A high-precision measurement method and system for the rotation angle of an object surface

Abstract

The present invention discloses a method and system for high-precision measurement of rotation angle of an object surface. The method comprises: making information rings on the surface of a target to be measured and numbering them; acquiring a sequence of images and performing polar coordinate conversion on them, and splicing the converted sequence of images to obtain a reference image; calibrating the reference image to obtain the global position of the reference image and the circumference of the target to be measured; acquiring a real-time measurement image and performing rough registration during the measurement process, and calculating the global position of the real-time measurement image in combination with the global position of the reference image; acquiring a final measurement image and performing precise registration, and calculating the angle of the final measurement image in combination with the global position of the real-time measurement image and the circumference of the target to be measured. The system comprises: a numbering module, an acquisition module, a calibration module, a rough registration module, a precise registration module and a calculation module. By using the present invention, high-precision measurement of one-dimensional rotation angle parameters of an object can be achieved.

Description

Technical Field

[0001] The present invention relates to the field of measurement technology, and in particular to a high-precision measurement method and system for the rotation angle of an object surface. Background Art

[0002] The principle of angle measurement is to install the angle measuring device on the object to be measured, and drive the installation axis of the angle sensor to rotate through the rotation of the object to be measured, so as to measure the change of angle; thus, it can provide high-precision displacement or angle data for precision instruments, precision equipment or high-precision weapons and equipment; therefore, it is of great significance to national defense, aerospace, precision manufacturing, automation industry and other industries.

[0003] At present, there are three ways to measure the rotation angle of an object: mechanical, electromagnetic and optical. Mechanical and electromagnetic measurement methods are mostly manual measurements, and the measurement accuracy is affected by the external environment such as temperature and pressure. The photoelectric encoder is an advanced digital angle measurement device that integrates light, machinery and electricity. It converts optical signals into electrical signals through mechanical devices, thereby directly or indirectly measuring various physical quantities such as rotational angular displacement, position and speed. Compared with other sensors for similar purposes, although the photoelectric encoder has the advantages of simple and diverse structural forms, small size, high measurement accuracy and resolution, and strong anti-interference ability, its structure is complex and the manufacturing process requirements are high, so it is expensive and the finished product is easy to damage. Even in actual application, its measurement accuracy will be limited by various conditions and factors, such as code disc scale line error, mechanical structure splicing error, and the influence of signal delay in the use of the encoder. Therefore, it becomes very difficult to further improve the measurement accuracy.

[0004] Imaging encoders are also used. Chinese patent document No. CN201280065242.8 discloses a high-resolution encoder device for measuring the absolute rotation angle of the rotating shaft. Although this method has the characteristics of simple structure and low cost, it still has shortcomings: first, this type of measurement method uses sensors to sense the characteristics of different sections on a circular track, which will lead to reduced working stability due to its high output impedance and poor load capacity. At the same time, the output characteristics are nonlinear. Although it can be improved by a differential structure, it cannot be completely eliminated, which will affect the measurement accuracy. Second, this type of measurement method still uses absolute measurement technology, and its measurement principle is to obtain each position of the encoder by obtaining the open and dark of each engraved line on the disk; such an encoder is determined by the mechanical position of the code disk, which will inevitably lead to its measurement accuracy being affected by installation errors, scale errors, light source errors, etc. Summary of the invention

[0005] In order to solve the above technical problems, an object of the present invention is to provide a method and system for high-precision measurement of the rotation angle of an object surface, which can achieve high-precision measurement of one-dimensional rotation angle parameters of the object.

[0006] The first technical solution adopted by the present invention is: a method for high-precision measurement of the rotation angle of an object surface, comprising the following steps:

[0007] Make information rings on the surface of the target to be measured and number them;

[0008] Obtaining a sequence of images, performing polar coordinate transformation on the sequence of images, and splicing the transformed sequence of images to obtain a reference image;

[0009] Calibrate the reference image to obtain the global position of the reference image and the circumference of the target to be measured;

[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 aligned, and the global position of the real-time measurement image is calculated in combination with the global position of the reference image;

[0011] The final measurement image is obtained and precisely registered with the reference image, and the global position of the final measurement image is calculated in combination with the global position of the real-time measurement image;

[0012] The angle of the final measurement image is obtained by calculating the global position of the final measurement image and the circumference of the target to be measured.

