JP2005164294A - Stripe image measurement data synthesis method using evaluation value - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a stripe image measurement data synthesis method which can obtain the result having high accuracy, by the synthesis using the data in which the measurement is performed at the best accuracy, from the plurality of measurement data pieces. <P>SOLUTION: The method includes: a step for obtaining a plurality of phase distribution data pieces by performing phase analysis of a plurality of phase shifted stripe image data pieces; a step for determining the evaluation value for each pixel, from the associated stripe image data with respect to each phase distribution data piece; and a step for the synthesis by selecting, for each pixel, one having the high evaluation value from the plurality of phase distribution data pieces. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a fringe image measurement data combining method for combining a plurality of fringe image measurement data.

  For non-contact shape measurement equipment using the grid projection method, non-contact stress / strain measurement equipment using interference fringe analysis, displacement measurement equipment using digital holography, etc., combining multiple fringe image measurement data obtained by changing measurement conditions Is expected to give better results.

  However, in the conventional fringe image analysis method, for example, when combining a plurality of fringe image measurement data obtained by changing the measurement conditions, the average value of these data is the result. In other words, there is a problem in that it is impossible to determine for each pixel a point where measurement can be accurately performed and a place where there is noise.

  As a conventional technique for solving this problem, Japanese Patent Application Laid-Open No. 2001-241930 “Method for analyzing fringe images” divides a region to be analyzed into a plurality of regions, and calculates the boundary between them. Thus, a synthesis method is disclosed. However, this conventional analysis method has a problem that it cannot be combined pixel by pixel.

  It seems that it is possible not only to simply average the data but also to obtain good results by determining the evaluation value for each pixel and adopting data with a high evaluation value. As a conventional technique for evaluating an image, Japanese Patent Application Laid-Open No. 07-325922 “Image Evaluation Method” passes a density distribution of a cross section of an image through a band filter having a characteristic representing the sensitivity of the human eye to a spatial frequency. A method for calculating an evaluation value representing image quality from a gray level distribution of the output of a filter is disclosed. This conventional image evaluation method is characterized in that the image quality of a digital image can be evaluated and displayed using an evaluation value. However, this conventional image evaluation method can be used only for the evaluation of the image quality of a digital image taken by a camera, and there is a problem that the phase data analyzed in the fringe image measurement cannot be evaluated.

  In view of the above, the present invention provides a fringe image measurement data synthesis method capable of obtaining an accurate result by combining data that can be measured with the highest accuracy among a plurality of measurement data. With the goal.

  A fringe image measurement data synthesis method according to the present invention includes a step of obtaining a plurality of phase distribution data by phase analysis of a plurality of phase-shifted fringe image data, and fringe image data related to each of the plurality of phase distribution data. And determining the evaluation value for each pixel, and synthesizing the plurality of phase distribution data by selecting one having a high evaluation value corresponding to each pixel.

  According to another embodiment of the fringe image measurement data synthesis method of the present invention, in the fringe image measurement data synthesis method according to claim 1, the phase distribution data and the evaluation are performed before the step of synthesizing the plurality of phase distribution data. The method further includes resampling the value at the same sampling point.

  According to still another embodiment of the method for synthesizing fringe image measurement data according to the present invention, the evaluation value is determined from a predetermined maximum luminance value even if the luminance of the pixel of the plurality of related phase-shifted fringe images is one. A fringe image measurement data synthesizing method characterized in that it is set to 0 if it is higher or smaller than the minimum luminance value, and if not, it is determined as a value corresponding to the amplitude of the fringe of the fringe image.

  Still another embodiment of the method for synthesizing fringe image measurement data according to the present invention is that a first-order frequency obtained by Fourier-transforming the evaluation value and the luminance of the pixel of the plurality of related phase-shifted fringe images. A fringe image measurement data synthesizing method characterized in that the fringe image measurement data is determined as a ratio between a component and a high frequency component.

  Still another embodiment of the fringe image measurement data synthesis method according to the present invention is obtained by photographing a plurality of interference fringe images obtained by phase-shifting the plurality of phase-shifted fringe image data. It is characterized by being.

