JP2014085157A - Strain measuring method and strain measuring system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a strain measuring method and strain measuring system capable of improving the measuring accuracy of strain of a measuring object by certainly image-recognizing a mark put on a surface of the measuring object.SOLUTION: The strain measuring method includes: an imaging process S13 of irregularly arranging a plurality of marks on a surface, imaging the surface of a measuring object in chronological order from the time before deformation to the time after the deformation, and acquiring an initial image before the deformation, an intermediate image during the deformation, and a final image after the deformation; an image selecting process S14 of selecting N images including the initial image, intermediate image, and final image; and a strain calculation process S16 of identifying, from the image X, by image recognition, the position indicating the highest correlation with each of a plurality of areas into which the image X(where, i is 1 or more and N-1 or less) is divided, calculating the moving amount of each area in the image Xand image X, and adding the plurality of moving amounts of the initial image to the final image to each other to calculate the strain of the measuring object.

Description

  The present disclosure relates to a strain measurement method and a strain measurement system for measuring a strain on a surface of a measurement object.

  In general, as one method for measuring strain generated in a measurement object, a method is known in which strain is obtained using images of the measurement object before and after deformation captured by an imaging device. Since strain measurement using an image can measure strain without contact, there is no need to install a measuring device such as a strain gauge on the measurement object, and there is an advantage that strain in a minute region can be measured.

  In addition, from the viewpoint of clearly recognizing the displacement of the surface of the measurement object, a mark is attached to the surface of the measurement object with a seal or paint, and the amount of displacement of the mark is determined from images before and after the deformation of the measurement object including the mark. A method for calculating and obtaining strain is also known. For example, in Patent Document 1, three marks are attached on the curved surface of a measurement object so that each mark is positioned at the apex position of an equilateral triangle. A method is disclosed in which a distortion of a measurement object is calculated by imaging from a location.

  Patent Document 2 discloses a method of measuring strain from a surface image before and after deformation of a measurement object to which irregular array marks are attached. In this method, as shown in FIG. 11, the surface of the measurement object 50 before and after deformation is photographed with a camera 60. Then, the pre-deformation image 51 is divided into small areas 52, and the movement amount of each area 52 in the post-deformation image is measured from the correlation between the pre-deformation image and the post-deformation image. Based on the movement amount of each area 52, the amount of displacement of the calculation target area in the post-deformation image from the pre-deformation image is calculated to obtain the strain.

JP-A-62-231104 JP 2012-132786 A

By the way, when measuring the amount of displacement of an image using the mark marked on the measurement object, it is important to recognize the mark accurately in order to measure the distortion with high accuracy. Although the method described in Patent Document 1 uses only three marks, image recognition is easy. However, since it is necessary to accurately attach the three marks to the surface of the measurement object, it takes time to attach the marks. In addition, there is a problem that the operation is complicated and the work is complicated.
In this regard, in the method described in Patent Document 2, the marks may be irregularly arranged, saving time and shortening time. Since the marks are irregularly arranged, the displacement amount of the image can be measured. However, when the measurement object is greatly deformed, the target mark is deformed from the pre-deformation image 52 shown in FIG. 12A to the post-deformation image 53 shown in FIG. 12B. In some cases, the image recognition accuracy may be reduced. For this reason, there is a possibility that the measurement accuracy of the strain is also lowered.

  An object of at least one embodiment of the present invention is to provide a strain measurement method and a strain measurement system capable of reliably recognizing a mark attached to the surface of a measurement object and improving the measurement accuracy of strain in the measurement object. It is.

A strain measurement method according to at least one embodiment of the present invention is a strain measurement method for measuring strain on a surface of a measurement object, wherein a plurality of marks are irregularly arranged on the surface. The surface is imaged in time series from before deformation to after deformation, and includes at least three initial images of the measurement object before deformation, at least one intermediate image being deformed, and a final image after deformation. An imaging step of acquiring images, and an image selection step of selecting N images (where N is 3 or more) including the initial image, the intermediate image, and the final image acquired in the imaging step; The position showing the best correlation with each area obtained by dividing the image X _i selected in the image selection step (where i is 1 to N−1) into a plurality of areas is identified from the image X _{i + 1} by image recognition. Then The image X _i and the amount of movement of each area in the image X _{i + 1} is calculated, strain and adding a plurality of the movement amount from the initial image to the final image to calculate the strain of the object to be measured is calculated A process.