[0013] Furthermore, the information ring includes a region boundary and a coding region, the region boundary is used to locate the coding region, the coding region includes a texture region and a coding mark, and the texture region corresponds to the coding mark one by one.

[0014] Furthermore, the step of making information rings on the surface of the target to be measured and numbering them specifically includes:

[0015] Make an information ring on the surface of the target to be measured;

[0016] Detect all the coded marks in the information ring, and image the marks on the surface of the target to be detected to obtain a mark image;

[0017] Determine the coordinates of the marked image, and intercept a certain range of images with the center of the mark as the reference point as the feature image corresponding to the mark;

[0018] The marked feature images are numbered using the coding marks of the coding areas.

[0019] Furthermore, the step of acquiring the sequence images, performing polar coordinate conversion on them, and splicing the converted sequence images to obtain the reference image specifically includes:

[0020] Acquire a sequence of images and perform polar coordinate transformation on a certain point in the sequence of images to obtain a polar coordinate transformation relationship;

[0021] Calculate the polar coordinate transformation origin of the sequence image and perform origin correction to obtain the origin coordinates;

[0022] Perform polar coordinate transformation on the sequence images according to the transformation relationship between the origin coordinates and the polar coordinates;

[0023] The sequence images after polar coordinate conversion are stitched together to obtain the reference image.

[0024] Furthermore, the step of calibrating the reference image to obtain the global position of the reference image and the circumference of the target to be measured specifically includes:

[0025] Selecting a reference image with overlapping regions;

[0026] Select multiple sub-regions from the overlapping region, and estimate the displacement of each sub-region relative to the reference image to obtain the global position of the reference image;

[0027] The initial reference image of the target to be measured and the first reference image with the same angle as the initial reference image are obtained, and the circumference of the target to be measured is calculated in combination with the global position of the reference image.

[0028] Furthermore, the step of acquiring a real-time measurement image during the measurement process, roughly registering the position of the real-time measurement image on the reference image, and calculating the global position of the real-time measurement image in combination with the global position of the reference image specifically includes:

[0029] Acquire real-time measurement images during the measurement process;

[0030] Locate the coding area by measuring the area boundaries in the image in real time;

[0031] Obtaining the angular position of the real-time measurement image according to the coding mark of the coding area;

[0032] The global position of the real-time measurement image is calculated by combining the angular position of the real-time measurement image and the global position of the reference image.

[0033] Furthermore, the coarse registration of the position of the real-time measurement image on the reference image further includes:

[0034] An oblique line or a curved line is added to the surface of the information ring, and the global position of the real-time measurement image on the reference 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.

[0035] Furthermore, the step of obtaining the final measurement image and accurately registering it with the reference image, and calculating the global position of the final measurement image in combination with the global position of the real-time measurement image specifically includes:

[0036] Obtain the final measurement image;

[0037] Constructing the geometric relationship between the final measured image and the reference image based on the image transformation model;

[0038] Calculating the displacement parameters of the final measurement image and the reference image based on the geometric relationship between the final measurement image and the reference image based on the image matching algorithm;

[0039] The global position of the real-time measurement image is combined with the displacement parameter calculation of the final measurement image and the reference image to obtain the global position of the final measurement image.

[0040] Furthermore, it also includes performing polar coordinate conversion according to a lookup table, and the formula of the lookup table is as follows:

[0041]

[0042] In the above formula, (u,v) is the new labeled image point, (x,y) is the original labeled image point, s is the decimal of x, t is the decimal of y, P m (s) is the interpolation function of s, m is the interpolation node of s, m = -1, 0, 1, 2, 3, P n (t) is the t interpolation function, n is the interpolation node of t, n = -1, 0, 1, 2, 3.

[0043] The second technical solution adopted by the present invention is: a high-precision measurement system for the rotation angle of an object surface, comprising:

[0044] A numbering module is used to make information rings on the surface of the target to be measured and number them;

[0045] An acquisition module is used to acquire a sequence of images, perform polar coordinate conversion on them, and splice the converted sequence of images to obtain a reference image;

[0046] A calibration module is used to calibrate the reference image to obtain the global position of the reference image and the circumference of the target to be measured;

[0047] A coarse registration module is used to obtain a real-time measurement image during the measurement process, and to perform coarse registration on the position of the real-time measurement image on the reference image, and to calculate the global position of the real-time measurement image in combination with the global position of the reference image;

[0048] A precise registration module is used to obtain the final measurement image and perform precise registration with the reference image, and calculate the global position of the final measurement image in combination with the global position of the real-time measurement image;

[0049] The calculation module is used to calculate the angle of the final measurement image according to the global position of the final measurement image and the circumference of the target to be measured.