  According to the present invention, when there are a plurality of measurement data at the same position, it is possible to synthesize using only the data that is most accurate for each pixel. Further, the evaluation value can be determined by a method according to the measurement method. Furthermore, according to the present invention, the composition when the measurement conditions are changed is simplified.

When a sinusoidal fringe pattern is projected onto the measurement object, the fringe intensity distribution I (x) at the point x is expressed by the following equation (1).

I (x) = a (x) cos [φ (x) + α] + b (x) (1)

Here, a (x) is the stripe contrast distribution, b (x) is the bias component of the stripe intensity distribution, α is the initial phase, and φ (x) is the phase. If this phase φ (x) can be analyzed with high accuracy, the surface shape of the object to be measured can be accurately known. In order to obtain the phase value, generally, a phase shift method or a Fourier transform phase shift that is resistant to noise is used. However, in fringe image measurement, errors tend to occur if the fringe contrast is low. In addition, when the brightness exceeds that which can be photographed by the camera, the error becomes large in the phase shift method. It is necessary to evaluate whether or not the phase analyzed at each measurement point can be measured with high accuracy. Therefore, in the method for synthesizing the fringe image measurement data according to the present invention, an evaluation value is given to the analyzed phase distribution by the following method.

A method for determining the evaluation value will be described. FIG. 1 is a flowchart illustrating such an evaluation value determination method. An evaluation value is determined for each pixel from the phase-shifted luminance data for one period. The luminance values of images obtained continuously by shifting the phase n times are represented by I ₁ , I ₂ ,. . . , When I _n, the evaluation value may be defined by the evaluation function as the following equation (2).

E (i, j) = f {I ₁ (i, j), I ₂ (i, j),. . . , I _n (i, j)}
(2)

Here, (i, j) represents the horizontal i-th and vertical j-th pixels of the camera to be photographed. E (i, j) is an evaluation value.

ここで、ａ（ｉ，ｊ）は投影した格子の振幅、Ｉ_ｎ（ｉ，ｊ）は投影した格子の明るさである。Ｉ_ｍａｘとＩ_ｍｉｎは、それぞれカメラが撮影できる最高輝度値と最低輝度値とを表す。撮影されるｎ枚の輝度データのうち、１つでもＩ_ｍｉｎより低い輝度値、又はＩ_ｍａｘより高い輝度値がある場合、評価値を０とする。そうでない場合、評価値を縞の振幅とする。この例の場合、コントラストが良いと評価値が高くなる。合成する際には、評価値の高い方を選ぶ。 The first example of the evaluation function that determines the evaluation value by the amplitude of the fringes is defined as follows.

Here, a (i, j) is the amplitude of the projected grating, and I _n (i, j) is the brightness of the projected grating. I _max and I _min represent a maximum luminance value and a minimum luminance value that can be captured by the camera, respectively. If at least one of the n pieces of luminance data to be photographed has a luminance value lower than I _min or a luminance value higher than I _max , the evaluation value is set to 0. Otherwise, the evaluation value is the amplitude of the stripe. In this example, the evaluation value increases when the contrast is good. When combining, select the one with the highest evaluation value.

ここで、Ｆ（１）と

は、それぞれ周波数１の成分と高次周波数成分である。この例の場合、評価値が高いものはノイズに強い（ＳＮ比が高い）ということになる。合成する際には、評価値の高い方を選ぶ。 The second example of the evaluation function that determines the evaluation value by the ratio between the primary frequency and the high-order frequency of the fringes is defined as follows.

Where F (1) and

Are a frequency 1 component and a higher-order frequency component, respectively. In this example, a high evaluation value is resistant to noise (high SN ratio). When combining, select the one with the highest evaluation value.

  The fringe image measurement data synthesizing method of the present invention synthesizes phase distribution data obtained from a plurality of fringe image measurement data based on the evaluation values as described above. FIG. 2 is a flowchart for explaining an example of a synthesis procedure using evaluation values by the fringe image measurement data synthesis method according to the present invention. In this example, the measurement is performed twice, but more measurements may be performed. Luminance data (stripe image measurement data) 1 and 2 is acquired by changing measurement conditions, and phase analysis is performed on each luminance data to obtain phase distribution data. In each of the luminance data, an evaluation value is obtained for each pixel by the method described above. For each pixel, the higher of the corresponding evaluation values of the two phase distribution data is selected as a synthesis result.

