pixels, height information corresponding to each pixel of a CCD for taking in interference fringes. Therefore, even if the interval between the interference fringes is narrow, shape measurement can be executed, therefore, when a curved surface shape is measured, a wide measuring range can be secured. On the other hand, a measuring object 7 is placed on a Y-axis stage 2 and on rotary stages 4, 5, and positioned oppositely to the white interferometer 6. The direct stages 1, 2, 3 and the rotary stages 4, 5 are controlled by using a control device, and a relative position attitude between the white interferometer 6 and the measuring object 7 is determined, so that the optical axis of the white interferometer 6 roughly agrees with the normal of the measuring object 7.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and a method for measuring the shape of a three-dimensional shape such as the surface of an aspheric lens mounted on a laser printer or the like. Is what you do.

[0002]

2. Description of the Related Art It is known that an aspherical f.theta. Lens used in an optical system of a laser printer or the like has a waviness of a small amplitude having a wavelength of about several mm on its surface, which may cause image deterioration. I have. In recent years, there has been an increasing demand for higher image quality, and it has been required to measure and evaluate the surface undulation shape in a range of several mm square which causes image degradation with nanometer accuracy.

Conventionally, in order to evaluate the shape of an aspherical fθ lens,
A method of measuring a shape by scanning a surface using a mechanical stylus or an optical stylus has been widely used. However, while these methods are suitable for measuring a relatively large area,
It is not suitable for scanning a small area at high speed. For example,
When trying to acquire data in a range of several mm square at 0.1 mm intervals, it takes too much measurement time, so that measurement errors due to temperature drift cannot be ignored, and as a result, it is difficult to achieve nanometer measurement accuracy. there were.

As another method for measuring a three-dimensional fine shape, a microscope type laser interferometer utilizing light wave interference is well known. However, if an attempt is made to measure the shape of a curved surface such as an aspherical fθ lens, the interval between interference fringes becomes too dense and analysis becomes impossible. For this reason, measurement must be performed with a high-magnification objective lens, and as a result, there has been a problem that the measurement range in a single measurement is very narrow, not more than 1 mm square.

On the other hand, for example, an area to be measured is
Attempts have also been made for a method of dividing the data into a plurality of partial regions having mutually overlapping regions, measuring the divided regions with an interferometer, and performing a joining process so that the data of the overlap regions match. However, when the joining process is performed, the joining process is performed without separating the low-frequency shape component having a low S / N and the high-frequency shape component having a high S / N. In some cases, there was a disadvantage that the joining accuracy was not sufficiently obtained.

In the method disclosed in Japanese Patent Application Laid-Open No. H10-160428, data is moved only in the normal direction of the overlap area, and in other directions, the position and orientation of the interferometer at the time of measuring the surface shape. Assuming that the measured data is correct, the degree of freedom of the joining process is reduced to compensate for the lack of accuracy when the joining process is performed by relying only on the shape data. However, it has been difficult to achieve an accuracy on the order of nanometers due to the influence of measurement errors included in the measurement data of the position and orientation of the interferometer.

[0007]

According to the present invention, a surface to be measured having a three-dimensional shape is divided into a plurality of partial regions having mutually overlapping regions, and the measured surface shape data is included in the data. An object of the present invention is to obtain a sufficient accuracy for evaluating a range of several mm square on the order of nanometers by adopting a configuration for performing a joining process in consideration of a frequency component. Specifically, low-frequency shape components are removed from the surface shape data divided into a plurality of partial regions, only high-frequency components having a high S / N ratio are extracted, and joining is performed based on the extracted high-frequency components. The object is to make it possible to evaluate the surface roughness in the range of several mm square with an accuracy of nanometer. It is another object of the present invention to perform joining in consideration of not only high-frequency components but also low-frequency components, and obtain overall shape data of a surface to be measured while maintaining accuracy on the order of nanometers. Another object of the present invention is to improve the calculation speed and accuracy when performing the joining process. Specifically, high-speed calculation and improvement in processing accuracy are achieved by performing rough joining processing before joining by high-frequency shape components, and sufficient shape measurement accuracy for joining processing is ensured. It is an object of the present invention to realize a possible arithmetic processing method and to increase the speed and accuracy of the joining process by reducing the total number of surface shape data to be joined. It is another object of the present invention to form alignment marks on the surface to be measured to further improve the joining processing accuracy.

