A high-precision linear displacement measurement method and system

Abstract

The present invention discloses a high-precision linear displacement measurement method and system. The method includes: creating an information strip on the surface of the target to be measured, acquiring and calibrating a sequence of images to obtain a reference image; calibrating the reference image to obtain the pixel position of the calibration image and the scale factor of the actual physical position; acquiring a real-time measurement image during the measurement process and coarsely positioning its position on the reference image; finely registering the roughly positioned real-time measurement image, combining the pixel position calculation of the calibration image to obtain the pixel position of the real-time measurement image; acquiring a measurement image of the target to be measured, and calculating the position of the measurement image of the target to be measured based on the pixel position of the real-time measurement image and the scale factor of the actual physical position. The system includes: an acquisition module, a calibration module, a calibration module, a coarse positioning module, a fine registration module, and a calculation module. By using the present invention, linear displacement can be accurately measured.

Description

Technical Field

[0001] The present invention relates to the technical field of displacement measurement, and in particular to a high-precision linear displacement measurement method and system thereof. Background Art

[0002] High-precision displacement measurement technology is widely used in important fields such as semiconductor processing, precision machinery manufacturing, aerospace, servo systems, and biomedicine. High-precision displacement measurement not only has high measurement resolution and accuracy, reaching nanometer and sub-nanometer accuracy; it also has a long measurement stroke and a wide range. Therefore, the demand for high-precision and large-scale displacement measurement is becoming increasingly prominent.

[0003] Traditional displacement measurement uses grating displacement measurement technology, with the grating pitch as the measurement reference, and its measurement signal is a digital pulse. Currently, grating scale displacement sensors are widely used in the closed-loop servo system of CNC machine tools and can be used to detect linear displacement or angular displacement. Although grating scales have the characteristics of large detection range, high detection accuracy, fast response speed, and good measurement repeatability, their complex structure and high price lead to increased usage costs. In addition, the actual measurement accuracy will be affected by various factors and cannot meet the needs.

[0004] For example, Chinese patent document No. CN202111370430.9 discloses a spliced ​​absolute linear displacement sensor based on the combined modulation principle, which realizes ultra-large-scale absolute linear displacement measurement with a measurement range of several meters or even tens of meters, breaking the limitations of traditional scale manufacturing processes on increasing the measuring range, and has flexible application scenarios. However, this method uses absolute measurement technology, and its measurement accuracy will inevitably be affected by various factors such as installation error, light source error, and scale error; Chinese patent document No. CN202010706819.5 discloses a high-precision translation measurement solution based on image vision technology, which uses dual-camera recognition technology and combines coarse and fine registration to improve the speed and accuracy of image registration. However, this method requires two cameras to work, the system is complex and the cost is high; Chinese patent document No. CN202210024968.2 discloses a linear displacement measurement system and measurement method based on an image grating. During the measurement, a glass linear scale is placed on an LED backlight source, and is allowed to perform translational motion with the linear displacement slide in the displacement drive module. The image acquisition module is used to collect stripe images at different positions, and then the data processing and display module is used to perform phase correlation on the images before and after the translation, and finally the actual displacement value of the displacement stage is obtained. Although this method solves the problem of insufficient installation flexibility of traditional grating sensors, in actual applications, the lines in the glass linear scale will be blocked by dust or stained by oil stains, which seriously affects its measurement accuracy. In addition, the process of making the lines is complex and the cost of use is high. Summary of the Invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a high-precision linear displacement measurement method and system thereof, which can accurately measure linear displacement.