[0050] The method and system of the present invention have the following beneficial effects: firstly, the present invention produces an information ring on the surface of the target to be measured. Since the information ring identification can be random speckle information and bar code mark, number, two-dimensional code mark, etc., the method is more flexible and has wider applicability; secondly, in order to improve the efficiency of converting the image from the rectangular coordinate system to the polar coordinate system, a lookup table (LUT) is established to realize the rapid polar coordinate transformation of the image, thereby effectively saving the conversion time; then, the coding mark is quickly located through the region boundary, and the circular ring is quickly and roughly located by using the coding mark recognition; finally, accurate positioning is realized by matching the final measured image with the reference image, and the corresponding angle is obtained by the circular ring position, thereby realizing the high-precision measurement of the one-dimensional rotation angle parameter of the object. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of the steps of a method for high-precision measurement of rotation angle of an object surface of the present invention;

[0052] Figure 2 It is a structural block diagram of a high-precision measurement system for the rotation angle of an object surface of the present invention;

[0053] Figure 3 is a schematic diagram of an information ring in a specific embodiment of the present invention;

[0054] Figure 4 is a schematic diagram of polar coordinate conversion of a point in a marked image according to a specific embodiment of the present invention;

[0055] Figure 5 is a schematic diagram of polar coordinate conversion of a marked image according to a specific embodiment of the present invention;

[0056] Figure 6 is a schematic diagram of image parameters to be transformed according to a specific embodiment of the present invention;

[0057] Figure 7 is a schematic diagram of reference image calibration according to a specific embodiment of the present invention;

[0058] Figure 8 This is a schematic diagram of the first real-time measurement image coarse positioning according to a specific embodiment of the present invention;

[0059] Fig. 9 It is a schematic diagram of the second real-time measurement image coarse positioning according to a specific embodiment of the present invention. DETAILED DESCRIPTION

[0060] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0061] Reference Figure 1 The present invention provides a method for high-precision measurement of the rotation angle of an object surface, the method comprising the following steps:

[0062] S1. Make information rings on the surface of the target to be measured and number them;

[0063] S1.1. Make an information ring on the surface of the target to be measured;

[0064] Specifically, Figure 3 As shown, the information ring is a ring mark with an outer diameter of 30mm and an inner diameter of 24mm. The information ring includes a region boundary and a coding region. The region boundary is used to quickly locate the coding region. A coding region includes a coding mark and a texture region. The information ring is divided into 60 coding regions. Each coding region corresponds to an angle of 6°. The coding region is used as the scale line of the information ring on the top surface of the truncated cone. The coding mark adopts a six-bit binary code. Each coding region has six blocks. Each block can be white or black. Black represents 0 and white represents 1, so it can represent 60 scales. And ensure that there is at least one complete mark in the image when imaging the surface of the object to be measured.

[0065] Among them, the information ring logo can be random speckle information and bar code marks, numbers, QR code marks, etc. (spraying, pasting), which is more flexible and has wider applicability.

[0066] S1.2, detecting all the coding marks in the information ring, and imaging the marks on the surface of the target to be detected to obtain a mark image;

[0067] S1.3. Determine the coordinates of the marked image, and intercept a certain range of images with the center of the mark as a reference point as the feature image corresponding to the mark.

[0068] Specifically, firstly, all the markers on the surface of the target to be measured are imaged, and the coordinates of the marker image are determined using the centroid method. The centroid is the coordinate center of each point with a pixel value of 255. The calculation formula is as follows:

[0069]

[0070] In the above formula, n is the number of points with a pixel value of 255 in the current target, (x i ,y i ) are the image coordinates of each point.

[0071] Secondly, an image of a certain range (such as 200×200 pixels) is intercepted with the center of the mark as the reference point as the feature image corresponding to the mark (due to the randomness of texture information, the feature image of each mark is different).

[0072] S1.4. Use the coding mark of the coding area to number the marked feature image.