[0008]

[Measures taken to solve the problem]

[Means 1] Means taken for solving the above problems are a means for dividing a surface to be measured having a three-dimensional shape into a plurality of partial regions having mutually overlapping regions, and measuring the surface shape; The low-frequency shape component is removed from the shape data to extract only the high-frequency shape component, and the high-frequency shape data is relatively moved and connected so that the high-frequency shape components corresponding to the overlap area most closely match each other. By calculating to match, it is possible to measure with nanometer accuracy.

[0009]

Another means taken for solving the above problem is a means for dividing a surface to be measured having a three-dimensional shape into a plurality of partial regions having mutually overlapping regions and measuring the surface shape. And arithmetic means, and storage means, using the arithmetic means, removes low-frequency shape components from the surface shape data to extract only high-frequency shape components, and stores the surface shape in the storage means. The low-frequency component removed from the data is temporarily stored, and using the arithmetic means, a high-frequency shape component corresponding to the overlap region is subjected to a joining process between the high-frequency shape data so as to most closely match each other, In the storage means, temporarily store the relative movement amount performed at the time of the joining processing of the high-frequency shape data, using a calculating means, the temporarily stored relative movement, the temporarily stored low-period waveform After applying to the data, this is added again to the high-frequency shape data, and for the data to which the high-frequency shape data has been added again, under the condition that the in-plane coordinate value of each measurement point in the overlap region is not changed. By performing the joining process again, the entire shape data can be accurately reproduced, and the evaluation on the order of nanometers is made possible.

[0010]

Another means taken to solve the above problem is to measure a surface shape by dividing a surface to be measured having a three-dimensional shape into a plurality of partial regions having mutually overlapping regions. Means for measuring at least one of the relative attitude and the relative translation amount of each surface shape data, and each surface shape data are roughly connected in advance based on the relative posture and / or the relative translation amount, and this is initially initialized. State, and then using the high-frequency shape component as a clue,
That is, a precise joining process is performed to speed up the arithmetic processing. Further, the shape measuring means may measure a three-dimensional shape by analyzing interference fringe image data. Alternatively, a three-dimensional shape may be measured by scanning in a two-dimensional plane using a stylus or a non-contact displacement meter as the shape measuring means.

[0011]

Another means taken for solving the above problem is to use a light source having a continuous spectrum as a light source for forming an interference fringe image, thereby enabling a wide range of shape measurement at a time. As a result, the total number of surface shape data is reduced.

[0012]

Another means taken to solve the above problem is that, for each measurement point in the surface shape data, an average value of adjacent data within a certain range around the measurement point is used. Calculated, the average value is subtracted from the measurement points, and the averaging process and the subtraction process are performed for all the measurement points to remove low-frequency shape components, thereby improving the calculation speed and accuracy. is there. Alternatively, the surface shape data may be approximated by an m × n-order polynomial, and the low-frequency shape component may be removed by subtracting the approximate polynomial from each measurement point. It is also possible to convert the surface shape data to spatial frequency domain data, remove low-frequency components in the spatial frequency domain, and remove the low-frequency shape components again by inversely converting them to surface shape data. is there.

[0013]

Another means taken to solve the above problem is to form an alignment mark by forming a surface structure in which a predetermined shape is regularly or irregularly arranged on a surface to be measured of an object to be measured. This is to improve the joining processing accuracy. Further, the production cost may be reduced by forming an alignment mark on the surface of the molding die in advance.

[0014]

According to the above construction, since the joining process is performed only with the high-frequency shape component having a high S / N ratio, a high-precision arithmetic process can be performed. Further, the relative movement amount for each partial region when the joining process based on only the high-frequency shape component is performed is stored, and the low-frequency shape component is moved by the same amount by the relative movement amount, and then added to the high-frequency component. Since the joining process is performed again, accurate whole shape data including all frequency shape components can be obtained. In addition, since the general joining process is performed in advance before the joining process using the high-frequency shape component, the calculation can be speeded up and the processing accuracy can be improved. Further, since each surface shape data is measured by scanning with an interferometer or a stylus or a non-contact displacement meter in a two-dimensional plane, highly accurate measurement is possible. In addition, since measurement is performed by an interferometer using the principle of low coherent interference having a light source having a continuous spectrum, a wide range of shape measurement can be performed by one measurement, and as a result, the total number of surface shape data to be joined can be reduced. Can be. In addition, since it is possible to remove low-frequency components using local averaging processing, polynomial approximation, and filtering in the spatial frequency domain, it is possible to perform joining processing using only high-frequency components having a high S / N. it can. In addition, since the alignment mark is formed on the surface to be measured, it can be used as a reference for the position of the object to be measured during shape measurement, and a more accurate joining process can be performed.

