A displacement measurement method and device based on grating image

Abstract

The present invention relates to a displacement measurement method and device based on a grating image, firstly, a template image is obtained; secondly, the template image is projected along the grating stripe direction to obtain a template vector; then, a grating image after the displacement of the object to be measured is collected, the grating image is projected along the grating stripe direction, a vector to be measured is constructed, and a normalized correlation coefficient between the vector to be measured and the template vector is calculated, wherein the pixel position corresponding to the maximum value of the correlation coefficient is the pixel position of the target feature after the displacement of the object to be measured; finally, the pixel position of the target feature after the displacement of the object to be measured is subtracted from the initial pixel position of the target feature to obtain the relative pixel displacement of the target feature, and based on the relative pixel displacement and the initial position of the object to be measured, the absolute position of the object to be measured after the displacement is obtained. The present invention only requires a grating light source and a camera for collecting grating images to perform displacement measurement, and is suitable for displacement measurement in a narrow space.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a displacement measurement method and device based on grating images. Background Art

[0002] The existing tools for measuring displacement using grating images are usually grating rulers. The basic principle is as follows: after the incident light beam passes through the grating, it is transmitted to the surface of the object to be measured, and the grating image of the object to be measured is obtained. When the object to be measured is displaced, the grating image will change, wherein the object to be measured refers to the target object whose displacement needs to be measured. Among them, the grating image can be obtained through reflection of the object to be measured, or it can be obtained through transmission of the object to be measured. When the object to be measured is displaced, the grating image will change. According to the change of the grating image, the displacement of the object to be measured can be obtained. Grating rulers can be divided into absolute grating rulers and incremental grating rulers. Although the absolute grating ruler can obtain the actual position, its manufacturing cost is high and the reading speed is slow, which is not as fast and stable as the incremental grating ruler. However, the incremental grating ruler has the problem of not being able to determine the initial position. When the power is off or it needs to be restarted due to a fault, the incremental grating ruler must be returned to zero. In addition, the weight and size of the grating ruler may not be suitable for displacement measurement in a small space. Summary of the invention

[0003] The present invention aims at the technical problems existing in the prior art and provides a displacement measurement method and device based on grating images. The purpose of the present invention is to calculate the relative displacement according to the position difference between the one-dimensional projection of the grating image and the one-dimensional projection of the grating image of the object to be measured with a calibrated position when the grating image moves. In this way, both the relative position and the absolute position can be obtained. At the same time, the method and device have the characteristics of fast speed, high precision and online measurement.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In one aspect, the present invention provides a displacement measurement method based on a grating image, comprising the following steps:

[0006] S1, obtaining a template image, including: calibrating the initial position of the object to be measured, and collecting a grating image of the object to be measured, wherein the grating image includes a target feature, and the target feature is a mark in the grating image that is different from the grating stripes, and obtaining the initial pixel position of the target feature;

[0007] S2, calculating the grayscale average value of each column of pixels in the template image in sequence along the grating stripe direction to construct a template vector;

[0008] S3, collecting a grating image of the object to be measured after displacement;

[0009] S4, calculating the grayscale average of each column of pixels in the grating image after the object to be measured is displaced along the grating stripe direction, and constructing a vector to be measured;

[0010] S5, calculating the normalized correlation coefficient between the vector to be measured and the template vector, wherein the pixel position corresponding to the maximum value of the normalized correlation coefficient is the pixel position of the target feature of the object to be measured after displacement;

[0011] S6, subtracting the pixel position of the target feature after the object to be measured is displaced from the initial pixel position of the target feature to obtain a relative pixel displacement of the target feature, and obtaining an absolute position of the object to be measured after the displacement based on the relative pixel displacement and the initial position of the object to be measured.