[0073] Specifically, since the coding mark adopts six-bit binary code, when the number is 1, the corresponding binary code is 000001, when the number is 2, the corresponding binary code is 000010, and so on; therefore, the numbers from 0 to 60 can be numbered according to the corresponding binary codes and distinguish the characteristic images of the marks.

[0074] S2, acquiring a sequence of images, performing polar coordinate transformation on the sequence of images, and splicing the transformed sequence of images to obtain a reference image;

[0075] Specifically, since image pixels are arranged along the x or y axis rather than at an angle, in order to facilitate image registration, the image needs to be converted from the rectangular coordinate system to the polar coordinate system, that is, the rotational motion is converted into a linear displacement motion, and the angular difference between different images is converted into pixel displacement on the axis.

[0076] S2.1, acquiring a sequence of images and performing polar coordinate transformation on a certain point in the sequence of images to obtain a polar coordinate transformation equation;

[0077] Specifically, Figure 4 As shown, a point P(x, y) in the original marked image can be converted to a point P(θ, ρ) in the polar coordinate system. According to the rectangular coordinate system, θ is the angle of P(x, y), corresponding to the rotation angle of the rotating body, and ρ is the distance from point P (x, y) to the origin of the coordinate system. The conversion is expressed as:

[0078]

[0079] In the above formula, (x0, y0) is the origin of the rectangular coordinate system, so Figure 5 As shown, a part of the annular ABCD can be converted into a rectangular ABCD. The circular pattern can be converted into a rectangular pattern through polar coordinate transformation, and the circular motion can be converted into translational motion, which is convenient for image matching and angle measurement. It has the advantages of high execution efficiency, good real-time performance, simplicity and reliability.

[0080] It should be noted that, since each sequence image has a complete coding mark, it can also be called a marked image.

[0081] S2.2, calculating the polar coordinate transformation origin of the sequence image and performing origin correction to obtain the origin coordinates;

[0082] Specifically, from the polar coordinate transformation in step S2.1, if we want to transform the labeled image, we must know the origin (x0, y0), however, the origin cannot be found directly in the labeled image because it is not in the labeled image, but it can be estimated based on the point relationship between adjacent labeled images.

[0083] Assume that a point in the i-th labeled image is P i (x i ,y i ), the same point in the jth labeled image is P j (x j ,y j ). At this time, P j The relationship is:

[0084] (x i -x0) 2 +(y i -y0) 2 =(x j -x0) 2 +(y j -y0) 2 ;

[0085] P i (x i ,y i ) and P j (x j ,y j ) The two positions of the same point in different labeled images are named point pairs. The origin (x0, y0) can be solved by two point pairs. In order to obtain high-precision results, multiple point pairs can be used.

[0086] In order to obtain dot pairs, multiple points can be selected in a marked image, and the corresponding points can be found by combining the normalized cross correlation (NCC) image matching algorithm with the quadratic surface fitting method, so that the angular difference of the matching image is slightly less than 10 degrees. The pixel accuracy of this algorithm can reach single digits. The NCC algorithm is expressed as:

[0087]

[0088] Where R and I are the two labeled images to be matched, and the matching point (x, y) is the point with the highest correlation coefficient C.

[0089] S2.3. Perform polar coordinate transformation on the sequence images according to the conversion relationship between the origin coordinates and the polar coordinates to obtain the sequence images and their polar coordinates after polar coordinate transformation.

[0090] Specifically, Figure 6As shown, after calculating the origin (x0, y0), it is necessary to transform all the labeled images obtained by step S2.1. First, calculate the range of θ and ρ of the labeled image to be transformed, name the range of θ as (θ1, θ2), and the range of ρ as (ρ1, ρ2), and get the maximum and minimum values ​​of θ and ρ to calculate the boundary. The expression is as follows:

[0091]

[0092] Furthermore, assuming that the pixel width and height of the new labeled image are w and h, and a point (u, v) of the new labeled image is represented in pixel form as:

[0093]

[0094] In the above formula, θ is the angle of change of point P (x, y), θ1 is the minimum value of θ, θ2 is the maximum value of θ, ρ is the distance from point P (x, y) to the origin of the coordinate system, ρ1 is the minimum value of ρ, ρ2 is the maximum value of ρ, is equivalent to .

[0095] In order to simplify the formula, α is used instead of (θ2-θ1) / w, and β is used instead of (ρ2-ρ1) / h.

