KR101011925B1 - An apparatus for measuring three-dimensional shape - Google Patents

Abstract

The present invention provides a technique capable of performing three-dimensional measurement with a short measurement time.

The optical path forming unit 5 receives the broadband light from the broadband light source 1, branches into the reference light path and the measurement light path, and enters the light path forming unit 5 to combine the reflected light from the reference mirror and the reflected light from the object under test to capture the light. ) On the other hand, the optical path varying means 8 changes the optical path length of the measurement optical path. The imaging means acquires interference fringe data including an interference fringe by imaging the output from the optical path forming portion at a timing at which aliasing occurs for the change in the optical path length. The optical path length detecting means 20 is configured to obtain a specific optical path length indicating a characteristic value of the interference fringe, excluding a frequency component generated by aliasing from the interference fringe data acquired by the imaging means.

Description

Three-dimensional shape measuring device {AN APPARATUS FOR MEASURING THREE-DIMENSIONAL SHAPE}

The present invention relates to a three-dimensional shape measuring device for three-dimensionally measuring the shape of an object to be measured using interference phenomena caused by broadband light (for example, white light) having a plurality of spectra (hereinafter, referred to as wavelengths). . In particular, one side of the broadband light is incident on a reference optical path having a reference mirror at a far end, and the other side of the broadband light is incident on a measurement optical path having a measurement object at a far end, and from the reference mirror (reflected mirror) and the object to be measured. In the interference section (interferometer) which generates interference by each returning light, the shape of the object to be measured is measured based on the optical path length in which the interference fringe obtained by changing the optical path length of either the reference optical path or the measurement optical path. A three-dimensional shape measuring apparatus, which relates to a technique for shortening the time for obtaining an optical path length in which an interference fringe occurs.

In general, in the shape measuring apparatus using the above-mentioned interference phenomenon, when the optical path lengths of both the reference optical path and the measurement optical path become the same, the interference fringe shows the maximum luminance. In other words, by changing the optical path length of either the reference optical path or the measurement optical path (hereinafter, the optical path length of the reference optical path is fixed and the optical path length of the measurement optical path is changed). The optical path length (the amount of change in the optical field: hereinafter referred to as the 'specific optical path length') at the position where the maximum luminance is measured is measured as the displacement of the measured object in the direction of change of the optical path length (Patent Document 1).

In patent document 1, based on the interference light obtained by changing an optical path length with time change, it isolates into B component (blue band component), G component (green band component), and R component (red band component). Then, the phase change of the interference fringe is detected with respect to the change in the optical path, respectively, and the optical path length of the portion where the phases of the three characters coincide is recognized as the specific optical path length at the position where the interference fringe shows the maximum luminance. The shape measurement is performed from the recognized specific optical path length.

[Patent Document 1]

Japanese Patent Application 2006-371632

Generally, since the shape measuring apparatus measures a large number of objects to be measured, it is desired to shorten the measurement time even a little. In shortening, there are variable time (or speed) of the optical path length, the imaging time of the camera, the number of imaging times, and the like. However, the imaging time is limited by the minimum exposure time inherent to the imaging device of the camera.

Therefore, below, the prior art is considered from the viewpoint of measurement time. In the case of Patent Literature 1, the phase of the interference fringe is specified on the basis of the data of the interference light, and the specific optical path length is obtained. The data of the (optical field region) is changed by FFT processing, separated into respective band components in the frequency domain, and an interference fringe of each band component is obtained in the time domain again, and the coincidence points of the phases are obtained. At this time, in order to reproduce the interference fringe from the acquired digital data, it is generally performed by sampling theorem and the like at a repetition acquisition timing in which at least three points of data can be acquired in one cycle of the interference fringe to be reproduced. You need to convert it to data.