[0012] Furthermore, the step S3 further includes:

[0013] The grating image after the displacement of the object to be measured is uniformly divided into n sub-images along a direction perpendicular to the grating stripes, where n is a positive integer;

[0014] Steps S4-S5 are respectively performed on the n sub-images to obtain n normalized correlation coefficient maximum values ​​corresponding to the n sub-images, and the pixel position corresponding to the maximum value among the n normalized correlation coefficient maximum values ​​is the pixel position of the target feature after the object to be measured is displaced.

[0015] Furthermore, the method also includes preprocessing the template vector and / or the vector to be tested to eliminate vector distribution distortion caused by uneven grayscale.

[0016] Further, the preprocessing is shown as follows

[0017]

[0018] represents the vector before preprocessing, represents the vector after preprocessing, * represents the convolution operation, G represents the Gaussian kernel with a size of 3×T0, and T0 represents the grating stripe period.

[0019] Furthermore, step S5 also includes performing a detailed search on the pixel position of the target feature after the displacement of the object to be measured, so that the matching accuracy reaches the sub-pixel level, and the final pixel position of the target feature is obtained.

[0020] Furthermore, the method of refining the search includes enhancing the correlation coefficient or cubic spline interpolation.

[0021] Furthermore, the step S5 further includes: before calculating the normalized correlation coefficient between the vector to be measured and the template vector, performing offset subtraction on the template vector and the vector to be measured respectively.

[0022] Furthermore, the offset subtraction operation is obtained according to the following formula:

[0023]

[0024] In this formula, represents the vector obtained after offset subtraction, represents the one-dimensional vector constructed after calculating the grayscale average value of each column of pixels, c represents the number of columns of the one-dimensional vector, T0 represents the grating stripe period, k represents an integer, and ; express Before removal k The remaining elements after column T0, express After removal k The remaining elements after column T0.

[0025] Furthermore, the initial position of the target feature pixel can be obtained manually or calculated by template matching.

[0026] Furthermore, the grating stripe period can be obtained manually or by calculating the Fast Fourier Transformation (FFT) of the image.

[0027] Furthermore, the grayscale average value of each column of pixels is calculated as shown in the following formula:

[0028]

[0029] In this formula, Represents a one-dimensional vector constructed after calculating the grayscale average value of each column of pixels. r , c Respectively refer to the number of rows and columns of the image to be calculated, f ( i ,0) means the image is at coordinate ( i ,0) is the pixel gray value.

[0030] The beneficial effects of the present invention are as follows: (1) since the grating image is divided into n sub-images, if the grating image has stripe deformation and blurring caused by unevenness on the surface of the object to be measured, it can prevent the vector to be measured from being affected by these abnormal stripes during the subsequent projection operation, thereby improving the accuracy of displacement measurement;

[0031] (2) Since the grayscale average value of each column of pixels is calculated, the grating image is reduced from two dimensions to one dimension, so that the image-based displacement measurement method provided by the present invention has a faster computing speed; under 16G memory, Intel® Core™ i7-8700, and a CPU with a main frequency of 3.2GHz, the displacement time of a grating image of size 512×512 relative to a template image is determined by the method provided by the present invention to be less than 15ms, and online detection can be achieved;

[0032] (3) In the preferred embodiment of the present invention, cubic spline interpolation is used to perform a more refined position search, which can extend the search accuracy to sub-pixel accuracy. Combined with the design of the hardware optical path, the displacement detection accuracy can be improved to the micron level, which can meet the needs of small-range displacement detection.

[0033] (4) The present invention only requires a grating light source and a camera for collecting grating images to perform displacement measurement. Therefore, the size and weight of the device are relatively small compared to the grating ruler, and is suitable for displacement measurement in a small space.