[0006] The first technical solution adopted by the present invention is: a high-precision linear displacement measurement method, comprising the following steps:

[0007] Making information strips on the surface of the target to be measured and acquiring sequential images with overlapping areas;

[0008] Calibrate the sequence images to obtain the reference image;

[0009] Calibrate the reference image to obtain the ratio coefficient between the pixel position of the calibration image and the actual physical position;

[0010] During the measurement process, a real-time measurement image is acquired, and the position of the real-time measurement image on the reference image is roughly located. Combined with the pixel position of the calibration image, the pixel offset between the reference image and the real-time measurement image is obtained;

[0011] Perform fine registration on the real-time measurement image after rough positioning, and calculate the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image;

[0012] The measurement image of the target to be measured is obtained, and the measurement image position of the target to be measured is calculated based on the pixel position of the real-time measurement image and the actual physical position ratio coefficient.

[0013] Furthermore, the information strip includes a texture area, a coding area and a region boundary, wherein the texture area is used to provide sufficient image information for high-precision image registration; the coding area is used to identify coding marks of different regions; and the region boundary is used to define the boundaries of different regions.

[0014] Furthermore, the step of calibrating the sequence of images to obtain a reference image specifically includes:

[0015] Determine the overlapping area of ​​adjacent images in the sequence of images, and perform distortion correction on the adjacent images according to the overlapping area of ​​the adjacent images;

[0016] The sequence images after image correction are stitched together to obtain a reference image.

[0017] Furthermore, the step of calibrating the reference image to obtain a ratio coefficient between the pixel position of the calibrated image and the actual physical position specifically includes:

[0018] Select reference images with a certain time interval and partial overlap as calibration images;

[0019] Select multiple sub-regions from the overlapping region and estimate the pixel offset of each sub-region relative to the calibration image to obtain the pixel position of the calibration image;

[0020] Obtaining the pixel movement distance of the calibration image according to the pixel position of the calibration image;

[0021] The actual movement distance of the calibration image is obtained and combined with the pixel movement distance of the calibration image to calculate the actual physical position scale coefficient.

[0022] Furthermore, the step of acquiring a real-time measurement image during the measurement process, roughly locating the position of the real-time measurement image on the reference image, and obtaining a pixel offset between the reference image and the real-time measurement image in combination with the pixel position of the calibration image specifically includes:

[0023] Acquire real-time measurement images during the measurement process;

[0024] Locate textured and coded regions using real-time measurements of region boundaries in images;

[0025] Obtain the rough pixel position of the real-time measurement image according to the coding area;

[0026] The pixel offset between the reference image and the real-time measurement image is obtained by combining the rough pixel position of the real-time measurement image and the pixel position of the calibration image.

[0027] Furthermore, the step of performing fine registration on the real-time measurement image after the rough positioning and calculating the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image specifically includes:

[0028] Constructing the geometric relationship between the real-time measurement image and the reference image based on the image transformation model;

[0029] Calculating the displacement parameters of the real-time measurement image and the reference image based on the geometric relationship between the real-time measurement image and the reference image based on the image matching algorithm to obtain the displacement parameters of the real-time measurement image and the reference image;

[0030] The pixel position of the real-time measurement image is obtained by combining the pixel offset between the reference image and the real-time measurement image and the displacement parameter calculation between the real-time measurement image and the reference image.

[0031] Furthermore, the calculation formula for the pixel position of the real-time measurement image is as follows:

[0032] L i =T d +t di +δc;

[0033] In the above formula, i is the i-th real-time measurement image, d is the d-th reference image closest to the i-th real-time measurement image, Td is the pixel position of the reference image, t di is the pixel offset between the reference image and the real-time measurement image, and δc is the incremental term of the real-time measurement image displacement parameter.

[0034] Furthermore, the calculation formula for measuring the image position of the target to be measured is as follows:

[0035] D=ρ z (L-L0)+D0;

[0036] In the above formula, ρ z is the actual physical position scale coefficient of the zth part of the measurement range, L is the pixel position of the measurement image of the target to be measured, L0 is the starting position of the measurement image of the target to be measured, and D0 is the starting position of the measurement range section of the target to be measured.

[0037] Furthermore, the coarse positioning of the real-time measurement image on the reference image further includes:

[0038] An oblique line or a curved line is added to the surface of the information strip, and the rough pixel position of the real-time measurement image is estimated according to the intersection position of the oblique line or the curved line and the center line of the real-time measurement image.