In general, an interference fringe measured by an interference method in a shape measuring device is represented by a change in luminance with respect to an optical path length changed in the model as shown in FIG. 7 is represented on the coordinate of the frequency band amplitude as shown in FIG. At this time, the envelope width Δt (for example, half-value width: width of the horizontal axis of the portion where the peak luminance value is 1/2) of the interference fringe of FIG. 7 (A) is the frequency of FIG. It is known to have a correlation with the bandwidth ΔF (for example, half width: width of the horizontal axis of the portion where the peak amplitude is 1/2). Therefore, according to the envelope width Δt, as shown in Fig. 7B, the bandwidth ΔF is narrowed, whereby a condition in which the space of ΔFc exists in the lower frequency band can be obtained. In other words, as shown by the dashed-dotted line in Fig. 7B, a condition that the broadband does not become until the frequency approaches zero (direct current component) can be obtained.

In summarizing the sampling, if the sampling frequency Fs is a frequency sufficiently higher than the highest frequency component included in the interference fringe, the frequency component of the original interference fringe can be reproduced as shown in Fig. 7B, but the sampling frequency Fs As lowering, aliasing occurs. That is, as shown in Fig. 7C, the original desired frequency component (solid line portion) and the bending frequency component (dashed line portion) due to aliasing are lowered (as if being divided). As shown in Fig. 7 (D), they are close to each other, and as shown in Fig. (E), the height is inverted.

Therefore, the present inventors paid attention to the following. That is, since there is a space of ΔFc in the band characteristic of the interference fringe itself as shown in Fig. 7B, by selecting the sampling frequency Fs, the desired frequency component and the bending frequency component are shown in Fig. 7E. The frequency can be separated. Therefore, as shown in Fig. 7F, the frequency component is removed by the filter, and the frequency axis is extended only as long as the frequency is lowered by further lowering the sampling frequency.

By doing so, a desired frequency component of the interference fringe can be obtained even at a frequency lower than the frequency satisfying the sampling theorem. In the end, the inventors focused on reducing the measurement time by reducing the data acquisition frequency by lowering the sampling frequency Fs.

The present invention provides a technique capable of performing three-dimensional measurement with a short measurement time.

In order to achieve the above object, it is necessary to consider the relationship between the envelope width Δt, the bandwidth ΔF, and the highest frequency Fh of the interference fringe, but in general, the period of the interference fringe is a source of generating original interference. It is about 1/2 of the period? Of the center wavelength of the light source of the broadband light source. The envelope width Δt of the interference fringe at that time depends on the bandwidth ΔF of the broadband light source as described above. In general, since ΔF < [1 / (λ / 2)] &lt; Fh is satisfied, the space ΔFc (= Fh-ΔF) can be sufficiently provided in the lower frequency band as shown in Fig. 7B.

Specifically, the invention described in claim 1 is further divided into a broadband light source 1 that outputs broadband light having a plurality of spectra, and the broadband light into a measurement light path in which a reference light path having a reference mirror and an object to be measured are arranged. And an optical path forming unit 5 which combines the reflected light from the reference mirror with the reflected light from the irradiation position of the irradiated range of the object to be irradiated and outputs the combined light; The interference fringe is formed by imaging the output from the optical path forming portion at a predetermined sampling timing with respect to the optical path variable means 8 for changing the optical path length and the optical path length by the optical path variable means. When the imaging means 10 which acquires the interference fringe data included and the characteristic value of the interference fringe are shown from the interference fringe data output from the imaging means In the three-dimensional shape measuring device for measuring the shape of the object to be measured, comprising: an optical path length detecting means (14) for obtaining a specific optical path length;

The predetermined sampling timing at the time of imaging by the imaging means is a timing at which aliasing occurs for an interference fringe included in the output of the optical path forming portion, and is obtained by the actual frequency component to be obtained in the frequency domain and the frequency generated by the aliasing. It becomes the timing which can be separated into a component, and,

The optical path length detecting means converts the interference fringe data acquired by the imaging means at the predetermined sampling timing into data in the frequency domain, and removes the unnecessary interference due to the aliasing, thereby creating a new interference fringe by the actual frequency component. An interference fringe data selecting section 14a for selecting data and an optical path length calculating section 14c for obtaining the specific optical path length representing the characteristic value of the interference fringe based on the new interference fringe data.