[0034] In another aspect, the present invention provides a displacement measurement device based on a grating image, comprising:

[0035] The template image acquisition module is used to acquire the template image, including: calibrating the initial position of the object to be measured, and collecting the grating image of the object to be measured, wherein the grating image includes a target feature, and the target feature is a mark in the grating image that is different from the grating stripes, and acquiring the initial pixel position of the target feature;

[0036] A template vector construction module is used to calculate the grayscale average value of each column of pixels in the template image in sequence along the grating stripe direction to construct a template vector;

[0037] A grating image acquisition module for the object to be measured, used for acquiring a grating image of the object to be measured after displacement;

[0038] A module for constructing a vector to be measured, used to sequentially calculate the grayscale average value of each column of pixels in the grating image after the object to be measured is displaced along the grating stripe direction, and construct the vector to be measured;

[0039] A pixel position calculation module, used to calculate the normalized correlation coefficient between the vector to be measured and the template vector, wherein the pixel position corresponding to the maximum value of the normalized correlation coefficient is the pixel position of the target feature after the displacement of the object to be measured;

[0040] The displacement calculation module is used to subtract the pixel position of the target feature after the object to be measured is displaced from the initial pixel position of the target feature to obtain the relative pixel displacement of the target feature, and based on the relative pixel displacement and the initial position of the object to be measured, obtain the absolute position of the object to be measured after the displacement. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flow chart of a displacement measurement method based on a grating image provided by an embodiment of the present invention;

[0042] Figure 2 A raster image used in embodiments of the present invention; Figure 2 The mark 1 in the figure is the target feature in the grating image, and the mark 2 represents the grating fringe period;

[0043] Figure 3 The template image captured by the embodiment of the present invention;

[0044] Figure 4 A line graph drawn by a template vector calculated by an embodiment of the present invention;

[0045] Figure 5 The image to be tested is collected by the embodiment of the present invention;

[0046] Figure 6 The original vector to be tested and the preprocessed vector to be tested in the embodiment of the present invention;

[0047] Figure 7 A structural diagram of a displacement measurement device based on grating images provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0049] like Figure 1 As shown, the displacement measurement method based on grating image provided by the embodiment of the present invention comprises the following steps:

[0050] S1, obtain a template image, calibrate the initial position of the object to be tested, and collect a grating image of the object to be tested, the grating image includes a target feature, the target feature is a mark in the grating image that is different from the grating stripes, and obtain the initial pixel position P0 of the target feature. R0 and a grating stripe period T0; the template image can be obtained by intercepting grating stripes containing the target feature and its adjacent multiple periods in the grating image; the grating stripe period T0 refers to the pixel width of a black stripe and its adjacent white stripe; the target feature refers to a mark designed on the grating, and illustratively, the target feature is a specially designed black stripe that is wider (width is greater than the grating stripe period) in the grating image, located in the center of the grating, such as Figure 2 As shown, Figure 2 Mark 1 in the figure is the target feature in the grating image, and mark 2 represents the grating stripe period.

[0051] The initial pixel position P0 of the target feature can be obtained manually or calculated by template matching. The grating fringe period T0 can be obtained manually or by calculating the Fast Fourier Transformation (FFT) of the template image.

[0052] S2, calculate the grayscale average value of each column of pixels in the template image in turn along the grating stripe direction, that is, perform a projection operation on the template image to construct a one-dimensional vector .

[0053] The projection operation is shown as follows:

[0054]

[0055] In this formula, Represents a one-dimensional vector constructed after calculating the grayscale average value of each column of pixels. r , c Respectively refer to the number of rows and columns of the image to be calculated, f ( i ,0) means the image is at coordinate ( i ,0), that is, the grayscale value of the pixel at the ith row and the 0th column.

[0056] S3, collect the grating image of the object to be tested after displacement; project the grating image of the object to be tested after displacement to obtain the vector to be tested V, calculate the normalized correlation coefficient between the vector to be tested and the template vector, where the pixel position corresponding to the maximum value of the correlation coefficient is the pixel position of the target feature after the object to be tested is displaced P max .

[0057] The projection operation in step S3 is processed in the same manner as the projection operation in step S2, and the grayscale average value of each column of pixels in the grating image after the object to be measured is calculated.

[0058] Projection operation is performed on the template image and the grating image after the object to be measured is displaced, and the acquired two-dimensional image data is reduced to one dimension, which greatly improves the calculation speed of displacement measurement based on the grating image.