[0039] The second technical solution adopted by the present invention is: a high-precision linear displacement measurement system, comprising:

[0040] An acquisition module, used for making information strips on the surface of the target to be measured and acquiring a sequence of images with overlapping areas;

[0041] A calibration module is used to calibrate the sequence images to obtain a reference image;

[0042] A calibration module is used to calibrate the reference image to obtain the pixel position of the calibration image and the ratio coefficient of the actual physical position;

[0043] The coarse positioning module is used to obtain the real-time measurement image during the measurement process, and to coarsely locate the position of the real-time measurement image on the reference image. The pixel offset between the reference image and the real-time measurement image is obtained by combining the pixel position of the calibration image.

[0044] The fine registration module is used to perform fine registration on the real-time measurement image after rough positioning, and calculate the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image;

[0045] The calculation module is used to obtain the measurement image of the target to be measured, and calculate the measurement image position of the target to be measured based on the pixel position of the real-time measurement image and the actual physical position ratio coefficient.

[0046] The beneficial effects of the method and system of the present invention are as follows: first, the present invention produces an information strip on the surface of the target to be measured and obtains a sequence of images with overlapping areas, calibrates the sequence of images, and obtains a reference image; secondly, the reference image is calibrated to obtain a pixel position of the calibration image and a ratio coefficient of an actual physical position; then, a real-time measurement image in the measurement of the target to be measured is obtained, and the real-time measurement image is coarsely positioned, and the pixel position of the calibration image is combined with the pixel position calculation to obtain a pixel offset between the reference image and the real-time measurement image; then, the real-time measurement image after the coarse positioning is finely aligned, and the pixel offset calculation of the reference image and the real-time measurement image is combined to make the pixel position of the obtained real-time measurement image more accurate; finally, a measurement image of the target to be measured is obtained. Since the real-time measurement image is an image obtained during the measurement of the target to be measured, the measurement image position of the target to be measured can be calculated based on the pixel position of the real-time measurement image and the ratio coefficient of the actual physical position. The method makes the linear displacement measurement more accurate, solves the problem of low working efficiency of the existing method, and has the advantages of high execution efficiency, good real-time performance, simplicity and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of the steps of a high-precision linear displacement measurement method of the present invention;

[0048] Figure 2 This is a structural block diagram of a high-precision linear displacement measurement system of the present invention;

[0049] Figure 3 is a schematic diagram of an information bar according to a specific embodiment of the present invention;

[0050] Figure 4 is a schematic diagram of a reference image according to a specific embodiment of the present invention;

[0051] Figure 5 2 is a schematic diagram of pixel position estimation of a calibration image according to a specific embodiment of the present invention;

[0052] Figure 6 This is a schematic diagram of the first real-time measurement image coarse positioning process according to a specific embodiment of the present invention;

[0053] Figure 7 This is a schematic diagram of a second real-time measurement image coarse positioning process according to a specific embodiment of the present invention;

[0054] Figure 8 It is a schematic diagram of the relationship between the reference image and the real-time measurement image in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0055] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are provided for ease of description only and do not limit the order of the steps. The order of execution of the steps in the embodiments can be adaptively adjusted based on the understanding of those skilled in the art.

[0056] Reference Figure 1 The present invention provides a high-precision linear displacement measurement method, which includes the following steps:

[0057] S1. Making information strips on the surface of the target to be measured and obtaining a sequence of images with overlapping areas;

[0058] S1.1. Make an information strip on the surface of the target to be measured;

[0059] Specifically, such as Figure 3 As shown, the information strip produced by the preferred embodiment has a measuring length of 300 mm, and the information strip includes three parts: one part is a texture area, which is used to provide sufficient image information for high-precision image registration; one part is a binary coding area, which is used to identify coding marks of different areas; the last part is a region boundary part, which is used to define the boundaries of different areas; the coding area is used as the scale line of the information strip, wherein the coding adopts a ten-bit binary coding, each column of the coding has 10 blocks, each block can be black or white, black represents 0, and white represents 1, and 1024 scales can be represented by binary, and it is ensured that there is at least 1 complete mark in the image when imaging the surface of the object to be measured.