According to the strain measurement method, the mark attached to the surface of the measurement object can be reliably recognized, and the strain measurement accuracy in the measurement object can be improved. That is, the surface of the measurement object is imaged in time series from before deformation to after deformation, and includes at least three images including an initial image before deformation, at least one intermediate image being deformed, and a final image after deformation. Get the images. From these three or more images, N images (where N is three or more) are selected, and among these N images, the images X _i and X _{i + 1} that are adjacent in time series are compared with each other. The amount of movement for each area is calculated. Therefore, it is possible to compare images having a deformation amount smaller than the deformation amount between the initial image and the final image, and the shape of the mark is not greatly changed, so that the image recognition can be reliably performed. In particular, it is possible to perform image recognition with high accuracy even for distortions accompanied by significant deformation. Since the distortion of the measurement object is calculated by adding a plurality of movement amounts from the initial image to the final image, the measurement accuracy of the distortion in the measurement object can be improved.

In some embodiments, when two or more intermediate images are acquired in the imaging step, the intermediate image may be selected in the image selection step corresponding to a time-series interval stored in advance. Good.
For example, when a large number of intermediate images are acquired, calculating the movement amount for all the images increases the processing load and takes time. Therefore, in the above-described embodiment, an intermediate image is selected corresponding to a time series interval stored in advance. Thereby, the processing load of image recognition can be reduced, and the processing time can be shortened.
As the time-series interval, an intermediate image for each preset time interval may be selected, or an intermediate image for each preset number of intervals may be selected.

In some embodiments, the when the intermediate image in the imaging step is acquired two or more, in the distortion calculating step, said said each area image X _{i + 1} of the image X _i by the image recognition When the correlation with each area cannot be obtained, the process may return to the image selection step and select an image before the image X _{i + 1} .
Thus, if the correlation between each area and each area of the image X _{i + 1} of the image X _i by the image recognition can not be obtained, i.e., the deformation of the mark between the image X _i and image X _{i + 1} is from the image recognition limit Is larger, the process returns to the image selection step to select an image before the image X _{i + 1} . Thereby, an image can be recognized reliably and it contributes to the improvement of the measurement precision of distortion.

In some embodiments, in the strain calculation step, a calculation target area for calculating strain is selected from each area of the image after deformation, and a point separated by a predetermined distance with the selected calculation target area as a central point. The distance between the two areas after deformation may be calculated based on the respective displacement amounts in the two areas that are symmetrical, and the strain of the calculation target area may be calculated.
As described above, it is possible to calculate more accurate strain by calculating the strain based on the displacement amounts of the two areas separated from the calculation target area by a predetermined distance.

A strain measurement system according to at least one embodiment of the present invention is a strain measurement system for measuring strain on a surface of a measurement object, wherein a plurality of marks are irregularly arranged on the surface of the measurement object. At least three images including an imaging unit that images the surface, an initial image before deformation of the measurement object imaged by the imaging unit, at least one intermediate image being deformed, and a final image after deformation A storage unit for storing the image of the image, and N images (where N is 3 or more) including the initial image, the intermediate image, and the final image among at least three images stored in the storage unit. Image recognition unit for selecting an image to be selected and a position showing the best correlation with each area obtained by dividing the image X _i selected by the image selection unit (where i is 1 to N−1) into a plurality of areas By the picture X _{i + 1 is} specified, and the movement amount of each area in the image X _i and the image X _{i + 1} is calculated, and a plurality of the movement amounts from the initial image to the final image are added to calculate the measurement object. And a strain calculator that calculates strain.

According to the strain measurement system, the mark attached to the surface of the measurement object can be reliably recognized, and the measurement accuracy of strain on the measurement object can be improved. That is, N images (where N is 3 or more) are selected from three or more images obtained by capturing the surface of the measurement object in time series from before the deformation to after the deformation, and among these N images Then, the amount of movement of each area is calculated by comparing the image images X _i and X _{i + 1} adjacent in time series. Therefore, it is possible to compare images having a deformation amount smaller than the deformation amount between the initial image and the final image, and the shape of the mark is not greatly changed, so that the image recognition can be reliably performed. Since the distortion of the measurement object is calculated by adding a plurality of movement amounts from the initial image to the final image, the measurement accuracy of the distortion in the measurement object can be improved.