[0059] Under ideal conditions, the grating image will not have stripe deformation and blurring. However, due to the intensity of light and unevenness on the surface of the object to be measured, the grating image after displacement will have stripe deformation and blurring. Therefore, it is necessary to process the grating image after displacement and then determine the pixel position of the target feature. In this regard, the method of some embodiments of the present invention can improve the displacement measurement accuracy. The specific processing method is as follows:

[0060] The grating image after the displacement of the object to be measured is uniformly divided into n sub-images along a direction perpendicular to the grating stripes, where n is a positive integer;

[0061] After performing projection operations on the n sub-images respectively, the one-dimensional vector corresponding to each sub-image is obtained. , calculate the one-dimensional vector The normalized correlation coefficient of the template vector is obtained to obtain n normalized correlation coefficient maximum values ​​corresponding to the n sub-images, and the pixel position corresponding to the maximum value among the n normalized correlation coefficient maximum values ​​is the pixel position of the target feature of the object to be measured after displacement.

[0062] Preferably, in some embodiments, when constructing a template vector or a vector to be tested, after performing the projection operation, a preprocessing operation is further included on the projected vector. The preprocessing operation is shown in the following formula:

[0063]

[0064] represents the vector before preprocessing, represents the vector after preprocessing, * represents the convolution operation, G represents the Gaussian kernel with a size of 3×T0, and T0 represents the grating stripe period.

[0065] During the preprocessing process, the vector to be preprocessed is convolved to obtain the benchmark of the vector, and then the benchmark of the vector is subtracted from the vector to obtain a vector containing only the grating fringe information. The above method can eliminate the vector distribution distortion caused by uneven grayscale, avoid the influence of light source jitter, optical path hardware and other system noise, and improve the repeatability of the measurement system.

[0066] Preferably, in some embodiments, before calculating the normalized correlation coefficient between the vector to be measured and the template vector, offset subtraction is performed on the template vector and the vector to be measured respectively.

[0067] The offset subtraction operation is shown in the following formula:

[0068]

[0069] In this formula, represents the vector obtained after offset subtraction, Represents a one-dimensional vector constructed after calculating the grayscale average value of each column of pixels. c represents the number of columns of the one-dimensional vector, T0 represents the grating fringe period, k represents an integer, and ;in express Before removal k The remaining elements after column T0, express After removal k The remaining elements after T0 column. Further, k The preferred value is 1 or 2. By adopting the above method, the signal strength of the target feature can be highlighted, thereby improving the positioning accuracy.

[0070] S4, the pixel position of the target feature after the object to be tested is displaced P max Subtracting the initial pixel position P0 of the target feature, the relative pixel displacement of the target feature is obtained, and based on the relative pixel displacement and the initial position of the object to be measured, the absolute position of the object to be measured after displacement is obtained.

[0071] The relative pixel displacement is converted into relative physical displacement and then compared with the initial position P of the object to be measured. R0 After adding, the absolute position of the target feature after displacement is obtained. Among them, combined with the design of the hardware optical path, the actual physical size corresponding to a pixel can be obtained, so the relative pixel displacement can be converted into a relative physical displacement.

[0072] Preferably, in some embodiments, after obtaining the pixel position of the target feature after the displacement of the object to be measured P max Finally, the position is refined and searched to achieve sub-pixel matching accuracy, and the final pixel position of the target feature is obtained.

[0073] The refined search method can be implemented by using methods such as Enhanced Correlation Coefficient (ECC) or Cubic Spline Interpolation.

[0074] Using cubic spline interpolation for more detailed position search can expand the search accuracy to sub-pixel accuracy. Combined with the design of the hardware optical path, the displacement detection accuracy can be improved to the micron level, which can meet the needs of small-range displacement detection.

[0075] The above method is further described below with reference to specific examples.