[0060] The preferred measurement tool used in this embodiment is an LSM translation stage. The information bar on the LSM translation stage can be placed in the middle, edge, or other positions according to actual conditions. This method is highly flexible and has wide applicability. Furthermore, since the information bar identifier can be random speckle information or a number, a QR code, or other mark, the measurement range of the information bar can be changed according to the actual range of the translation stage.

[0061] S1.2. Acquire sequential images with overlapping regions;

[0062] Specifically, a single linear array camera is used to capture the inherent features or fabricated texture features of the surface of the target to be measured, thereby obtaining a sequence of images.

[0063] S2, calibrate the sequence images to obtain a reference image;

[0064] Specifically, in order to obtain a complete reference image, the sequence image is made to include images of each part of the information strip, so the obtained sequence image has a certain degree of overlap. Then, the overlapping area of ​​adjacent images in the sequence image is determined, and the adjacent images are distorted and corrected according to the overlapping area of ​​the adjacent images. Finally, the grayscale level of the image pixel at the splicing point is determined and the corrected adjacent images are spliced ​​based on the fusion algorithm to obtain the reference image.

[0065] Since the precise position of each reference image within the entire measurement range is obtained by splicing the information strips of the sequence images, higher measurement accuracy can be obtained.

[0066] Among them, the selection relationship between the reference image and the sequence image is as follows: Figure 4 As shown, a reference image is selected from the sequence of images, and the pixel position of the dth reference image is described as T d (0≤d≤M). The parameter d can be described as:

[0067] d=i / α(0≤i≤N);

[0068] In the above formula, i is the reference image index, N is the number of reference images, and α is a positive integer greater than 1. For example, when α=2 is set, it means that a reference image is selected from every two sequential images.

[0069] S3. Calibrate the reference image to obtain a ratio coefficient between the pixel position of the calibration image and the actual physical position;

[0070] S3.1. Selecting reference images with a certain time interval and partial overlap as calibration images;

[0071] Specifically, reference images with a certain time interval and partial overlap are selected as calibration images, and the overlapping area of ​​adjacent calibration images is ensured to be larger than the calculation window of image registration.

[0072] S3.2. Select multiple sub-regions from the overlapping region and estimate the pixel offset of each sub-region relative to the calibration image to obtain the pixel position of the calibration image;

[0073] Specifically, such as Figure 5 As shown, it is assumed that the reference image moves along the x direction with the slide rail, and the reference image is captured within the measurement range at a certain interval, and the global coordinate x of its center point is expressed as T i (0≤i≤N), select the i-th reference image and the (i+1)-th reference image as the calibration image, in the overlapping area of ​​the i-th reference image and the (i+1)-th reference image, obtain the two calibration images through the ICGN algorithm, select several sub-regions, and estimate the pixel offset of each sub-region relative to the calibration image The calculation formula is as follows:

[0074]

[0075] In the above formula, j is the sub-region index, R is the sub-region number, is the x-coordinate of the j-th subregion in the i-th calibration image, is the x-coordinate of the j-th sub-region in the (i+1)-th calibration image.

[0076] Thus, the pixel offset E between the i-th calibration image and the (i+1)-th calibration image is obtained. i,i+1 The calculation formula is estimated to be as follows:

[0077]

[0078] Assume T0 = 0, according to the pixel offset E of the calibration image i,i+1 , the pixel positions of all calibration images are obtained as:

[0079] T i =T i-1 +E i-1,i (1≤i≤N);

[0080] These positions are used as the initial values ​​of , and the exact value of is calculated by minimizing the energy equation as follows:

[0081]

[0082] In the above formula, s is the image index that has an overlapping area with the i-th calibration image, and S is the number of images that have an overlapping area with the i-th calibration image.