[0015]

Next, an embodiment will be described with reference to the drawings. FIG. 1 is a diagram showing a schematic configuration of a shape measuring apparatus of the present invention. 1 is an X-axis stage movable in the X-axis direction, 2 is a Y-axis stage movable in the Y-axis direction, and 3 is a Z-axis stage movable in the Z-axis direction. Furthermore, this shape measuring device
It has a θx axis stage 4 that can rotate around the X axis and a θy axis stage 5 that can rotate around the Y axis. White interferometer 6 internally provided with a plane reference surface as a partial surface measuring means
Is attached to the Z-axis stage 3. The white light interferometer has a feature that height information corresponding to each pixel of a CCD for capturing interference fringes can be directly obtained without information on adjacent pixels (for example, Japanese Patent Application Laid-Open No. 8-502829).
No.). Therefore, as compared with a laser interferometer that requires information on adjacent pixels when reconstructing a shape, the shape can be measured even when the interval between interference fringes is small, so that a wider measurement range can be measured when measuring a curved surface shape. Can be secured. On the other hand, the DUT 7 is mounted on the Y-axis stage 2 and the rotation stages 4 and 5 and positioned so as to face the white interferometer 6. The rectilinear stages 1, 2 and 3 and the rotary stages 4 and 5 are controlled by using a control device (not shown). The relative position and orientation of the interferometer 6 and the DUT 7 are determined.

FIG. 2 is a view of the surface of the device under test 7 viewed from the Z-axis direction, showing how the surface to be measured is divided into partial areas. Each of the partial areas is divided and set so as to generate an overlapping area, and thereafter, measurement is performed for each of the divided areas.

Next, the flow of the shape measuring method according to the present embodiment will be described with reference to the flowchart shown in FIG. First, the device under test 7 is set on the rotating stage θy (step S1), and the surface to be measured is divided into partial regions of an appropriate size (step S2). Next, the interferometer is positioned with respect to each partial area based on the information on the design shape of the device under test (step S3), and the positioning information at that time is stored in a storage device (step S4). The shape is measured (Step S5). When all the partial areas have been measured (Step S6), the measurement data of each partial area is relatively moved based on the positioning information stored in Step S4 (Step S7), and the general joining is completed. Next, adjacent partial areas having mutually overlapping areas are selected, and coordinate conversion processing for joining one partial area to another partial area is performed according to the flow described below. First, a low-frequency shape component is separately extracted from each of the selected two partial regions (step S8) and stored in a storage device (step S9), and then the low-frequency shape component is subtracted from the original measurement data. Thus, high-frequency shape components are extracted (step S10). In general, in the case of a mirror-like surface such as a lens surface, the joining accuracy is higher when the joining process is performed after extracting only the high-frequency components than when the joining process is performed while including the low-frequency components. Therefore, in step S11, the high-frequency components extracted in step S10 are connected to each other in the shared overlap region (step S11), and in step 12, the relative movement amount at that time is stored in the storage device (step S1).
2). Next, the low frequency shape component stored in the storage device in step 9 is subjected to coordinate transformation corresponding to the relative movement amount stored in step S12 (step S13), and is added to the high frequency component (step S14). Then, the joining process is performed again (step S15). However, the joining performed in step S15 optimizes only the θx, θy, and Z coordinates so as not to change the in-plane coordinate values of the common overlap area, that is, the X and Y coordinates. When all the partial regions have been connected (step S16), an analysis process is performed on the connected shape data (step S17), and an analysis result such as a shape error is output (step S18).