[0076] (1) Collect a grating image with a calibrated position. For example, a grating image with 512×512 pixels, where the center of the grating image corresponds to the target feature, that is, the pixel position of the target feature P 0=256, initial position of the object to be measured P R0 =156.35 μm , grating fringe period T 0=6, that is, one grating stripe period contains 6 pixels, such as Figure 2As shown; a sub-image is selected near the pixel position of the target feature on the image as the template image, such as Figure 3 shown.

[0077] (2) Project the template image along the vertical direction to obtain a one-dimensional vector, and T 0=6 is used for offset subtraction to obtain the template vector, which is displayed in the form of a line graph as follows Figure 4 .

[0078] (3) Collect the grating image after the object to be measured is displaced, such as Figure 5 As shown, there is obvious grayscale unevenness in the image due to different light intensities.

[0079] (4) Divide the image into 4 sub-images by row, and project each sub-image into Figure 6 As shown in the left column, it can be seen that the grayscale distribution of the image is uneven; after preprocessing and offset subtraction operations, the vector to be tested is obtained and drawn into a line graph as shown Figure 6 Shown in the right column.

[0080] (5) Calculate the NCC of the template vector and the four vectors to be tested respectively, find their maximum values ​​and corresponding pixel positions, and obtain Table 1.

[0081] Table 1 NCC maximum value and corresponding pixel position

[0082]

[0083] (6) According to the data in Table 1, the pixel position corresponding to the maximum NCC value of vector 4 is selected as the initial point of the refined search, and the cubic spline interpolation method is used for further search to obtain the pixel position of the final target feature. P max =172.4.

[0084] (7) Obtain the relative pixel displacement of the target feature ΔP = P max - P 0 = 172.4 - 256 = -83.6 pixels. According to the optical path design of this embodiment, the actual physical size corresponding to one pixel is 1 pixel ≈1.4 μm , so the absolute position of the object to be measured after displacement can be calculated P R = P R0 + ΔP R = 156.35+(-83.6×1.4)=39.31μm.

[0085] Based on the above method, an embodiment of the present invention further provides a displacement measurement device based on grating images, such as Figure 7 As shown, including:

[0086] The template image acquisition module is used to acquire the template image, including: calibrating the initial position of the object to be measured, and collecting the grating image of the object to be measured, wherein the grating image includes a target feature, and the target feature is a mark in the grating image that is different from the grating stripes, and acquiring the initial pixel position P0 of the target feature;

[0087] The template vector construction module is used to calculate the grayscale average value of each column of pixels in the template image in turn along the grating stripe direction to construct the template vector ;

[0088] A grating image acquisition module for the object to be measured, used for acquiring a grating image of the object to be measured after displacement;

[0089] A module for constructing a vector to be measured, used to sequentially calculate the grayscale average value of each column of pixels in the grating image after the object to be measured is displaced along the grating stripe direction, and construct the vector to be measured;

[0090] A pixel position calculation module, used to calculate the normalized correlation coefficient between the vector to be measured and the template vector, wherein the pixel position corresponding to the maximum value of the normalized correlation coefficient is the pixel position of the target feature after the displacement of the object to be measured;

[0091] The displacement calculation module is used to calculate the pixel position of the target feature after the object to be measured is displaced. P max Subtracting the initial pixel position P0 of the target feature, the relative pixel displacement of the target feature is obtained, and based on the relative pixel displacement and the initial position of the object to be measured, the absolute position of the object to be measured after displacement is obtained.