[0096] The polar coordinate form is:

[0097]

[0098] The corresponding relationship between the points (x, y) obtained from the newly marked image points (u, v) is as follows:

[0099]

[0100] In the above formula, θ0 is the central angle between θ1 and θ2, and x0 is the half-width of the original labeled image pixel.

[0101] According to the point (u, v) in the new labeled image, the pixel position (x, y) in the original labeled image is obtained. The corresponding relationship is as follows:

[0102]

[0103] According to the above formula, the position of (x, y) is calculated, and the grayscale value is obtained by interpolating each point (u, v) in the new marked image. In this way, a new marked image is obtained, and the polar coordinates of the marked image are obtained after transformation.

[0104] As a further preferred embodiment of the method, a lookup table is further included to improve the efficiency of converting the image from the rectangular coordinate system to the polar coordinate system. By establishing a lookup table (LUT), the image is quickly transformed into polar coordinates, which effectively saves the conversion time. The specific steps of constructing the lookup table are as follows:

[0105] For each point (u,v) in the new labeled image, we need to get the grayscale value at point (x,y) from the original labeled image. This point is a sub-pixel position, and we get its value from its 16 adjacent integer points through bicubic interpolation:

[0106]

[0107] In the above formula, s and t are decimals of x and y, m is the interpolation node of x, m = 0, 1, 2, 3, n is the interpolation node of y, n = 0, 1, 2, 3, and P m (u) is a weight function and also an interpolation function expression, expressed as:

[0108]

[0109] Depend on It can be seen that when switching to the marker image position, all marker images remain unchanged. Therefore, a lookup table (LUT) can be used to achieve fast conversion.

[0110] For each point (u, v) in the new labeled image, an 18-item LUT is constructed for it, represented as follows:

[0111]

[0112] In the above formula, (u,v) is the new labeled image point, (x,y) is the original labeled image point, s is the decimal of x, t is the decimal of y, P m (s) is the interpolation function of s, m is the interpolation node of s, m = -1, 0, 1, 2, 3, P n (t) is the t interpolation function, n is the interpolation node of t, n = -1, 0, 1, 2, 3.

[0113] Take advantage of LUTs before the conversion process and apply them during the conversion to save time.

[0114] S2.4. Splice the sequence images after polar coordinate conversion to obtain a reference image to calibrate the coding mark.

[0115] S3, calibrating the reference image to obtain the global position of the reference image and the circumference of the target to be measured;

[0116] Specifically, Figure 7As shown, first assume that the reference image moves in the x direction as it rotates, and the x coordinate of the pixel center is used to represent the reference image as T i (0≤i≤N), in the overlapping area of ​​the i-th reference image and the j-th reference image, select several sub-areas, and use the ICGN algorithm to estimate the image shift of the two reference images for each sub-area The details are as follows:

[0117]

[0118] In the above formula, k is the sub-region index, R is the sub-region number, is the x-coordinate of the k-th sub-region in the i-th reference image, is the x-coordinate of the k-th subregion in the j-th reference image, and the displacement E between the i-th reference image and the j-th reference image i,j Estimated to be:

[0119]

[0120] Set T0 = 0, and according to the reference image displacement E i,j , the global position of all benchmark images is obtained as:

[0121] T i =T i-1 +E i-1,i (1≤i≤N);

[0122] Secondly, find the rth reference image T0 at the same angle as the initial mark image T0 of the target to be measured r , estimate the circumference P of the circle, and the calculation formula is as follows:

[0123] P=T r -T0;

[0124] It should be noted that T r With T i The calculation formula is the same.

[0125] Further, these results are used as T i and the initial value of P, and calculate T by minimizing the energy equation i The exact value of and P is expressed as:

[0126]

[0127] In the above formula, s is the index of the image that has an overlapping area with the i-th reference image, and S is the number of images that have an overlapping area with the i-th reference image.

[0128] S4, acquiring a real-time measurement image during the measurement process, and coarsely registering the position of the real-time measurement image on the reference image, and calculating the global position of the real-time measurement image in combination with the global position of the reference image;

[0129] Specifically, each real-time measurement image contains a coding mark. For each coding mark, for example, the nth coding mark, an image containing the coding mark is selected and named as the i-th real-time measurement image, such as Figure 8 As shown, during the measurement process, a real-time measurement image is acquired, and the center position of the coding mark in the x direction is estimated through the cross-sectional boundary. When the center C of the nth coding mark is determined n When , a marked image area is selected as the marked area corresponding to the nth coded mark, and the center M of the marked area is n Angular position as a coding mark.