In the invention according to claim 2, in the invention according to claim 1, the interference fringe data selection unit performs Fourier transform of the interference fringe data output from the image pickup means into the number of sampling data based on the sampling timing. New interference fringe data of the frequency domain is selected by transforming the data of the domain and excluding unnecessary components by the aliasing on the frequency domain,

The optical path detection means converts the number of sampling data of the new interference fringe data in the frequency domain into the number of sampling data based on the sampling theorem, and then inversely transforms the new interference fringe data in the frequency domain. The specific optical path length indicating the characteristic value of the interference fringe is obtained by converting the envelope data of the new interference fringe into envelope data of the new interference fringe.

In the invention according to claim 3, in the invention according to claim 1 or 2, the timing at which the aliasing occurs with respect to the change in the optical path length by the optical path varying means means the optical path by the optical path varying means. The change in the field was set to the interval exceeding λ / 6 when the center wavelength of the broadband light was λ.

In the invention according to claim 4, in the invention according to claim 1 or 2, the image pickup means captures the image at a unique minimum exposure time, and the speed at which the optical path variable means changes the optical path length is v. When it did, it was set as the structure which image | photographs the time interval whose change of the said optical path length exceeds (lambda) / (6v) as said timing.

In the invention as set forth in claim 5, in the invention as set forth in claim 1 or 2, the optical path length detecting means further comprises: wavelength selection for extracting at least two wavelength components from the new interference fringe data selected by the interference fringe data selection unit; It has a part 14d, and the said optical path length calculation part was set as the structure which calculates | requires the optical path length which the phase difference of the extracted at least 2 wavelength component becomes substantially zero as said specific optical path length.

According to the present invention, since the interference fringe data is acquired (sampled) at the timing at which aliasing occurs, the unnecessary frequency component generated due to the aliasing is removed, and the displacement is measured. It is possible to lengthen the period, that is, reduce the number of times of data acquisition, reduce the optical processing time by that amount, and reduce the measurement time. That is, if it is expressed as [the optical field variable time + data acquisition frequency x (exposure time + acquisition process time)] as a main measurement time, a measurement time can be shortened only by the time to which the frequency of data acquisition is reduced.

EMBODIMENT OF THE INVENTION Embodiment which concerns on this invention is described using drawing. 1 is a diagram illustrating a functional configuration of the first embodiment. 2 is a diagram for explaining an interference fringe. FIG. 3 is a diagram showing a second embodiment in which the optical path length detecting means of FIG. 1 is changed. FIG. 4 is a diagram for explaining an operation of the interference fringe data selection unit of FIG. 3. 5 and 6 are diagrams for explaining the operation of the optical path varying means of FIG. 3, FIG. 5 shows an interference fringe of each frequency component, and FIG. 6 shows its phase characteristics. FIG. 7 is a diagram for explaining an operation of obtaining an interference fringe except for the effect of aliasing in the embodiment of FIG. 1. On the other hand, Figure 7 is also a view for explaining the background of the present invention in the column of the problem to be solved.

[One. Whole structure of 1st Embodiment]

As described above, in the first embodiment, aliasing is generated from the respective data obtained by imaging by the camera 10 of the image pickup means at a timing later than the timing satisfying the sampling theorem, that is, the timing at which aliasing occurs. Except for unnecessary components, data equivalent to an interference fringe obtained at a timing satisfying the original sampling theorem is extracted. Then, the optical path length in which the peak value of the intensity of the extracted interference fringe is generated (which is also good in the amount of change when the optical path length is changed until the interference fringe is generated) is obtained to obtain a specific optical path length (displacement).

In the following description, when the optical path length in which the interference fringe is generated when the optical path length of the measurement optical path is changed (the amount of change when the optical path length is changed until the interference fringe is generated in the optical path length) is called a specific optical path length. This indicates the displacement of the shape of the object under test.

In Fig. 1, the light source 1 uses a white light source that emits low coherency light having a plurality of wavelength components over a wide band in order to cause interference. The collimator lens 2 collects white light (broadband light) from the light source 1 and sends it to the beam splitter 3. The beam splitter 3 converts the direction of the white light and sends it to the objective lens 4. The objective lens 4 sends white light as parallel light to the beam splitter 5 (optical path forming unit). The beam splitter 5 splits the white light received from the objective lens 4 in two directions, and sends one to the object 7 as measurement light (the optical path from the beam splitter 5 to the object 7). The other is sent to the reference mirror 6 as the reference light (the light path from the beam splitter 5 to the reference mirror 6 is referred to as a reference light). In this example, the beam splitter 5 is fixed between the reference mirror 6, that is, the optical path length of the reference optical path is of a fixed length. Instead of the beam splitter 5, a translucent mirror may be configured.