  When combining measurement data measured in different directions, re-sampling is performed so that the xy coordinates are equidistant before combining. FIG. 3 is a flowchart for explaining another example of the synthesis procedure by the fringe image measurement data synthesis method according to the present invention when such re-sampling is performed. In this example, the measurement is performed twice, but more measurements may be performed. Luminance data 1 and 2 are acquired by changing the measurement conditions, and phase analysis is performed on each luminance data to obtain phase distribution data. In each of the luminance data, an evaluation value is obtained for each pixel by the method described above. Next, each phase distribution data and evaluation value are resampled at the same equal interval. FIG. 4 is a diagram for explaining an algorithm for resampling such phase distribution data. a shows the measurement points of the phase distribution data before re-sampling, finds the three measurement points Pa, Pb, Pc closest to the sampling point Ps selected so that the xy coordinates are equidistant on the xy plane, The intersection of the plane passing through these three points and the straight line extending perpendicularly in the z direction from the xy coordinates of the sampling point Ps is taken as the z coordinate of the sampling point. FIG. 4b is a diagram showing the measurement points after re-sampling in this way. Similar re-sampling is performed for the evaluation value. Returning to FIG. 3, for each pixel, the higher evaluation value after the corresponding re-sampling of the two phase distribution data is selected as a synthesis result.

  An example in which the above-described fringe image measurement data synthesis method according to the present invention is used for actual object measurement will be described. For example, when measuring metal parts or the like, there is an influence due to specular reflection. According to the fringe image measurement data synthesis method according to the present invention, measurement is performed from two different directions using two cameras, and the measurement data in these two directions is synthesized. In these two-direction measurement data, the place where specular reflection occurs varies, so by adopting the data that can be measured accurately out of the two-direction measurement data for each pixel, the effect of specular reflection is consequently obtained. Can be avoided. FIG. 5 is a diagram illustrating an optical system that performs such measurement. Light is projected onto the object 1 from the projector 2, the object 1 is photographed from different directions by the cameras 3 and 4, and each luminance data obtained is subjected to phase analysis by the computer 5 to obtain a height distribution (phase distribution). An evaluation value is calculated from the luminance data, the height distribution and the evaluation value are resampled, and a height distribution is synthesized based on the evaluation value.

  FIG. 6 shows the actually measured object 1, where a is a photograph of the object 1 and b is a perspective view showing the dimensions of each part of the object 1. Such an object 1 was measured according to the procedure described with reference to FIG. FIG. 7 shows the results obtained in this way. A and c are heights obtained by phase analysis of the luminance data obtained by photographing with the cameras 3 and 4 by the computer 5 and re-sampling, respectively. B and d are evaluation value distributions obtained from the luminance data corresponding to a and b, respectively, and f is an evaluation value used for composition obtained by selecting the higher evaluation value for each pixel. E is a height distribution obtained by combining the height distributions of b and d based on the evaluation value distribution of f. FIG. 8 is a graph showing cross-sectional data taken along the line A-A ′ of FIG. a and b are height distributions obtained from the cameras 3 and 4, respectively, and c is a combined height distribution. The part which was not measured by the influence of the specular reflection and shadow seen in a and b is not seen in c. Therefore, it can be seen that the effect of specular reflection could be removed by the fringe image measurement data synthesis method according to the present invention.

  Another example in which the fringe image measurement data synthesis method according to the present invention is used for actual object measurement will be described. In the fringe image measurement, since the fringes projected on the object are photographed and analyzed by a camera, generally the shape behind the object cannot be measured. Therefore, the shape of the entire circumference of the object can be measured by combining the results obtained by rotating the object and measuring from several directions. At this time, when there are a plurality of data for the same position, it is possible to measure the shape of the entire circumference with high accuracy by combining them using the fringe image measurement data combining method according to the present invention. FIG. 9 is a diagram for explaining an optical system for measuring the entire circumference. An object 11 is placed on a rotary stage 12, light is projected by a projector 13, and the rotary stage 12 is rotated by 60 °, for example, and photographed by a camera 14, and the obtained luminance data is subjected to phase analysis by a computer 15. Then, a height distribution is obtained, an evaluation value is calculated from the luminance data, the height distribution and the evaluation value are resampled, and a height distribution is synthesized based on the evaluation value.