[0130] Set the distance from the center of the marking area to the center of the encoding mark to a fixed value of D mc , and press M n The location of the selected marked area is:

[0131] M n =C n +D mc ;

[0132] The x position of the center of the nth marked area in the i-th real-time measurement image is defined as m n,i , mark the center of the area M n The precise global x position of the real-time measurement image is expressed as:

[0133] M n =T i -w / 2+m n,i ;

[0134] In the above formula, w is the image width, T i is the global position of the center of the reference image corresponding to the i-th real-time measurement image.

[0135] In addition, the coarse registration of the position of the real-time measurement image on the reference image is not unique, e.g. Fig. 9 As shown, it is also possible to obtain the cross-sectional area based on the above method according to its global position C n , with C n 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 precise position m of the coding mark of the i-th real-time measurement image. n,i .

[0136] 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, so it can be used to roughly locate the position of the real-time measurement image on the reference image.

[0137] Assume that the intersection point P of the oblique line and the central axis is (x, y), the width of the measurement range is W, the height is H, the width of the real-time measurement map 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:

[0138]

[0139] 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, H0 is the distance between the intersection point P and the upper boundary of the measurement range, is equivalent to .

[0140] S5, obtaining the final measurement image and accurately registering it with the reference image, and calculating the global position of the final measurement image in combination with the global position of the real-time measurement image;

[0141] S5.1, obtaining the final measurement image;

[0142] Specifically, the reference image marked area is first recorded into the final measured image of the area near the rough position of the marked area, and then the image registration algorithm used is the inverse combination matching algorithm combined with the Gauss-Newton (ICGN) algorithm proposed by Baker and Matthews, which has very good accuracy, up to 0.001 pixel.

[0143] S5.2. Constructing the geometric relationship between the final measurement image and the reference image based on the image transformation model;

[0144] The ICGN algorithm generally uses an affine transformation model to represent the geometric changes between two images, but for angle measurement, the geometric relationship between the marked area of ​​the reference image and the final measured image is a simple one-dimensional displacement that can be described by no more than two parameters:

[0145]

[0146] The reference image marked area and the final measured image are named R(x,y) and I(x,y) respectively. c and f are two parameters of image displacement, which are expressed as s(c,f). Then the marked area is expressed as:

[0147]

[0148] S5.3, calculating the displacement parameters between the final measurement image and the reference image based on the geometric relationship between the final measurement image and the reference image based on the image matching algorithm;

[0149] In order to obtain the correct displacement parameters, the cost function must satisfy its minimum value, as shown below:

[0150]

[0151] Using the least squares method:

[0152]

[0153] In the above formula, s i is the displacement parameter of the i-th image, Q is the energy, is the partial derivative of energy.

[0154] Furthermore, the result of R(x,y,s)-I(x,y) is:

[0155]

[0156] In the above formula, is the displacement parameter change, is the partial derivative of the displacement parameter.

[0157] Furthermore, we can get:

[0158]

[0159] and The transformation parameters can be calculated by the least squares method, and the expression is:

[0160]

[0161] In the above formula, H is the Hessian matrix, which is expressed as:

[0162]

[0163] From the above formula, we can get And apply it to the real-time measurement image I(x,y), and obtain a new real image through parameter transformation Then repeat the above process until the error is lower than the expected error value.

[0164]

[0165] 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.

[0166] Since the value obtained by the fine image displacement is an incremental term based on the coarse position, the final position of the marked area in the measurement is:

[0167]

[0168] S5.4. The global position of the real-time measurement image is combined with the displacement parameter calculation of the final measurement image and the reference image to obtain the global position of the final measurement image.

[0169] Specifically, the specific formula for measuring the global position of the final image is as follows:

[0170] L=M n +w / 2-m n .

[0171] S6. Calculate the angle of the final measurement image according to the global position of the final measurement image and the circumference of the object to be measured.

[0172] Specifically, based on the global position L of the final measurement image and the circumference P of the target to be measured, the calculation formula for the rotation angle of the target to be measured is finally obtained as follows:

[0173] θ=L*360 / P.