The measurement optical path is configured to irradiate the desired irradiation range on the surface of the measurement target object 7 with white light at the same time.

The object to be measured 7 is mounted on the piezo 8 as the optical path variable means. The piezo 8 is constituted by a piezoelectric element, and the Z to be measured is continually moved to the XY plane (surface orthogonal to the surface of FIG. 1) by the instruction from the optical path control means 16. The optical path length of the measurement optical path is variably controlled at a predetermined speed by displacing (moving) in the axial direction (up and down direction of the paper in Fig. 1).

On the other hand, here, as the variable method for changing the optical path length in the present invention, it is continuous variable, and the variable speed is described to be constant, but it is variable even in the fine step shape compared to the data acquisition timing by a camera or the like described later. good.

The piezo 8 is a means (optical-field variable means) which changes the optical-field length of a measurement optical path with respect to the fixed position of the beam splitter 5 by control of the optical-field control means 16. FIG. On the other hand, although the optical path length of the reference optical path is described here by changing the optical path length of the measurement optical path, in order to generate an interference fringe, the piezo 8 is attached to the reference mirror 6 to fix the measurement optical path. The optical path length of the reference optical path can be changed.

The white light reflected from each of the reference mirror 6 and the measurement target object 7 (hereinafter sometimes referred to as a "return light") is combined (synthesized) by the beam splitter 5 to form an objective lens ( 4) more condensed. The return light passes through the beam splitter 3 and becomes parallel light by the imaging lens 9 and is input to the camera 10.

At this time, according to the instruction from the optical path control means 16, the camera 10 picks up the return light according to the distance (or time interval at which the piezo 8 changes the optical path length of the measurement light path). By doing so, the interference fringe by the return light is imaged (actually, the imaging is only imaging the return light, but among them includes the interference fringe by the return light that appears when the image data is developed later, 'Interference pattern is captured'. The picked-up interference fringe is stored in the memory 13. At this time, since the measurement light path is configured to irradiate the entire desired irradiation range of the measurement target object 7 simultaneously with white light as described above, each irradiation position of the irradiation range, that is, the position to be measured (hereinafter, 'measurement The interference fringe corresponding to the return light from the position.

On the other hand, as a modification of the optical system of FIG. 1, an optical system in which the objective lens 4 is arranged in each of the measurement optical path and the reference optical path instead of the position of FIG. 1 can be configured. It is not limited to the optical system of 1. However, the following description is demonstrated along FIG.

The timing of the photographing by the camera 10 and the timing of the memory by the memory 13 are both output in synchronization with the optical path length control means 16 as the timing of acquiring the data of the present invention. That is, the optical path control means 16 variably instructs the piezo 8 at a predetermined speed, generates a timing signal at a predetermined time interval, and sends the timing signal to the camera 10 and the memory 13 at that timing. Allows data acquisition. That is, the camera 10 and the memory 13 store and store the image pickup data (which becomes the luminance data indicating the luminance of the returned light) of the returned light at the timing of the timing signal.

Generally, the interference fringe by broadband light has a repetition period of half of the center wavelength (lambda) of broadband light. In order to acquire and reproduce the waveform of this λ / 2 cycle, in order to prevent aliasing from occurring, three data are normally required for the cycle, so that data is acquired at a repetition timing faster than λ / (2 × 3). It is necessary to do

In the present invention, since data is acquired at the timing at which aliasing occurs, the timing of photographing by the camera 10 and the storage timing of the memory 13 are repeated (longer) than the λ / 6 repetition period (hereinafter, 'sampling timing'. Fs' may be called). Specifically, when the optical path control means 16 drives the piezo 8 to vary the speed at which the optical path length is v, the optical path control means 16 displays a control signal at a time interval later than λ / 6v. (10) and the memory 13, and data is acquired at the timing of the control signal. When the optical path control means 16 drives the piezo 8 in steps instead of analog, it generates a timing signal with a time interval faster than lambda / 6v, and instructs the piezo 8 to vary the optical path length.