According to at least one embodiment of the present invention, N images (where N is 3 or more) are selected from three or more images obtained by capturing the surface of the measurement object in time series from before deformation to after deformation. Then, among the N images, the image X _i and the image X _{i + 1} that are adjacent in time series are compared, and the movement amount of each area is calculated. Therefore, it is possible to compare images having a deformation amount smaller than the deformation amount between the initial image and the final image, and the shape of the mark is not greatly changed, so that the image recognition can be reliably performed. Since the distortion of the measurement object is calculated by adding a plurality of movement amounts from the initial image to the final image, the measurement accuracy of the distortion in the measurement object can be improved.

(A), (b) is a schematic arrangement drawing of the distortion | strain measurement system concerning 1st embodiment of this invention. 1 is a schematic configuration diagram of a strain measurement system according to a first embodiment of the present invention. (A) is a figure explaining the comparison of an image, (b) is a figure which shows the change of the distortion shape of each image. It is a figure which shows the state which divided | segmented the 2nd image into several areas, respectively. It is a figure which shows an example of the method of comparing a target image and the image for a comparison. It is the schematic for demonstrating the correlation degree of a target image and the image for a comparison. It is a conceptual diagram which shows the method of calculating distortion from displacement amount. It is a figure which shows an example of the data table of distortion | strain (epsilon) xt of a X-axis direction. It is a figure which shows an example of the distribution map of distortion | strain (epsilon) xt of a X-axis direction. It is a figure which shows the example of a processing flow of distortion calculation. It is a figure explaining the conventional strain measuring method. (A) is an image of the mark before deformation, and (b) is an image of the mark after deformation.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the dimensions, materials, shapes, relative arrangements, and the like of the components described below as the embodiments or shown in the drawings as the embodiments are not intended to limit the scope of the present invention. It is just an example.

  Fig.1 (a) and FIG.1 (b) are front views of schematic arrangement | positioning of the strain measurement system 1 concerning 1st embodiment of this invention, respectively. FIG. 2 is a schematic configuration diagram of the strain measurement system 1 according to the first embodiment of the present invention.

As shown in these drawings, the strain measurement system 1 according to the present embodiment includes an imaging device 3 that images the measurement target region S on the surface of the measurement object 2 from the front, and an image X _i captured by the imaging device 3. And a measuring device 10 that calculates strain generated in the measurement object 2 based on the above.

The imaging device 3 is arranged on the front side of the measurement object 2 and captures an image X for calculating the strain of the measurement object 2. The imaging device 3 is a so-called digital camera, and includes a camera body 3a having an imaging device 3c such as a CCD or CMOS, and a lens 3b that forms an image of a subject on the imaging device 3c. In the present embodiment, a digital camera is used for the imaging device 3, but the present invention is not limited to this, and a digital video camera, a scanner, or the like may be used.
Here, the image X is captured when three or more along the sequence, including the initial image X _1, intermediate image X _i, the final image X _N. N is the number of images selected by an image selection program to be described later, and is 3 or more. Moreover, i is 1 or more and N-1 or less, and is the order of the images along the time series among the images selected by the image selection program. In the case of the intermediate image X _i , i is 2 or more and N−1 or less.

  The measurement device 10 is a nonvolatile device such as an I / F unit 16 (see FIG. 2) that captures a plurality of images X captured by the imaging device 3 and a hard disk drive that stores the plurality of images X captured by the I / F unit 16. Storage unit 11, a non-volatile memory 12 such as a RAM, a calculation unit 13 such as a CPU that reads and executes a program or the like held in the storage unit 11, and an input unit such as a keyboard or a mouse 14 and an output unit 15 such as a display or a printer. As the measuring device 10, for example, a personal computer can be applied.

The strain calculation processing by the measuring apparatus 10 is realized by the calculation unit 13 reading and executing the initial image X ₁ , the intermediate image X _i , the final image X _N and the various programs 11 a to 11 c stored in the storage unit 11. Is done.

