JP2007071663A - Automatic measuring processor of minute displacement of one point in diffusing surface - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically measure the irregularity and vertical displacement of fabricating surface of a product during fabrication which is not a mirror surface but a diffusing surface of an object to be measured by non-contact, with high precision and in real time. <P>SOLUTION: The interference pattern on concentric circles caused by reflection light from the measuring surface, that is, the diffusing surface and a reference surface is read with an optical device. The moving quantity of sinking in circle center direction or floating from the circle center is calculated and the displacement is given to the reference surface in the direction vertical to the surface so as to cancel the moving quantity. As the displaced quantity is the same as the displacement of the measuring surface in the direction vertical to the surface, the displaced quantity is shown as the displacement of the measuring surface in the direction vertical to the surface. In this manner, automatic measurement becomes possible. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

The present invention relates to a technique for automatically measuring a vertical displacement of a surface in a non-contact manner.

According to Non-Patent Document 1, it is possible to obtain reflected light from the diffusing surface as if it was reflected from the mirror surface by reducing the spot diameter of the irradiated laser even if it is a diffusing surface. Interference fringes are formed by interference between the reflected light from the light and the reference light, and the displacement in the direction perpendicular to the surface of the measurement surface is displayed as subtraction or upwelling movement of the interference fringes on the concentric circle. A method for calculating the displacement of the measurement surface from the amount of movement has already been proposed.
Laser Research, Vol. 32, pp 538-542, 2004

  In response to the demand to automatically measure the vertical displacement of the processed surface of the product being processed in a non-contact manner with high accuracy, in the micro displacement measurement using a laser, the measured object is a mirror surface to obtain regular reflected light. Although it is necessary, in the measurement of the actual product, the condition that the object to be measured is a mirror surface is very few, and it is often a diffusion surface.

  Displacement can be measured without contact on the diffusion surface, but at this stage, it is necessary to visually measure the movement distance of the interference fringes and calculate the displacement of the measurement surface against the calibration data prepared in advance. is there. This is a cumbersome operation, and it is necessary to enable automatic measurement in order to eliminate this.

  The above problem is solved by reading concentric interference fringes generated by reflected light from the diffusing surface, which is the measurement surface, and the reference surface with an optical device, and subtracting or rising from the center of the circle on this image data. The movement amount is calculated and the reference surface is displaced in the direction perpendicular to the plane so as to cancel this. The amount of displacement is the same as the displacement in the direction perpendicular to the surface of the measurement surface. It is shown as the displacement of the surface perpendicular to the surface. This allows automatic measurement.

  In the part that calculates the sinking of the concentric interference fringes toward the center of the circle or the movement of upwelling from the center of the circle, the approximate circle center coordinates are visually given in advance from the concentric interference fringes on the screen. An intersection with the arc in the Y direction is obtained from an arbitrary point in the arc on the center X coordinate, and a point having the maximum radius is obtained by scanning in the X direction. This midpoint becomes a new circle center coordinate.

  After obtaining the center of the circle instead of the Hough transform in this way, the arc is recognized again on the image data, the amount of movement is calculated in real time, and the displacement of the reference surface that offsets the displacement of the measurement surface is given. Drive signal is calculated and applied to the drive device. If the displacement given to the reference plane at this time is displayed, the measurement is automated.

  It becomes possible to automatically measure the unevenness and surface vertical displacement of the processed surface of the product being processed, which is a diffusion surface that is not a mirror surface, in real time with high accuracy without contact. This greatly changes the processing and inspection methods.

  Embodiments of the present invention will be described with reference to the drawings. Each figure is not drawn to an accurate scale, and there are parts exaggerated to make the drawing easier to see.

1. Automation Algorithm FIG. 1 is a diagram showing an example of a system configuration for performing automatic measurement according to this embodiment.
FIG. 2 is a photograph of concentric interference fringes formed by reflected light from the measurement surface, which is a diffusing surface, and the reference surface. FIG. 3 is a flowchart showing operations necessary for performing automatic measurement.

  As shown in FIG. 1, in extracting circle information, an interference fringe image captured from a CCD camera is binarized by binarization processing, and then a circle is detected by scanning pixels vertically from the center of the circle. The circle information extracted from the image before the measurement surface is displaced is determined as a reference, and the displacement of the circle information can be detected by comparing the circle information with the image after the displacement.

  Based on the detected displacement of the circle information, a PZT drive voltage is applied to the PZT. The PZT is attached to the reference surface and moves the reference surface in the optical axis direction. Compared with the reference image, the displacement of the measurement surface can be determined by automatically controlling the applied voltage of PZT on the reference surface until there is no difference between the movement of the measurement surface and the difference. The reference plane will always move in real time to compensate for the movement of the measurement plane. It is possible to detect a displacement that is applied to the PZT from the extraction process of the circle information and the displacement distance within the time until the correction of the reference surface distance is completed once is smaller than ½ wavelength (± ¼ wavelength). If this loop is repeated, the measurement range is theoretically unlimited.