[0083] S3.3, obtaining the pixel movement distance of the calibration image according to the pixel position of the calibration image;

[0084] Specifically, the calculation formula for the pixel movement distance of the calibration image is as follows:

[0085] T ij =T j -T i ;

[0086] In the above formula, T i is the pixel position of the i-th calibration image, T j is the pixel position of the jth calibration image.

[0087] S3.4. Obtain the actual movement distance of the calibration image and calculate the actual physical position scale coefficient based on the pixel movement distance of the calibration image.

[0088] Specifically, laser interferometer is used to measure D ij From the calibration image Ti To the calibration image T j The actual moving distance is used to obtain the actual physical position proportional coefficient ρ Z :

[0089] ρ z =D ij / T ij ,(z=0,1,...,Z);

[0090] In the above formula, ρ z is the actual physical position scale factor of the z-th part of the measuring range. z means dividing the measuring range into z parts and obtaining a scale factor for each part to reduce the cumulative error. The number of sections can be determined according to the measuring range. For example, when the measuring range is narrow, select the entire range as a part and use a unique scale factor.

[0091] S4. During the measurement process, a real-time measurement image is acquired, and a position of the real-time measurement image on the reference image is roughly located. In combination with the pixel position of the calibration image, a pixel offset between the reference image and the real-time measurement image is obtained;

[0092] S4.1. During the measurement process, obtain a real-time measurement image and use the region boundaries in the real-time measurement image to locate the texture region and the coding region. Quickly locate the coding mark through the coding region, and use the coding mark recognition to quickly and roughly locate the information strip.

[0093] Specifically, such as Figure 6 As shown, in the positioning process, the section boundary is first detected, a complete section is found in the real-time image, the code k is identified from the coding mark of the section based on both sides of the section, the pixel grayscale values ​​are accumulated in the area along the y direction of the gap section area between the code and the texture, and the section boundary is detected according to the data curve.

[0094] S4.2. Obtaining a rough pixel position of the real-time image according to the encoded region;

[0095] Specifically, according to the code k and its global pixel position c k , estimate the rough pixel position L of the i-th real-time measurement image i .

[0096] Among them, code k is the identification code k in the cross-section coding mark obtained by real-time measurement image, c k That is, the position of code k in the global pixel, which corresponds to the rough pixel position L of the i-th real-time measurement image i .

[0097] In addition, there is not only one way to measure the coarse positioning of the image in real time, such as Figure 7As shown, a slant line or curve can be added to the texture area of ​​the information bar to replace the coding mark method, and the cross-sectional area is obtained according to its global pixel position c k , with c k Draw a central axis, add a slant line or a curve represented by a sin function in the cross-sectional area, and obtain the intersection point of the central axis to estimate the rough pixel position L of the i-th real-time measurement image. i .

[0098] It should be noted that when the position of the real-time measurement image is different, the height of the corresponding oblique line point is different, which can be used to perform coarse positioning of the real-time measurement image position.

[0099] Assume that the intersection point P of the oblique line and the central axis is (x, y), the measurement range width is W, the height is H, the real-time measurement image width is w, the height is h, and the angle between the oblique line and the horizontal line is θ. The corresponding relationship is as follows:

[0100]

[0101] In the above formula, tanθ is W0 is the distance between the starting boundary of the real-time measurement image and the starting boundary of the measurement range, and H0 is the distance between the intersection point P and the upper boundary of the measurement range. Is equivalent to .

[0102] S4.3. Calculate the pixel offset between the reference image and the real-time measurement image by combining the rough pixel position of the real-time measurement image and the pixel position of the calibration image.

[0103] Specifically, select a reference image I that is closest to the i-th real-time measurement image d , based on the rough pixel position L of the real-time measurement image i and the pixel position T in the selected reference image d , we can get the pixel offset between the real-time measurement image and the reference image:

[0104] t di =T d -L i .