The storage unit 11 captures the measurement object 2 and stores the captured image X in the storage unit 11, and three or more images X captured in time series by the imaging program 11 a. A strain calculation program 11c that calculates a displacement amount of the measurement object 2 based on _i and calculates a strain based on the displacement amount is stored.

  Next, processing executed by the imaging program 11a, the image selection program 11b, and the distortion calculation program 11c will be described.

<About processing executed by the imaging program 11a>
The calculation unit 13 receives values relating to imaging conditions such as the imaging range, imaging distance, and focal length of the imaging device 3 from the user by the input unit 14, and performs processing for changing the setting of the imaging device 3 to each received value. Run.
Subsequently, the calculation unit 13 transmits an imaging signal for performing imaging to the imaging device 3. Imaging is executed by the imaging device 3 that has received this imaging signal, and a plurality of images X obtained by imaging the surface of the measurement object 2 are acquired. A plurality of image X includes initial image _{X 1,} intermediate image X _i, the final image _{X N.} The acquired plurality of images X are all stored in the storage unit 11 via the I / F unit 16.
The calculation unit 13 performs a process of imaging the measurement object 2 before, during and after the deformation of the measurement object 2. Thereby, a plurality of images along the time series including the images X before, during and after the deformation are acquired.

<Processing Performed by Image Selection Program 11b>
As shown in FIG. 3A, the calculation unit 13 obtains an initial image X ₁ before deformation, an intermediate image X _{i being} deformed, and a final image X _N after deformation from a plurality of images stored in the storage unit 11. Select N images to include. For example, when N is three, one initial image, one intermediate image X _i , and one final image X _N are selected. Of course, a plurality of intermediate images X _i may be selected. As a result, as shown in FIG. 3B, an intermediate image X _i having a deformation amount smaller than the deformation amount between the initial image X ₁ and the final image X _N after deformation is selected, and image recognition described later is ensured. Can be performed. FIG. 3A is a diagram for explaining comparison of images, and FIG. 3B is a diagram showing a change in the distortion shape of each image.
Further, when the intermediate image X _i in the storage unit 11 has a plurality of storage may select an intermediate image X _i corresponding to the spacing of the pre-stored time series. As the time-series interval, an intermediate image X _i for each preset time interval may be selected, or an intermediate image X _i for each preset number of intervals may be selected. For example, when strain measurement is performed a plurality of times under the same conditions, a time interval (or number interval) at which image recognition is possible is stored at the time of the previous strain measurement, and an intermediate image is selected using this time interval. . Thereby, even when a large number of intermediate images X _i are acquired, it is possible to avoid calculating the movement amount for all the images, and to reduce the processing load of image recognition described later. Time can be shortened.
The selected image X may be stored in the storage unit 11.

<About processing executed by the strain calculation program 11c>
As a first step, the calculation unit 13 reads the image X _i and the image X _{i + 1} from the N images selected by the image selection program from the storage unit 11, and the image X as shown in FIG. _{The process} of comparing _i with the image X _{i + 1} and calculating the displacement amount of the surface of the measurement object 2 is executed. Note that the image X _i and the image X _{i + 1} are images that are adjacent in time series among the selected N images.

For the calculation of the displacement amount, a digital image correlation method capable of calculating the displacement amount from the luminance value distribution of the image X _i and the image X _{i + 1} was used.

Specifically, as shown in FIG. 4, first, the image X _i and the image X _{i + 1} are each divided into a plurality of areas E _ij (for example, 36 × 36 pixels).
Next, one area E _ij (for example, area E ₂₂ ) is selected from the plurality of areas E _ij of the image X _i . Hereinafter, one area E _ij selected from the area E _ij of the image X _i is referred to as a target image Xaa.
Further, a comparative image Xbb is extracted by designating a cut-out section (for example, 72 × 72 pixels) slightly larger than the area E _ij for the image X _{i + 1} . Note that the size of the cut-out section can be set regardless of the size of the area E _ij and is set as appropriate according to the processing capability of the calculation unit 13 and the like.

  FIG. 5 is a diagram illustrating an example of a method for comparing the target image Xaa and the comparison image Xbb. FIG. 6 is a schematic diagram for explaining the degree of correlation between the target image Xaa and the comparison image Xbb. In the following description, as shown in FIG. 5, the left-right direction in the figure is the X-axis direction and the up-down direction in the figure is the Y-axis direction.