[0092] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0093] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A displacement measurement method based on grating image, characterized in that: The following steps are involved: S1, obtaining a template image, including: calibrating the initial position of the object to be measured, and collecting a grating image of the object to be measured, wherein the grating image includes a target feature, and the target feature is a mark in the grating image that is different from the grating stripes, obtaining the initial pixel position of the target feature, and obtaining the template image by intercepting the grating stripes containing the target feature and its adjacent multiple periods in the grating image; S2, calculating the grayscale average value of each column of pixels in the template image in sequence along the grating stripe direction to construct a template vector; S3, collecting a grating image of the object to be measured after displacement; S4, calculating the grayscale average of each column of pixels in the grating image after the object to be measured is displaced along the grating stripe direction, and constructing a vector to be measured; S5, calculating the normalized correlation coefficient between the vector to be measured and the template vector, wherein the pixel position corresponding to the maximum value of the normalized correlation coefficient is the pixel position of the target feature of the object to be measured after displacement; S6, subtracting the pixel position of the target feature after the object to be measured is displaced from the initial pixel position of the target feature to obtain a relative pixel displacement of the target feature, and obtaining an absolute position of the object to be measured after the displacement based on the relative pixel displacement and the initial position of the object to be measured.

2. The method according to claim 1, characterized in that The step S3 further comprises: The grating image after the displacement of the object to be measured is uniformly divided into n sub-images along a direction perpendicular to the grating stripes, where n is a positive integer; Steps S4-S5 are respectively performed on the n sub-images to obtain n normalized correlation coefficient maximum values ​​corresponding to the n sub-images, and the pixel position corresponding to the maximum value among the n normalized correlation coefficient maximum values ​​is the pixel position of the target feature after the object to be measured is displaced.

3. The method according to claim 1, characterized in that Also includes: The template vector and / or the vector to be tested are preprocessed to eliminate the vector distribution distortion caused by grayscale unevenness.

4. The method according to claim 3, characterized in that The preprocessing is shown in the following formula: represents the vector before preprocessing, represents the vector after preprocessing, * represents the convolution operation, G represents the Gaussian kernel with a size of 3×T0, and T0 represents the grating stripe period.

5. The method according to claim 1, characterized in that Step S5 also includes: performing a detailed search on the pixel position of the target feature after the displacement of the object to be measured, so that the matching accuracy reaches the sub-pixel level, and obtaining the final pixel position of the target feature.

6. The method according to claim 5, characterized in that The method of refining the search includes enhancing the correlation coefficient or cubic spline interpolation.

7. The method according to claim 1, characterized in that The step S5 further includes: before calculating the normalized correlation coefficient between the vector to be measured and the template vector, performing offset subtraction on the template vector and the vector to be measured respectively.

8. The method according to claim 7, characterized in that Subtracting the misalignment, we get: In this formula, represents the vector obtained after offset subtraction, represents the one-dimensional vector constructed after calculating the grayscale average value of each column of pixels, c represents the number of columns of the one-dimensional vector, T0 represents the grating fringe period, k represents an integer, and ; express Before removal k The remaining elements after column T0, express After removal k The remaining elements after column T0.

9. A displacement measurement device based on grating image, characterized in that: include: The template image acquisition module is used to acquire the template image, including: calibrating the initial position of the object to be tested, and collecting the grating image of the object to be tested, wherein the grating image includes a target feature, and the target feature is a mark in the grating image that is different from the grating stripes, and acquiring the initial pixel position of the target feature, and obtaining the template image by intercepting the grating stripes containing the target feature and its adjacent multiple periods in the grating image; A template vector construction module is used to calculate the grayscale average value of each column of pixels in the template image in sequence along the grating stripe direction to construct a template vector; A grating image acquisition module for the object to be measured, used for acquiring a grating image of the object to be measured after displacement; A module for constructing a vector to be measured, used to sequentially calculate the grayscale average value of each column of pixels in the grating image after the object to be measured is displaced along the grating stripe direction, and construct the vector to be measured; A pixel position calculation module, used to calculate the normalized correlation coefficient between the vector to be measured and the template vector, wherein the pixel position corresponding to the maximum value of the normalized correlation coefficient is the pixel position of the target feature after the displacement of the object to be measured; The displacement calculation module is used to subtract the pixel position of the target feature after the object to be measured is displaced from the initial pixel position of the target feature to obtain the relative pixel displacement of the target feature, and based on the relative pixel displacement and the initial position of the object to be measured, obtain the absolute position of the object to be measured after the displacement.

Images