2. Radial extraction algorithm diagram of concentric center coordinates and its concentric circle 4 is a diagram showing how to determine the radius calculation of first circle and the second circle that the circle center coordinates of concentric interference fringes.

  In general, the Hough transform is often used as a method for extracting circle information. However, in the basic Hough transform, when a screen with N dots in the vertical and horizontal directions is handled, the number of processes is in the order of the fourth power of N. Since it increases, it is difficult to increase the processing speed. Another Hough transform can reduce the processing time. Nevertheless, since it takes 0.8 to 288 seconds to detect a circle with an image size of 256 × 256, it does not fundamentally reduce the processing time.

  In the present invention, a technique is used in which interference fringes are swept up and down from a certain center point, are scanned in the vertical direction from the center, and circle information of the valley portion is detected. Since the method itself is simple scanning, circle information can be extracted by scanning only the vertical size of the image at the maximum.

  This circle detection method uses the property of interference fringes, that is, the nature of fringing up and down around a certain point. Since it is unlikely that the center position of the interference fringes will be shifted due to the measurement of the same measurement object, the center position of the fringes is determined in advance by visual observation. Then, scanning is started in the vertical direction from the input center point. Since the valleys of the interference fringes of the binary image are black pixels and the valleys have a width, the point where the black pixels appear continuously in the vertical direction is considered as a contact point with the circle. The number of points that appear continuously in the vertical direction is given an optimum value for detecting a circle. Here, the smallest circle is the first circle, and the second smallest circle is the second circle. The detection of the second circle is obtained by scanning further outward from the contact point of the circle detected in the first circle.

  In FIG. 4, since it is expected that the center point is visually deviated from the original center point, the scanning range is expanded in the X direction from the input center point, and two intersection points with the circle are obtained. And the middle point of the line that maximizes the distance between the intersections is the new center coordinate of the circle.

  After the central coordinates are determined, the radius of the first circle is obtained from the intersection of the first circle on the investigation line, the scan is further extended from the intersection, the intersection with the second circle is obtained, and the radius of the second circle is obtained from this. .

  In the present invention, detection of minute displacement is performed by performing image processing on the interference fringe image, but as can be seen from the image of FIG. 2 which is the interference fringe, the interval between the bright and dark fringes near the center of the circle is larger. , Clearly reflected. Therefore, in image processing, circle information near the center of the circle is used. Instead of using the interference fringes themselves, the dark (valley) portions of the fringes are used. The reason why the fringes themselves are not used for image processing is that the brightness of the interference fringes is not stable. Although it depends largely on the shooting situation, in the experiment, the circle information can be obtained stably from the first valley and the second valley from the center of the interference fringes.

3. Implementation Apparatus FIG. 5 shows an experimental apparatus for automatic measurement. In FIG. 1, a screen is arranged at a place where interference occurs, and a bright and dark pattern of interference fringes is projected and taken into a personal computer by a CCD camera. A screen is not always necessary, but when a screen is not used, measurement using a photoelectric conversion element is required. By using a digital image processing system, the temporal response is somewhat degraded, but two-dimensional changes in concentric interference fringes can be followed.

  An image taken into the personal computer from the CCD camera is extracted with concentric circle information by image processing, then compared with a reference image, and an applied voltage is applied to the reference plane drive PZT so as to correct the difference. The applied voltage is controlled based on a graph based on hysteresis characteristics. In this experiment, the experiment was performed using a diffusion surface driven in the direction perpendicular to the surface by PZT as a measurement surface whose displacement is known.

  The optical system used in the experiment uses a He—Ne gas laser (wavelength: 633 nm) as a light source and expands it into a parallel beam having a diameter of 20 mm by a beam expander. Thereafter, the beam is converged by a projecting lens having a focal length f = 150 mm. The spot diameter at the focal point is 6.04 μm. The resolution of the CCD camera is 640 × 640.

4). Operation Result FIG. 6 shows an example of a screen when automatic measurement is activated. FIG. 7 shows the result of binarization using threshold processing for the grayscale concentric interference fringes of FIG. FIG. 8 is a graph showing the reference plane driving PZT applied voltage and the reference plane vertical displacement. FIG. 9 is a graph showing the PZT applied voltage for driving the measurement surface and the reference surface vertical displacement. Table 1 shows an example of calculating the center coordinates of the concentric interference fringes and the radii of the first and second circles. Table 2 shows an example of the result of automatic measurement with negative feedback.

  As shown in FIG. 6, when the automatic measurement is operated, the screen displays the approximate center coordinate display inputted in advance by looking at the concentric interference fringes and the first interference fringes before the measurement surface is displaced. The calculated first and second circle radii (reference values in FIG. 6), and the first and second circle radii calculated for the interference fringes after the measurement surface is displaced (in FIG. 6). (Current value), the voltage applied to PZT to move the reference plane in the vertical direction at that time, and the displacement calculated from the voltage using FIG. 9 are shown. The radii of the first circle and the second circle are theoretically the same because negative feedback is applied before and after the measurement surface is displaced. Therefore, the displacement of the measurement surface is equal to the displacement of the reference surface.