  As illustrated in FIG. 5, the calculation unit 13 superimposes the target image Xaa on the comparison image Xbb and shifts the position of the target image Xaa up and down, left and right, and determines the position that obtains the best correlation with the target image Xaa. It is specified from Xbb. As shown in FIG. 6, the degree of correlation increases as the value of the correlation function S = f (x, y) decreases, and the degree of correlation of the luminance value distribution increases, and the accuracy of position detection increases. Since the actual displacement amount may be smaller than the size of one pixel, the luminance value distribution is interpolated using a spline function or the like so that the displacement amount can be detected with a resolution of one pixel or less, and the function S Position coordinates and X axis directions, Y axis directions, and XY directions (directions inclined by 45 ° from the X axis toward the Y axis) are calculated (Δx, Δy, Δxy), respectively.

The process of superimposing the target image Xaa on the comparison image Xbb described above is repeatedly performed for all the areas E _ij , in the X axis direction, the Y axis direction, and the XY direction for each area E _{ij in the} entire image X _{i + 1} . Displacement amounts (Δx, Δy, Δxy) are respectively calculated.

  Subsequently, the calculation unit 13 stores the calculated displacement amounts (Δx, Δy, Δxy) in the memory 12. The calculated displacement amounts (Δx, Δy, Δxy) may be stored in the storage unit 11.

  In this embodiment, the case where the digital image correlation method is used as the image analysis method has been described. However, the present invention is not limited to this method, and a general image analysis method may be used.

Next, as a second step, the calculation unit 13 executes a strain calculation process using each displacement amount calculated as described above.
FIG. 7 is a conceptual diagram illustrating a method for calculating strain from the amount of displacement.
As shown in FIG. 7, the calculation unit 13 reads out the displacement amounts (Δx, Δy, Δxy) in the X-axis direction, the Y-axis direction, and the XY direction from the memory 12 and calculates the strain from the image X _{i + 1.} A target area E _ij (for example, E ₄₄ ) is selected, and two areas E _ij (for example, symmetric) separated by a predetermined distance L with the selected calculation target area E _ij (for example, E ₄₄ ) as a central point. In the case of calculating the X-axis direction, a process of calculating the strain of the calculation target area E _ij (for example, E ₄₄ ) is executed based on the respective displacement amounts (Δx, Δy, Δxy) in E ₁₄ , E ₇₄ ). . Here, the distance between the two areas E _ij existing in a point-symmetric position with the calculation target area E _ij as the center point is defined as a distance GL between the scores. Since the two areas E _ij exist at point symmetry positions with respect to the calculation target area E _ij , the inter-score distance GL is twice the predetermined distance L.

Calculation unit 13, the value of the score distance GL accepted from the user at the input unit 14 such as a keyboard, X-axis direction in the calculation target area E _ij, strain each of the Y-axis direction and the XY direction (εx, εy, εxy) The process of calculating is executed.

  Below, the calculation method of each distortion ((epsilon) x, (epsilon) y, (epsilon) xy) is demonstrated.

---- When calculating strain in the X-axis direction ---
As shown in FIG. 7, when calculating the strain εx in the X-axis direction of the area E _ij (for example, E ₄₄ ), ± 0. 0 in the X-axis direction based on the inter-score distance GL set by the user. The strain εx is calculated from the displacement amount of both areas E _ij (for example, E ₁₄ and E ₇₄ ) separated by 5 GL. The strain εx is calculated from the following equation (1).
εx = (Δx (x + 0.5GL, y) −Δx (x−0.5GL, y)) / GL
Expression (1) where εx is the strain in the X-axis direction in each area E _ij , Δx is the amount of displacement in the X-axis direction in each area E _ij , and GL is the distance between scores.

--- When calculating strain in the Y-axis direction ---
In addition, as shown in FIG. 7, when calculating the strain εy in the Y-axis direction of the area E _ij (for example, E ₄₄ ), both distances ± 0.5 GL apart in the Y-axis direction are calculated based on the inter-score distance GL. The strain εy is calculated from the displacement amount of the area E _ij (for example, E ₄₁ , E ₄₇ ). The strain εy is calculated from the following equation (2).
εy = (Δy (x, y + 0.5GL) −Δy (x, y−0.5GL)) / GL
(2) where εy is the strain in the Y-axis direction in each area E _ij , and Δy is the amount of displacement in the Y-axis direction in each area E _ij .