[0105] It should be noted that the calibration image obtained by calibrating the reference image in step S3 is equivalent to a temporary variable. The pixel position of the reference image is obtained through calibration. At this time, the reference image to be calibrated is called the calibration image, and the pixel position of the obtained calibration image is called the pixel position of the reference image.

[0106] S5. Perform fine registration on the real-time measurement image after rough positioning, combine the pixel offset calculation of the reference image and the real-time measurement image, and obtain the pixel position of the real-time measurement image. Compared with the existing method of traversing and searching the image, this method achieves high-precision positioning of the real-time measurement image by matching the real-time image with the reference image, solving the problem of low working efficiency of the existing method, and having the advantages of high execution efficiency, good real-time performance, simplicity and reliability;

[0107] S5.1. Constructing the geometric relationship between the real-time measurement image and the reference image based on the image transformation model;

[0108] Specifically, such as Figure 8 As shown, the marked area of ​​the reference image is named R(x, y), and the real-time measurement image is named I(x, y). The preferred image registration algorithm of this embodiment is the inverse synthesis matching strategy combined with the Gauss-Newton (ICGN) algorithm proposed by Baker and Matthews, which has an excellent accuracy of 0.001 pixels. The ICGN algorithm generally uses an affine transformation model to represent the geometric changes between the two images. However, for LSM translation stage measurement, the geometric relationship between the marked area and the real-time measurement image is a simple one-dimensional displacement, which can be described by no more than two parameters. The specific geometric relationship is as follows:

[0109]

[0110] In the above formula, c and f are two parameters for real-time measurement of image displacement, which are expressed as s(c,f). Then the marked area is expressed as:

[0111]

[0112] S5.2. Calculate displacement parameters between the real-time measurement image and the reference image based on a geometric relationship between the real-time measurement image and the reference image using an image matching algorithm to obtain the displacement parameters between the real-time measurement image and the reference image;

[0113] In order to obtain the correct displacement parameters between the real-time measurement image and the reference image, the cost function must meet its minimum value, which is expressed as follows:

[0114]

[0115] Using image matching algorithm:

[0116]

[0117] In the above formula, s i is the displacement parameter of the i-th real-time measurement image, Q is the energy, is the partial derivative of energy.

[0118] Furthermore, the result of R(x,y,s)-I(x,y) is:

[0119]

[0120] In the above formula, is the displacement parameter change, is the partial derivative of the displacement parameter.

[0121] Furthermore, we can obtain:

[0122]

[0123] and The transformation parameters can be calculated using the image matching algorithm, and the expression is:

[0124]

[0125] In the above formula, H is the Hessian matrix, which is expressed as:

[0126]

[0127] From the geometric relationship formula between the marked area and the real-time measurement image, we can get And apply it to the real-time measurement image I(x,y), and obtain a new real-time measurement image through parameter transformation Then repeat the above process until the error is lower than the expected error value ε. The formula is as follows:

[0128]

[0129] Since the marked area remains unchanged throughout the process, and H -1 The value of also remains unchanged and does not need to be recalculated.

[0130] Since the translation motion of this embodiment only includes two translation parameters, the six affine parameters of the image matching algorithm are reduced to two translation parameters through the above steps, and the solution is based on the backward Gauss-Newton iteration method, thereby obtaining a translation image matching method with higher accuracy, faster speed and better reliability.

[0131] S5.3. Calculate the pixel offset between the reference image and the real-time measurement image and the displacement parameter between the real-time measurement image and the reference image to obtain the pixel position of the real-time measurement image.

[0132] Specifically, the value obtained due to the precise image displacement is an incremental term based on the rough position t di , so the final position of the image mark area in real-time measurement is:

[0133] L i =T d +t di +δc;

[0134] In the above formula, i is the i-th real-time measurement image, d is the d-th reference image closest to the i-th real-time measurement image, T d is the pixel position of the reference image, t di is the pixel offset between the reference image and the real-time measurement image, and δc is the incremental term of the real-time measurement image displacement parameter.