Next, the low frequency component extraction method shown in step S8 of FIG. 3 will be described with some examples. The first is an extraction method using a local averaging process. This is to calculate a local average value between adjacent data and use it as a low-frequency shape component. Specifically, for a specific data point, an m × n matrix area is defined around that point, and m This is a method in which the average value of × n is used as the low-frequency shape component of the center data point. The second is an extraction method based on polynomial approximation. In this method, measured surface shape data is approximated to an m × n-order polynomial, and is used as a low-frequency shape component. The third is an extraction method using filter processing in the spatial frequency domain. This is a method in which the frequency analysis of the measured surface shape data is performed, high-frequency components are removed in the spatial frequency domain, and the result of inverse conversion to the surface shape data again is used as a low-frequency shape component. When actually performing the process of removing the low-frequency shape component, the measured surface shape data is converted into spatial frequency domain data, the low-frequency component is removed from the data, and the inverse conversion to the surface shape data is performed again. If it is performed, it becomes the surface shape data of only the high frequency shape component.

The above is one embodiment of the present invention, but the present invention is not limited to this. For example, in the embodiment of FIG. 1, a white light interferometer is used as the shape measuring means, but an interferometer using monochromatic light such as a laser as a light source may be used. In addition, a probe such as a mechanical stylus or an optical displacement meter equipped with a high-speed scanning head for measuring a fine shape, which is suitable for scanning a relatively narrow area at high speed, is used for a white interferometer. It is also possible to use it instead.

Further, in order to perform the joining process with high accuracy, there is a method of making an alignment mark on the surface to be measured. For example, when a plastic lens is assumed as the object to be measured, it is conceivable to form a regular or irregular arrangement of convex or concave shapes on the surface and use it as an alignment mark. In order not to degrade the original optical performance of the lens, it is desirable that the size of the alignment mark has a diameter, depth or height of several tens μm or less, and preferably several μm or less. Although the alignment mark may be formed directly on the surface of the molded lens, it is advantageous in terms of production cost to form the alignment mark on the surface of the molding die in advance.

[0021]

According to the present invention, since the joining process is performed using only the high-frequency components of the measured surface shape data, that is, the joining process is performed so that the surface roughness components are most overlapped, the range of several mm square is obtained. It is possible to evaluate the surface roughness with nanometer accuracy. In addition, since the high-frequency components are joined so as to overlap the most, the low-frequency components are added and the joining process is performed again, so the joining process accuracy is improved, and the entire shape data including the low-frequency components can be accurately analyzed. Can be reproduced well. Also, the relative attitude of each surface shape data,
Since the relative translation amount is measured and the rough joining is first performed based on this information, the processing speed is increased and the joining accuracy is also improved at the same time. In addition, since each surface shape data is measured by scanning an interferometer or a stylus or a non-contact displacement meter in a two-dimensional plane, high-precision measurement is possible and the accuracy of the joining process is improved. In addition, by measuring with an interferometer based on the principle of low coherent interference, it is possible to measure a wide range of shapes at once, so the total number of surface shape data to be joined is reduced, and the accuracy and calculation speed of the joining process Is improved. Also, low-frequency components are removed by using local averaging, polynomial approximation, and filtering in the spatial frequency domain.
The joining process can be performed only with the high-frequency component having a high S / N, and the processing accuracy is improved. In addition, when forming a regular or irregular minute fine shape on the surface of a lens or a lens molding die that does not substantially affect the original optical performance, and performing the joining process. Since the alignment mark is used, the joining processing accuracy is further improved.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a shape measuring device.

FIG. 2 is a state diagram in which a surface to be measured is divided into a plurality of partial regions.

FIG. 3 is a flowchart illustrating a shape measuring method. 1 to 3... X-axis stage 2... Y-axis stage 3... Z-axis stage 4. ... Θy-axis stage 6... White interferometer 7.

Continued on the front page F term (reference) 2F065 AA45 AA53 BB05 BB27 CC22 FF01 FF04 FF52 GG04 GG24 JJ03 JJ26 PP12 QQ00 QQ17 QQ23 QQ25 QQ27 QQ33 QQ42 TT02 2F069 AA51 AA66 GG01 GG04 GG07 GG07 GG07 GG04 BB62 BB74 FF05 HH00

Claims (14)