  First, FIG. 10 shows the result of measurement from one direction. a is a photograph of the object 11 to be measured, b is an image taken by projecting a grating, c is a phase distribution before phase connection, d is a phase distribution after phase connection, and e is resampling. F is an evaluation value distribution.

  FIGS. 11 a to 11 f are diagrams showing height distributions obtained by phase-connecting measurement data obtained by rotating the object 11 by 60 ° and measuring from six directions and re-sampling.

  FIG. 12 shows the result of combining the measurement data in the six directions shown in FIG. a is a height distribution (cylindrical coordinate system) after synthesis, and b is an evaluation value distribution after synthesis. It can be seen that the shape of the entire circumference could be measured with high accuracy by using the fringe image measurement data synthesis method of the present invention.

It is a flowchart explaining an evaluation value determination method. It is a flowchart explaining an example of the procedure of the synthesis | combination by the fringe image measurement data synthesis method by this invention. It is a flowchart explaining the other example of the procedure of the synthesis | combination by the fringe image measurement data synthesis method by this invention. It is a figure explaining the resampling algorithm of phase distribution data, a shows the measurement point of the phase distribution data before resampling, and b shows the measurement point after resampling. It is a figure explaining the optical system at the time of using the fringe image measurement data synthesis method by this invention for actual object measurement. a is a photograph of the object 1 in FIG. 5, and b is a perspective view showing dimensions of each part of the object 1. a and c are height distributions, b and d are evaluation value distributions, f is an evaluation value distribution obtained by selecting a higher evaluation value for each pixel, and e is obtained by synthesis. Height distribution. It is a graph which shows the section data taken along the A-A 'line of Drawing 7e. It is a figure explaining the optical system which performs a perimeter measurement. a is a photograph of an object to be measured, b is an image taken by projecting a grating, c is a phase distribution before phase connection, d is a phase distribution after phase connection, and e is resampled It is a height distribution, and f is an evaluation value distribution. FIGS. 4A to 4F are diagrams showing height distributions obtained by re-sampling by phase-connecting measurement data obtained by rotating the object 11 by 60 ° and measuring from six directions. a is a height distribution (cylindrical coordinate system) after synthesis, and b is an evaluation value distribution after synthesis.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,11 Object 2,13 Projector 3,4,14 Camera 5,15 Computer 12 Rotary stage

Claims (5)

  Analyzing the phase of each of the plurality of phase-shifted fringe image data to obtain a plurality of phase distribution data; and determining an evaluation value for each pixel from the related fringe image data for each of the plurality of phase distribution data; And a step of synthesizing the plurality of phase distribution data by selecting one having a high evaluation value corresponding to each pixel.   2. The fringe image measurement data synthesis method according to claim 1, further comprising the step of resampling the phase distribution data and the evaluation value at the same sampling point before the step of synthesizing the plurality of phase distribution data. The fringe image measurement data synthesis method.   3. The fringe image measurement data synthesis method according to claim 1, wherein the evaluation value is higher than a predetermined maximum luminance value, even if the luminance of the pixel of the plurality of related phase-shifted fringe images is at least one. If it is smaller than the minimum luminance value, 0 is set, and if not, it is determined as a value corresponding to the stripe amplitude of the stripe image.   3. The fringe image measurement data synthesis method according to claim 1, wherein the evaluation value is a first-order frequency component obtained by Fourier-transforming the luminance of the pixel of the plurality of related phase-shifted fringe images. And a fringe image measurement data synthesis method characterized in that it is determined as a ratio of a high frequency component.   5. The fringe image measurement data synthesis method according to claim 1, wherein a plurality of interference fringe images obtained by phase shifting the plurality of phase-shifted fringe image data are captured. 6. A fringe image measurement data synthesis method characterized by being obtained by the above.

Images