As a result, the memory 13 stores the captured image data using the time interval of the timing signal as an address. These timing advancing directions (ie, address directions) indicate the optical path length direction (Z-axis direction). At that time, the captured data is stored together with the measurement positions Xm and Yp. The information of the measurement positions Xm and Yp is the position of the pixel in the XY direction corresponding to the position of the imaging device of the camera 10. 2 shows an example of an interference fringe obtained by the processing of the signal processing means 20 described below from the data stored in the memory 13.

The signal processing means 20 includes an optical path length detecting means 14 and a displacement calculating means 15.

The optical path length detection means 14 is comprised from the interference fringe data selection part 14a and the optical path length calculation part 14c as shown in FIG. The interference fringe data selection unit 14a receives the captured image data from the memory 13, for example, data of the measurement positions Xm and Yp, and converts the data into the frequency domain data by FFT (Fourier transform). Then, as shown in Fig. 7E, since the data of the frequency component due to aliasing (the dotted line in Fig. 7E) and the frequency component of the original interference fringe exist, they are separated by a filter, Except for the frequency component, the component of the interference fringe (solid line part in Fig. 7E) is separated out.

Subsequently, as shown in Fig. 7F, the interference fringe data selection unit 14a samples the frequency axis at the original frequency axis, i.e., sampling timing such that no aliasing occurs according to the sampling theorem. The scale is returned to the frequency axis of the frequency domain (because it is amplified at a lower frequency than the original sampling frequency, so the frequency axis is displayed in a divided form, so it is returned). That is, it sends to the optical-field length calculation part 14c as interference fringe data (FIG. 2 is the data of the interference fringe at that time.).

On the other hand, in the waveform of the interference fringe of FIG. 2, the peak position of the center of the interference fringe, which is the characteristic value of the interference fringe, is the case where the optical path length of the reference optical path and the optical path length of the measurement optical path are the same. In addition, the wavelength of the interference fringe by white light is made by the synthesis | combination of each wavelength used as an element of broadband light, and becomes 1/2 of the wavelength (lambda) of the substantially center of those bands. In addition, the width of the interference fringe by the white light in the direction of the optical path length, that is, the width Δt of the envelope of the interference fringe is the degree of coherence of the white light (coherence length), that is, the frequency component in the frequency domain. It depends on the width ΔF. The lower the coherence (the longer the coherence length), that is, the larger the ΔF, the narrower the width Δt (see Fig. 7 (A), (B)). If the maximum frequency of the frequency component of the interference fringe is Fh, the width ΔF of the frequency component satisfies ΔF < [1 / (λ / 2)] &lt; Fh, and improves coherence. Lower surface (shorter coherence length)}, since the amplitude of the interference fringe becomes substantially constant and the peak disappears, the width ΔF is determined so that the width Δt of the interference fringe is such that the peak value of the interference fringe can be grasped. For example, in the model description, when the bandwidth of the light source 1 is determined such that the width ΔF is less than the highest frequency Fh / 2, as shown in Fig. 7E, the right half of the frequency band after sampling is shown. Only the actual frequency component of the interference fringe can be extracted by a high pass filter which is allocated to the actual frequency component of the interference fringe, the frequency component (aliased frequency component) by aliasing in the left half, and cuts Fh / 2. That is, the frequency component except the bending component which is an unnecessary component can be obtained.

The optical path length detecting means 14 performs an inverse FFT process on the actual frequency components excluding the frequency components (fracture components) due to aliasing, converts them into new interference fringes in the time domain, obtains the peak positions of the new interference fringes, and positions the positions. Determine the optical path length of the target (specific optical path length). The 'peak position' (or 'peak position') is a position on the horizontal axis at which the luminance (amplitude) of the interference fringes caused by white light is the maximum (hereinafter referred to as 'peak'). And the optical path length direction of the measurement optical path (Z-axis direction: up and down direction of the ground of FIG. 1), and the time axis direction when the optical path length is varied (time axis direction when the camera 10 is picked up at a predetermined time interval). .