--- When calculating strain in XY direction ---
Furthermore, when calculating the strain εxy the XY direction of the area E _ij, based on the scores distance GL, it calculates the more strain εxy displacement of the both areas E _ij away ± 0.5GL XY directions. The strain εxy is calculated from the following equation (3).
εxy = (Δxy (x + 0.5GL , y + 0.5GL) - Δxy (x-0.5GL, y-0.5GL)) / GL ··· (3) equation where, εxy: XY direction in each area E _ij , Δxy: a displacement amount in the XY direction in each area E _ij .

  Subsequently, the calculation unit 13 stores the respective strains (εx, εy, εxy) in the X-axis direction, the Y-axis direction, and the XY direction calculated by the above formulas (1), (2), and (3), respectively. 12. The data may be stored in the storage unit 11.

The first step and the second step described above are repeated to calculate each distortion between the N images selected by the image selection program 11b. That is, the strain between the initial image X ₁ and the intermediate image X _i, strain between the intermediate image X _i and an intermediate image X _{i + 1,} calculates all the distortion between the intermediate image X _{i + 1} and the final image X _N . Then, by adding the plurality of movement amount from the initial image X ₁ to the last image X _N, and calculates the strain of the measuring object.

Subsequently, the calculation unit 13 creates a strain data table based on the calculated strains (εxt, εyt, εxyt) in the X-axis direction, the Y-axis direction, and the XY direction, respectively.
In addition, distribution maps of the corrected strains (εxt, εyt, εxyt) in the X-axis direction, the Y-axis direction, and the XY direction are created.
That is, an X-axis direction strain data table and strain distribution diagram, a Y-axis direction strain data table and strain distribution diagram, and an XY direction strain data table and strain distribution diagram are created.

Subsequently, the calculation unit 13 associates the position information of the area E _ij with each created strain data table and each strain distribution diagram, and stores them in the storage unit 11. Details of each strain data table and each strain distribution diagram will be described later.
Each strain data table and each strain distribution diagram may be stored in the memory 12.

  Here, since the respective strains (εx, εy, εxy) in the X-axis direction, the Y-axis direction, and the XY direction are respectively calculated, the maximum principal strain, the minimum principal strain, and the maximum principal shear strain may be calculated, respectively. .

  Next, each strain data table and each strain distribution diagram stored in the storage unit 11 will be described with an example of the strain data table and strain distribution diagram in the X-axis direction.

FIG. 8 is a diagram illustrating an example of a strain data table in the X-axis direction.
As shown in the figure, the strain data table in the X-axis direction, a table that associates the area E _ij number of strain εxt the final image X _N. The area E _ij number is input in advance by the user via the input unit 14 before performing the measurement work.
Note that the strain data table in the Y-axis direction and the strain data table in the XY direction are also stored in the storage unit 11 in the same format.

FIG. 9 is a diagram illustrating an example of a strain distribution diagram in the X-axis direction.
As shown in the figure, strain distribution diagram of the X-axis direction is a contour map showing the correspondence to the area E _ij number of strain εxt the final image X _N. The area E _ij number is input in advance by the user via the input unit 14 before performing the measurement work.
Note that the strain distribution diagram in the X-axis direction and the strain distribution diagram in the XY direction are also stored in the storage unit 11 in the same format as the strain distribution diagram in the X-axis direction.

  Finally, the calculation unit 13 transmits each strain data table and each strain distribution diagram to the output unit 15, and displays each strain data table and each strain distribution diagram on the monitor of the output unit 15.

(Strain calculation processing flow)
Next, a processing flow for calculating the strain generated on the surface of the measurement object 2 using the strain measurement system 1 having the above-described configuration will be described with reference to FIG.