[Claims] 1. A means for measuring a surface shape by dividing a surface to be measured having a three-dimensional shape into a plurality of partial regions having mutually overlapping regions, and removing a low-frequency shape component from the surface shape data. And computing means for relatively moving and joining the high-frequency shape data so that the high-frequency shape components corresponding to the overlap region most closely match each other. Characteristic shape measuring device. 2. The apparatus according to claim 1, further comprising: means for dividing a surface to be measured having a three-dimensional shape into a plurality of partial regions having mutually overlapping regions to measure the surface shape; calculating means; Means for removing low-frequency shape components from the surface shape data to extract only high-frequency shape components; temporarily storing the low-frequency components removed from the surface shape data in the storage means; By using the means, so that the high-frequency shape components corresponding to the overlap region best match each other,
Performing a joining process between the high-frequency shape data; temporarily storing, in the storage unit, a relative movement amount performed during the joining process between the high-frequency shape data; Is performed on the temporarily stored low-frequency shape data,
This is again added to the high-frequency shape data, and the data to which the high-frequency shape data has been added again are subjected to the joining process again under the condition that the in-plane coordinate value of each measurement point in the overlap region is not changed. The shape measuring device according to claim 1, wherein: 3. A surface to be measured having a three-dimensional shape is divided into a plurality of partial regions having mutually overlapping regions to measure a surface shape, and at least a relative attitude and a relative translation amount of each surface shape data are measured. Means for measuring one of the above, the respective surface shape data based on the relative attitude or relative translation amount, or both, are roughly connected in advance to an initial state, and then, using the high-frequency shape component as a clue, The shape measuring apparatus according to claim 1, wherein a joining process is performed. 4. The shape measuring apparatus according to claim 1, wherein said shape measuring means measures a three-dimensional shape by analyzing interference fringe image data. 5. A light source for forming an interference fringe image,
5. The shape measuring apparatus according to claim 4, wherein a light source having a continuous spectrum is used. 6. The apparatus according to claim 1, wherein said shape measuring means measures a three-dimensional shape by scanning a stylus or a non-contact displacement meter in a two-dimensional plane. 3. Shape measuring device. 7. An average value of adjacent data within a specific range around the measurement point is obtained for each measurement point in the surface shape data, and the average value is subtracted from the measurement point. The averaging process and the subtraction process are performed on all measurement points to remove low-frequency shape components. 6. Shape measuring device. 8. The method according to claim 1, wherein the surface shape data is approximated by an m × n-order polynomial, and a low-frequency shape component is removed by subtracting the approximate polynomial from each measurement point.
The shape measuring apparatus according to claim 1, 2, 3, 4, 5, or 6. 9. A method of converting surface shape data to spatial frequency domain data, removing low-frequency components in the spatial frequency domain, and removing the low-frequency shape components by inversely converting the data back to surface shape data. The shape measuring device according to claim 1, 2, 3, 4, 5, or 6. 10. An object to be measured by any one of the shape measuring devices according to claim 1, wherein a predetermined shape is regularly or irregularly arranged on a surface to be measured of the object to be measured. An object to be measured, characterized by forming an aligned mark by forming a bent surface structure. 11. A molding die for forming the object to be measured according to claim 10. 12. The molding die according to claim 10, wherein the object to be measured is a lens. 13. A surface to be measured having a three-dimensional shape is divided into a plurality of partial regions having mutually overlapping regions and the surface shape is measured, and a low-frequency shape component is obtained from the surface shape data. Using a computing means to remove only high-frequency shape components, temporarily storing low-frequency components removed from the surface shape data in a storage means, using the computing means to remove a high-frequency component corresponding to the overlap region So that the shape components of
Performing a joining process between the high-frequency shape data, temporarily storing the relative movement amount performed during the joining process between the high-frequency shape data in the storage unit, and using the arithmetic unit, the temporarily stored relative movement. Is performed on the temporarily stored low-frequency shape data,
This is again added to the high-frequency shape data, and the data to which the high-frequency shape data has been added again are subjected to a joining process again under the condition that the in-plane coordinate value of each measurement point in the overlap region is not changed. A method for measuring a three-dimensional shape. 14. A surface to be measured having a three-dimensional shape is divided into a plurality of partial regions having mutually overlapping regions to measure a surface shape, and at least a relative attitude and a relative translation amount of each surface shape data are measured. One surface is measured, and the respective surface shape data are roughly connected in advance based on the relative posture or the relative translation amount, or both, to an initial state, and thereafter, using the high-frequency shape component as a clue, precise connection is performed. 2. A combination process is performed.
3. Shape measuring method.