On the other hand, since the captured image data stored in the memory 13 is stored at the timing (sampling timing) stored in the memory 13 (FIG. 2 shows them collectively and expressed continuously.), The optical path detection means 14 The interference fringe data selector 14a receives discrete data on the interference fringe. Since it is discrete in this way, the maximum position of an amplitude and the peak position of an envelope do not correspond, but since it is a smooth characteristic, you may calculate the peak position of an envelope by interpolation calculation from the maximum point of front and back amplitude. For example, as a method of obtaining peaks of an interference fringe from discrete image data, a technique of changing the optical path length stepwise and performing the following processing based on the discrete image data captured for each predetermined predetermined optical path length There is this. The direct current component is excluded by the digital high pass filter from the data of the interference fringe obtained from the imaging data. To rectify the power of the AC component. The integration data is integrated through a digital low pass filter through which a low repetition component is passed as compared with the rectified repetition component, thereby calculating the envelope data of the interference fringe. That is, normal envelope detection is performed. At this time, in accordance with the request for the fineness of the peak position, interpolation between the rectified repeating components is performed, for example, with a quadratic characteristic, and the interpolated repeating components are integrated to obtain envelope data. The position which becomes the peak of this envelope data is calculated | required. On the other hand, regardless of the timing (time interval) of the imaging data and the period of the interference fringe, the signal processing means 20 may obtain the peak position by discrete processing, as described in JP-A-9-318329. .

The displacement calculating means 15 is an optical path length of the peak position of the interference fringe at each measurement position of the measurement range of the measurement target 7, that is, each specific optical path length. Xm, Yp), the shape of the measurement object in the measured value with respect to the reference measurement position from the difference ts-t1 with the specific wavelength ts at the reference measurement position (Xs, Ys) obtained in the same manner as the specific wavelength t1. Find the displacement of.

Second Embodiment

It demonstrates based on FIG.3, FIG.4 and FIG.5 (However, FIG.4 is a figure for demonstrating an operation principle.). In the second embodiment, in which the first embodiment of FIG. 1 obtains the peak position of the interference fringe from the amplitude, the wavelengths of the phase and the wavelength of the green component and the wavelength of the plurality of frequency components, for example, the red component, are obtained. It is a form to obtain | require the optical path length with which phase corresponded as a specific optical path field. The signal processing means 20 of FIG. 1 is replaced with the signal processing means 20a of FIG. 3, and the camera 10 of FIG. 5 is used as a color camera. Therefore, in FIG. 1, the light source 1 is a light source including a component having a wavelength band of at least two waves in a plurality of wavelength components covering a wide band, and here, for example, a light source including respective color wavelength bands of red and green. Use You may synthesize | combine and use the light of each color wavelength of red and green. Configurations, operations, timings, and the like other than the signal processing means 20a are the same as in FIG. Hereinafter, the signal processing means 20a will be described.

In FIG. 3, the interference fringe data selection unit 14e is basically the same as the interference fringe data selection unit 14a of FIG. 1. However, since one of the main structures of the present invention is used, the interference fringe data selection unit 14e will be described with reference to FIG. 4. The interference fringe data selection unit 14e acquires data at the sampling timing (timing for a period longer than λ / 6) in which the camera 10 or the memory 13 generates aliasing as described above. 13) converts the time domain data from 13 into frequency domain data by FFT, filters out unnecessary components due to aliasing, obtains a frequency component of an interference fringe, and sends it to a wavelength selecting section 14d to send a wavelength selecting section 14d. , Return the frequency axis to the original interference pattern on the time axis.

The interference fringe data selection unit 14e selects, as a desired data, the data in the right half (data shown in light gray in FIG. 4 (D)) from the data of FIG. 4D obtained by the FFT as a desired data, Remove unnecessary components in the previous frequency band. Then, the frequency axis of Fig. 4D is extended (the right half indicated by the gray area of Fig. 4E) and returned to the wavelength selection unit 14d.