FIG. 10 is a diagram illustrating an example of a processing flow for calculating strain.
As shown in the figure, first, a marking step is performed in which a coating material is irregularly sprayed on the measurement target region of the measurement target 2 by spraying or the like (step S10). At this time, from the viewpoint of improving the accuracy of image recognition, it is preferable that at least one of the shape and the outer dimension of the mark is different. In the present embodiment, the case where the mark is marked by spraying paint is exemplified, but the present invention is not limited to this, and the powder marking agent may be attached to the surface of the measurement object 2 with chalk.

In parallel with the marking step, an installation step is performed in which the imaging device 3 for imaging the measurement object 2 is installed at a predetermined position and fixed so as not to move (step S11).
When the imaging device 3 is installed and the sprayed paint is sufficiently dried, the process proceeds to the imaging condition setting step (step S12).

Next, in the imaging condition setting step (step S <b> 12), the calculation unit 13 of the strain measurement system 1 inputs values such as “imaging range”, “imaging distance”, “focal length”, and the like related to the imaging conditions of the imaging device 3. It receives from a user in the part 14. Note that the imaging conditions of the imaging device 3 may be different from each other or may be exactly the same.
When the imaging conditions of the imaging device 3 are set, the process proceeds to the imaging process (step S13).

In the imaging step (step S <b> 13), the calculation unit 13 transmits an imaging signal to the imaging device 3. The imaging device 3 receives the imaging signal and images the surface of the measurement object 2 under the imaging conditions set in the imaging condition setting step (step S12). At this time, an initial image X ₁ before deformation, an intermediate image X _i during deformation, and a final image X _N after deformation are acquired.
Each acquired image is stored in the storage unit 11 by the calculation unit 13 via the I / F unit 16. If each image is acquired, it will transfer to an image selection process (step S14).

In the image selection step (step S14), at least 3 including an initial image X ₁ before deformation, an intermediate image X _{i being} deformed, and a final image X _N after deformation among the plurality of images stored in the storage unit 11. N or more images X are selected. At this time, a plurality of intermediate images X _i may be selected.
The N images X selected here may be temporarily stored in the storage unit 11. After selecting N images X, the process proceeds to an image recognition step (step S15). In this embodiment, image recognition is included in the distortion calculation step, but will be described separately for convenience of explanation.

In the image recognition step (step S15), the calculation unit 13 reads the image X _i and the image X _{i + 1} , compares these images, and detects the correlation between the image X _i and the image X _{i + 1} by image recognition. . At this time, if the image recognition fails, the process returns to the image selection step (step S14) to reselect the image. In this case, the image is reselected so that the time interval between the image X _i and the image X _{i + 1} is shortened so that the deformation amount of the mark is small. On the other hand, if the image recognition is successful, the strain calculation step (step S16) is then executed.

In the strain calculation step (step S16), the calculation unit 13 executes a strain calculation program on the image X _i and the image X _{i + 1} , thereby causing displacements (Δx, Δy in the X axis direction, the Y axis direction, and the XY direction). , Δxy) are calculated respectively. Subsequently, strains (εx, εy, εxy) in the X-axis direction, the Y-axis direction, and the XY direction are calculated using the calculated displacement amounts (Δx, Δy, Δxy).
The calculated strains (εx, εy, εxy) are stored in the memory 12 by the calculation unit 13. The data may be stored in the storage unit 11.
The above-described distortion calculation step (step S16) is repeated for all images selected in the image selection step (step S14), and the amount of movement of each area between the N images is calculated. Then, each strain of the measurement object is calculated by adding the calculated plurality of movement amounts. When each strain is calculated, the process proceeds to the output step (step S17).

  In the output step (step S <b> 17), the calculation unit 13 reads out each strain data table and each strain distribution diagram from the storage unit 11, transmits them to the output unit 15, and displays them on the monitor of the output unit 15.

According to the strain measurement system 1 having the above-described configuration, the surface of the measurement object 2 is N out of 3 or more images (where N is 3 or more) taken in time series from before deformation to after deformation. An image is selected, and among the N images, the image images X _i and X _{i + 1} that are adjacent in time series are compared, and the movement amount of each area is calculated. Therefore, the initial image X ₁ and a final image X _N can compare the images with each other smaller deformation amount than the amount of deformation of between, it can reliably perform image recognition since it does not change mark shape larger Become. And, since to calculate the strain of the measuring object by adding a plurality of movement amount from the initial image X ₁ to the last image X _N, we can improve the measurement accuracy of the strain at the measurement object 2.