[0135] S6. Obtain a measurement image of the target to be measured, and calculate the measurement image position of the target to be measured based on the pixel position of the real-time measurement image and the actual physical position ratio coefficient.

[0136] Specifically, the calculation formula for the target position to be measured is as follows:

[0137] D=ρ z (L-L0)+D0;

[0138] In the above formula, ρ z is the actual physical position scale coefficient of the zth part of the measurement range, L is the pixel position of the measurement image of the target to be measured, L0 is the starting position of the measurement image of the target to be measured, and D0 is the starting position of the measurement range section of the target to be measured.

[0139] Among them, since the real-time measurement image is the image obtained during the measurement process of the target, z represents dividing the measurement range into Z parts, and obtaining a proportional coefficient for each part, z = 0, 1, ..., Z, so ρ z It can be selected according to the position of the measurement image of the target to be measured. L can be calculated based on the calculation formula of the pixel position of the real-time measurement image. L0 is the center point position of the measurement image of the target to be measured in the cross-section coding area, and D0 is the coarse positioning part of the image. During the measurement process, the cross-section boundary is first detected, and a complete cross-section is found in the measurement image of the target to be measured, and the cross-section is used as the starting position.

[0140] like Figure 2 As shown, a high-precision linear displacement measurement system includes:

[0141] An acquisition module, used for making information strips on the surface of the target to be measured and acquiring a sequence of images with overlapping areas;

[0142] A calibration module is used to calibrate the sequence images to obtain a reference image;

[0143] A calibration module is used to calibrate the reference image to obtain the pixel position of the calibration image and the ratio coefficient of the actual physical position;

[0144] The coarse positioning module is used to obtain the real-time measurement image during the measurement process, and to coarsely locate the position of the real-time measurement image on the reference image. The pixel offset between the reference image and the real-time measurement image is obtained by combining the pixel position of the calibration image.

[0145] The fine registration module is used to perform fine registration on the real-time measurement image after rough positioning, and calculate the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image;

[0146] The calculation module is used to obtain the measurement image of the target to be measured, and calculate the measurement image position of the target to be measured based on the pixel position of the real-time measurement image and the actual physical position ratio coefficient.

[0147] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0148] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A high-precision linear displacement measurement method, characterized in that: The following steps are involved: Making information strips on the surface of the target to be measured and acquiring sequential images with overlapping areas; Calibrate the sequence images to obtain the reference image; Calibrate the reference image to obtain the ratio coefficient between the pixel position of the calibration image and the actual physical position; During the measurement process, a real-time measurement image is acquired, and the position of the real-time measurement image on the reference image is roughly located. Combined with the pixel position of the calibration image, the pixel offset between the reference image and the real-time measurement image is obtained; Perform fine registration on the real-time measurement image after rough positioning, and calculate the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image; The measurement image of the target to be measured is obtained, and the measurement image position of the target to be measured is calculated based on the pixel position of the real-time measurement image and the actual physical position ratio coefficient.

2. A high-precision linear displacement measurement method according to claim 1, characterized in that: The information strip includes a texture area, a coding area and a region boundary. The texture area is used to provide sufficient image information for high-precision image registration; the coding area is used to identify coding marks of different areas; and the region boundary is used to define the boundaries of different areas.

3. A high-precision linear displacement measurement method according to claim 1, characterized in that: The step of calibrating the sequence images to obtain the reference image specifically includes: Determine the overlapping area of ​​adjacent images in the sequence of images, and perform distortion correction on the adjacent images according to the overlapping area of ​​the adjacent images; The sequence images after image correction are stitched together to obtain a reference image.