  Table 1 shows, from the concentric circles binarized as shown in FIG. 4, the approximate center coordinates of the first and second circles given visually, the calculated center coordinates, and the calculated first coordinates. It is a result showing the radius of 1 circle and 2 circle

  Since the voltage applied to PZT on the reference surface during negative feedback is calculated in the program, the displacement distance of the reference surface is calculated from this drive voltage. The relationship between the driving voltage and the displacement distance of the reference surface is obtained in advance as shown in FIG. 8 because the distance by which one circular arc sinks or rises corresponds to 1/2 wavelength. Similarly, the measurement PZT that gives a pseudo displacement of an object is also obtained as shown in FIG. Although PZT generally has hysteresis as shown in FIG. 8, it is not necessary for the measurement surface and is not described here.

  As shown in Table 2, the detection result actually measured with negative feedback is 10.9671 [nm]. This is the value displayed after it has been confirmed that negative feedback is applied and the first or second concentric circle moves and overlaps. Therefore, it is considered that automatic measurement has been performed.

  On the other hand, the displacement calculated from the voltage applied to the measurement surface driving PZT using FIG. 9 was 61.7 nm. A large error occurs at the time of automatic measurement. For example, since the voltage applied to the reference surface increases in time series, that is, the movement is not in a single direction, when calculating the displacement from the voltage of the reference surface driving PZT, It is considered that an appropriate value is obtained within the hysteresis range shown in FIG. In order to solve this, it is necessary to use a reference surface driving element having no hysteresis.

  In addition, when a minute displacement is measured without driving the PZT on the measurement surface, the size of the interference fringe after the displacement should be the same as the reference image. Therefore, the reference surface driving PZT operates, and the displacement amount is detected to be about plus or minus 5 [nm]. This is thought to be due to room temperature and vibration.

  In the present invention, the automation of minute displacement measurement is realized for one point on the diffusion surface using a laser interferometer. Since this automatic measuring device does not use Hough transform for extracting concentric laser interference fringes, we were able to construct a high-speed system.

  FIG. 7 shows the result of binarizing the gray scale concentric interference fringes shown in FIG. A good binary image has been obtained by the binarization process, and circular information detection from the binary image has also obtained good results. The processing time from the start of binarization processing to the completion of circle information detection is measured with the GetTickCount () function, but is detected around 20 ms, and the processing target image size is larger than the Hough transform method. However, the processing speed has been shortened, and high speed has been realized.

  The ability to automatically measure the vertical displacement of the product and during processing makes it possible to incorporate product measurement and processing inspection into the production process, resulting in labor savings.

It is a figure showing the example of a system configuration for performing automatic measurement. It is a photograph of concentric interference fringes. It is a flowchart figure for performing automatic measurement. It is a figure showing how to obtain the circle center coordinates of concentric interference fringes and the radius calculation of the first circle and the second circle. It is a photograph when carrying out automatic measurement. It is a display example when the radius calculation and measurement surface displacement calculation of the first circle and the second circle of the concentric interference fringes during the automatic measurement operation are performed. It is the figure which made the binarization example of the concentric interference fringe. It is the graph showing PZT application voltage for reference surface drive, and a reference surface displacement. It is a graph showing PZT applied voltage for measurement surface drive and a reference surface displacement.

Explanation of symbols

1 Measurement surface 2 Reference surface 3 PZT
4 Beam splitter 5 Lens 6 Screen 7 CCD camera 8 Laser beam 9 Computer 10 Laser expander 11 Laser transmitter

Claims (2)

In a device that measures the displacement of the diffusion surface in the direction perpendicular to the surface
Interference fringes caused by reflected light from the diffusing surface and the reference surface;
An optical device that reads this,
This image is processed to calculate the subtraction of concentric fringes on the concentric circle caused by displacement of the diffusion surface or the movement of upwelling from the center of the circle and the movement of the interference fringes in the direction perpendicular to the reference plane. A portion for calculating the vertical displacement of the diffusion surface from the displacement applied to the reference surface in the direction perpendicular to the surface to be canceled by the displacement of
With
Measurement processing apparatus characterized by In the part for calculating the sinking of the concentric interference fringes according to claim 1 toward the center of the circle or the movement of upwelling from the center of the circle,
The approximate circle center coordinates are given in advance from the concentric interference fringes on the screen, and the intersection with the arc in the Y direction is obtained from an arbitrary point in the arc on the X coordinate of the circle center. The radius is maximized. Is obtained by scanning in the X direction. This midpoint becomes a new circle center coordinate. In this way, instead of the Hough transform, find the center of the circle,
A measurement processing apparatus characterized by that.