Moreover, since the surface of the measurement object 2 is imaged by the imaging device 3 and the distortion is calculated based on the captured images Xia and 4b, the distortion can be calculated even if the measurement object 2 is away. That is, even when the measurement object 2 is present in a high temperature or a dangerous area, the strain of the measurement object 2 can be calculated.
And since the distance of the distance GL between grades can be changed arbitrarily, it becomes possible to adjust the measurement area (gauge length) for measuring distortion freely. Therefore, when it is desired to detect a strain in a minute section, the distance between the score points GL can be shortened, and when a strain in a long section is to be calculated, the distance between the score points GL can be increased. Thereby, the distortion of the calculation target area _Eij can be accurately measured.

  As mentioned above, although embodiment of this invention was described in detail, it cannot be overemphasized that this invention is not limited to this, In the range which does not deviate from the summary of this invention, various improvement and deformation | transformation may be performed.

  In addition, as a measuring object of the strain in each embodiment mentioned above, steel materials, concrete structures, bedrock, stone materials, gypsum, plastic materials, etc. are mentioned, for example. In particular, since it can be measured with high accuracy even in the measurement of strain accompanied by a large deformation, it can be suitably applied to elastic materials such as rubber.

1 strain measurement system 2 measurement object 3 imaging apparatus 3a camera body 3b lens 3c imaging element X ₁ initial image X _i intermediate image X _N final image 10 measuring device 11 storage unit 11a captured program 11b image selection program 11c strain calculation program 12 memory 13 Calculation unit 14 Input unit 15 Output unit 16 I / F unit 21 Strain measurement system E Area GL Inter-score distance S Measurement target area

Claims (5)

A strain measurement method for measuring the strain on the surface of a measurement object,
The surface of the measurement object in which a plurality of marks are irregularly arranged on the surface is imaged in time series from before deformation to after deformation, an initial image before deformation of the measurement object, An imaging step of acquiring at least three images including at least one intermediate image and a deformed final image;
An image selection step of selecting N images (where N is 3 or more) including the initial image, the intermediate image, and the final image acquired in the imaging step;
The position showing the best correlation with each area obtained by dividing the image X _i selected in the image selection step (where i is 1 to N−1) into a plurality of areas is identified from the image X _{i + 1} by image recognition. In addition, the movement amount of each area in the image X _i and the image X _{i + 1} is calculated, and a plurality of the movement amounts from the initial image to the final image are added to calculate the distortion of the measurement object. A strain measurement method comprising: a strain calculation step. When two or more intermediate images are acquired in the imaging step,
The strain measurement method according to claim 1, wherein in the image selection step, the intermediate image is selected corresponding to a time-series interval stored in advance. When two or more intermediate images are acquired in the imaging step,
At the strain calculation step, if the correlation between said respective areas with each area the image X _{i + 1} of the image X _i by the image recognition can not be obtained, the return to the image selection process image X _{i + 1} The strain measurement method according to claim 1 or 2, wherein an earlier image is selected.   In the strain calculation step, a calculation target area for calculating a strain is selected from each area of the image after deformation, and the two areas that are point-symmetric with a predetermined distance from the selected calculation target area as a central point The distance between the two areas after deformation is calculated on the basis of the respective displacement amounts in the calculation, and the distortion of the calculation target area is calculated. The strain measurement method described. A strain measurement system that measures strain on the surface of a measurement object,
An imaging unit that images the surface of the measurement object in which a plurality of marks are irregularly arranged on the surface;
A storage unit for storing at least three images including an initial image before deformation of the measurement object imaged by the imaging unit, at least one intermediate image being deformed, and a final image after deformation;
An image selection unit that selects N images (where N is 3 or more) including the initial image, the intermediate image, and the final image from at least three images stored in the storage unit;
A position showing the best correlation with each area obtained by dividing the image X _i (where i is 1 or more and N−1 or less) selected by the image selection unit into a plurality of areas is identified from the image X _{i + 1} by image recognition. In addition, the movement amount of each area in the image X _i and the image X _{i + 1} is calculated, and a plurality of the movement amounts from the initial image to the final image are added to calculate the distortion of the measurement object. A strain measurement system comprising a strain calculation unit.