Here, using Fig. 4, a general principle explanation will be made. Fig. 4A is an example of the case where the interference fringes are formed by the wavelength components of red and green, and are measured at sampling timing T0 (timing faster than λ / 6v) that satisfies the sampling theorem. Fig. 4B is data converted into the frequency domain of Fig. 4A. In this case, the left half of data is selected and used by the filter. According to the present invention, for example, when data is acquired (sampled) at half of T0 of sampling timing that satisfies sampling theorem and at a timing at which aliasing occurs, instead of FIG. Can be obtained for data in the same time domain. As the data of the frequency domain, instead of the data of FIG. 4 (B), as shown in FIG. 4 (D), the data in which the distribution of left and right is changed as compared with the distribution of the frequency component of FIG. 4 (B) can be obtained. That is, unnecessary components on the right side of FIG. 4B appear at the frequency position on the left side of FIG. 4D by aliasing, and components of the original interference region on the left side of FIG. 4B are shown in FIG. 4D. It appears at the frequency position on the right. And the frequency axis of FIG. 4 (D) is half in relationship with a sampling frequency compared with the frequency axis of FIG. 4 (B). Therefore, since the component waveforms on the left side of FIG. 4 (B) and the component waveforms on the right side of FIG. 4 (D) are the same, if this is separated by a filter and the scale is changed as shown in FIG. Data of the frequency component of the left half of (B) can be obtained.

The timing relationship will be described in detail. For example, the time is determined by the number of sampling data m based on the timing of x times the sampling timing T0 based on the sampling theorem by the camera 10 (two times in FIG. 4C). If the interference fringe data of the area is converted into frequency domain data by FTT processing, the number of sample data n based on the sampling timing T0 based on the sampling theorem is m = n / x. The scale is reduced and treated. That is, it appears in a divided form as described above. Accordingly, the interference fringe data selection unit 14e changes the scale of the frequency axis of FIG. 4D, and is the same as the frequency axis based on the sampling theorem as shown in FIG. 4E (the same as in FIG. 4B). In order to process, the number of sample data handled in the FFT process is added to m, and only the number Δn = nm = n (x-1) / x is supplemented with data of value 0 (0 padding). That is, the number of sampling data decreases, and the empty n-m frequency region is filled with zero data. On the other hand, since the frequency axis scales only n / x in theory, the scale on this frequency axis may be changed by x times and handled without changing the number of sampling data. The relationship between the number of sample data and the frequency axis by this FFT is also the same as in the first embodiment.

In Fig. 3, the wavelength selector 14d receives the data of the frequency domain sent from the interference fringe data selector 14e, and the filter selects, for example, the blue (B) component and the green (G) frequency. The components are classified into three components, a red (R) component, and each component is sent to the optical path calculator 14f. In the optical path calculation unit 14f, the B phase calculation unit 14f1, the G phase calculation unit 14f2, and the R phase calculation unit 14f1m in the interference fringe data of the time domain obtained by receiving the respective components and performing inverse Fourier transform. ) Obtains the phase change of the corresponding component (for example, orthogonal demodulates to obtain the phase). 5 shows their respective phase changes (vertical axis) with respect to changes in the optical path length (horizontal axis). As shown in Fig. 6, the optical path length determining means 14f4 determines the optical path length in which the three components are in phase as the specific optical path length at the measurement position. At this time, when data is discrete and it is difficult to identify the optical path length of the coincidence point of the phase, a method such as interpolation is used as described above.

In the above configuration, the signal processing means 20 and 20a and the optical path control means 16 can be constituted by a CPU and a memory.

As described above, since data can be acquired and measured at the timing at which aliasing occurs, the data acquisition frequency can be reduced. Therefore, if it is expressed as [the optical field variable time + data acquisition frequency x (exposure time + acquisition process time)] as a main measurement time, a measurement time can be shortened only by the time whose data acquisition frequency is reduced.

1 is a diagram illustrating a functional configuration of the first embodiment.

2 is a diagram for explaining an interference fringe.

It is a figure which shows 2nd Embodiment which changed the optical path length detection means of FIG.

FIG. 4 is a diagram for describing an operation of the interference fringe data selector of FIG. 3.