4. A high-precision linear displacement measurement method according to claim 1, characterized in that: The step of calibrating the reference image to obtain a ratio coefficient between the pixel position of the calibration image and the actual physical position specifically includes: Select reference images with a certain time interval and partial overlap as calibration images; Select multiple sub-regions from the overlapping region and estimate the pixel offset of each sub-region relative to the calibration image to obtain the pixel position of the calibration image; Obtaining the pixel movement distance of the calibration image according to the pixel position of the calibration image; The actual movement distance of the calibration image is obtained and combined with the pixel movement distance of the calibration image to calculate the actual physical position scale coefficient.

5. A high-precision linear displacement measurement method according to claim 2, characterized in that: The step of acquiring a real-time measurement image during the measurement process, roughly locating the position of the real-time measurement image on the reference image, and obtaining a pixel offset between the reference image and the real-time measurement image in combination with the pixel position of the calibration image specifically includes: Acquire real-time measurement images during the measurement process; Locate textured and coded regions using real-time measurements of region boundaries in images; Obtain the rough pixel position of the real-time measurement image according to the coding area; The pixel offset between the reference image and the real-time measurement image is obtained by combining the rough pixel position of the real-time measurement image and the pixel position of the calibration image.

6. A high-precision linear displacement measurement method according to claim 1, characterized in that: The step of performing fine registration on the real-time measurement image after the rough positioning and calculating the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image specifically includes: Constructing the geometric relationship between the real-time measurement image and the reference image based on the image transformation model; Calculating the displacement parameters of the real-time measurement image and the reference image based on the geometric relationship between the real-time measurement image and the reference image based on the image matching algorithm to obtain the displacement parameters of the real-time measurement image and the reference image; The pixel position of the real-time measurement image is obtained by combining the pixel offset between the reference image and the real-time measurement image and the displacement parameter calculation between the real-time measurement image and the reference image.

7. A high-precision linear displacement measurement method according to claim 6, characterized in that: The calculation formula for the pixel position of the real-time measurement image is as follows: L i =T d +t di +δc; In the above formula, i is the i-th real-time measurement image, d is the d-th reference image closest to the i-th real-time measurement image, T d is the pixel position of the reference image, t di is the pixel offset between the reference image and the real-time measurement image, and δc is the incremental term of the real-time measurement image displacement parameter.

8. The high-precision linear displacement measurement method according to claim 1, characterized in that: The calculation formula for the measurement image position of the target to be measured is as follows: D=ρ z (L-L0)+D0; In the above formula, ρ z is the actual physical position scale coefficient of the zth part of the measurement range, L is the pixel position of the measurement image of the target to be measured, L0 is the starting position of the measurement image of the target to be measured, and D0 is the starting position of the measurement range section of the target to be measured.

9. A high-precision linear displacement measurement method according to claim 1, characterized in that: The coarse positioning of the position of the real-time measurement image on the reference image further includes: An oblique line or a curved line is added to the surface of the information strip, and the rough pixel position of the real-time measurement image is estimated according to the intersection position of the oblique line or the curved line and the center line of the real-time measurement image.

10. A high-precision linear displacement measurement system, characterized in that: include: An acquisition module, used for making information strips on the surface of the target to be measured and acquiring a sequence of images with overlapping areas; A calibration module is used to calibrate the sequence images to obtain a reference image; A calibration module is used to calibrate the reference image to obtain the pixel position of the calibration image and the ratio coefficient of the actual physical position; The coarse positioning module is used to obtain the real-time measurement image during the measurement process, and to coarsely locate the position of the real-time measurement image on the reference image. The pixel offset between the reference image and the real-time measurement image is obtained by combining the pixel position of the calibration image. The fine registration module is used to perform fine registration on the real-time measurement image after rough positioning, and calculate the pixel offset between the reference image and the real-time measurement image to obtain the pixel position of the real-time measurement image; The calculation module is used to obtain the measurement image of the target to be measured, and calculate the measurement image position of the target to be measured based on the pixel position of the real-time measurement image and the actual physical position ratio coefficient.

Images