FIG. 5 is a view for explaining the operation of the optical path length determining means of FIG. 3, and shows the interference fringe of each frequency component.

FIG. 6 is a diagram for explaining the operation of the optical path length determining means of FIG. 3 and shows the phase characteristics of each frequency component.

FIG. 7 is a diagram for explaining an operation of obtaining an interference fringe except for the effect of aliasing in the first embodiment. Moreover, it is also a figure for demonstrating the background of this invention in the column of the problem which this invention tries to solve.

<Description of the symbols for the main parts of the drawings>

1: light source 2: collimator lens

3: beam splitter 4: objective lens

5: beam splitter 6: reference mirror

7: measuring object 8: piezo

9: imaging lens 10: camera

13 memory 14 optical path detection means

14a, 14e: Interference pattern data selection section 14d: Wavelength selection section

14, 14f: optical field calculation unit 15: displacement calculation means

16: optical path control means 18: user interface

20,20a: signal processing means

Claims (5)

A broadband light source 1 for outputting broadband light having a plurality of spectra, and the broadband light is incident on a measuring light path in which the reference light path having a reference mirror and the object to be measured are arranged, and the reflected light and the irradiation from the reference mirror are irradiated. An optical path forming unit 5 for combining and outputting each of the reflected light beams from the irradiation position of the irradiated range of the object to be measured and an optical path length varying means 8 for changing the optical path length of either the reference optical path or the measurement optical path And imaging means (10) for acquiring interference fringe data including an interference fringe by imaging the output from the optical path forming portion at a predetermined sampling timing with respect to the change in the optical path length by the optical path variable means; Optical field length detection means (1) for obtaining a specific optical path length when the characteristic value of the interference fringe is represented from the interference fringe data output from the imaging means (1) A three-dimensional shape measuring device comprising 4) and measuring the shape of the object under measurement based on the obtained specific optical path length, The predetermined sampling timing at the time of imaging by the imaging means is a timing at which aliasing occurs for an interference fringe included in an output of the optical path forming portion, and is generated by the actual frequency component to be obtained in the frequency domain and the aliasing. The timing can be separated into frequency components, and The optical path length detecting means converts the interference fringe data acquired by the imaging means at the predetermined sampling timing into data in the frequency domain, and removes the unnecessary interference due to the aliasing, thereby creating a new interference fringe by the actual frequency component. An interference fringe data selection section 14a for selecting data and an optical path length calculation section 14c for obtaining the specific optical path length indicating a characteristic value of the interference fringe, based on the new interference fringe data. Three-dimensional shape measuring device. The data of the frequency domain according to claim 1, wherein the interference fringe data selector converts the data of the frequency domain by Fourier transforming the interference fringe data output from the image pickup means to the number of sampling data based on the sampling timing. New interference fringe data in the frequency domain is selected by excluding unnecessary components due to aliasing, The optical path detection means converts the number of sampling data of the new interference fringe data in the frequency domain into the number of sampling data based on the sampling theorem, and then inversely transforms the new interference fringe data in the frequency domain. Converting into envelope data of a new interference fringe, and obtaining the specific optical path length representing a characteristic value of the interference fringe based on the new interference fringe data in the time domain. 3. The timing according to claim 1 or 2, wherein the timing at which the aliasing occurs with respect to the change of the optical path length by the optical path varying means means that the change of the optical path length by the optical path varying means is the center wavelength of the broadband light. When it is set to λ, the three-dimensional shape measuring device characterized in that the interval over λ / 6. 3. The change in the optical path length according to claim 1 or 2, wherein the imaging means captures images with an intrinsic minimum exposure time and sets the speed at which the optical path length varying means changes the optical path length to v. A three-dimensional shape measuring apparatus characterized by imaging the time interval over / (6v) as said timing. 3. The optical path detection means according to claim 1 or 2 has a wavelength selection section 14d for extracting at least two wavelength components from the new interference fringe data selected by the interference fringe data selection section. And the optical path length calculating unit obtains the optical path lengths of which the phase difference becomes zero based on the extracted new interference fringe data of the time domain of the at least two wavelength components as the specific optical path length.

Images