[0174] This method quickly locates the coded mark through the area boundary, uses the coded mark recognition to quickly and roughly locate the ring, achieves precise positioning by matching the real-time measurement image with the mark image, and obtains the corresponding angle through the position of the ring, making the measurement faster and more convenient, and realizing high-precision measurement of the one-dimensional rotation angle parameters of the object.

[0175] like Figure 2 As shown, a high-precision measurement system for the rotation angle of an object surface comprises:

[0176] A numbering module is used to make information rings on the surface of the target to be measured and number them;

[0177] An acquisition module is used to acquire a sequence of images, perform polar coordinate conversion on them, and splice the converted sequence of images to obtain a reference image;

[0178] A calibration module is used to calibrate the reference image to obtain the global position of the reference image and the circumference of the target to be measured;

[0179] A coarse registration module is used to obtain a real-time measurement image during the measurement process, and to perform coarse registration on the position of the real-time measurement image on the reference image, and to calculate the global position of the real-time measurement image in combination with the global position of the reference image;

[0180] A precise registration module is used to obtain the final measurement image and perform precise registration with the reference image, and calculate the global position of the final measurement image in combination with the global position of the real-time measurement image;

[0181] The calculation module is used to calculate the angle of the final measurement image according to the global position of the final measurement image and the circumference of the target to be measured.

[0182] 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.

[0183] 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 may 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 method for high-precision measurement of rotation angle of an object surface, characterized in that: The following steps are involved: Make information rings on the surface of the target to be measured and number them; Obtaining a sequence of images, performing polar coordinate transformation on the sequence of images, and splicing the transformed sequence of images to obtain a reference image; Calibrate the reference image to obtain the global position of the reference image and the circumference of the target to be measured; 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 aligned, and the global position of the real-time measurement image is calculated in combination with the global position of the reference image; The final measurement image is obtained and precisely registered with the reference image, and the global position of the final measurement image is calculated in combination with the global position of the real-time measurement image; The angle of the final measurement image is obtained by calculating the global position of the final measurement image and the circumference of the target to be measured.

2. The method for high-precision measurement of rotation angle of an object surface according to claim 1, characterized in that: The information ring includes a region boundary and a coding region, the region boundary is used to locate the coding region, the coding region includes a texture region and a coding mark, and the texture region corresponds to the coding mark one by one.

3. The method for high-precision measurement of rotation angle of an object surface according to claim 2, characterized in that: The step of making information rings on the surface of the target to be measured and numbering them specifically includes: Make an information ring on the surface of the target to be measured; Detect all the coded marks in the information ring, and image the marks on the surface of the target to be detected to obtain a mark image; Determine the coordinates of the marked image, and intercept a certain range of images with the center of the mark as the reference point as the feature image corresponding to the mark; The marked feature images are numbered using the coding marks of the coding areas.

4. The method for high-precision measurement of rotation angle of an object surface according to claim 1, characterized in that: The step of acquiring a sequence of images, performing polar coordinate transformation on the sequence of images, and splicing the transformed sequence of images to obtain a reference image specifically includes: Acquire a sequence of images and perform polar coordinate transformation on a certain point in the sequence of images to obtain a polar coordinate transformation relationship; Calculate the polar coordinate transformation origin of the sequence image and perform origin correction to obtain the origin coordinates; Perform polar coordinate transformation on the sequence images according to the transformation relationship between the origin coordinates and the polar coordinates; The sequence images after polar coordinate conversion are stitched together to obtain the reference image.

5. The method for high-precision measurement of rotation angle of an object surface according to claim 1, characterized in that: The step of calibrating the reference image to obtain the global position of the reference image and the circumference of the target to be measured specifically includes: Selecting a reference image with overlapping regions; Select multiple sub-regions from the overlapping region, and estimate the displacement of each sub-region relative to the reference image to obtain the global position of the reference image; The initial reference image of the target to be measured and the first reference image with the same angle as the initial reference image are obtained, and the circumference of the target to be measured is calculated in combination with the global position of the reference image.

6. The method for high-precision measurement of the rotation angle of an object surface according to claim 2, characterized in that: The step of acquiring a real-time measurement image during the measurement process, roughly registering the position of the real-time measurement image on the reference image, and calculating the global position of the real-time measurement image in combination with the global position of the reference image specifically includes: Acquire real-time measurement images during the measurement process; Locate the coding area by measuring the area boundaries in the image in real time; Obtaining the angular position of the real-time measurement image according to the coding mark of the coding area; The global position of the real-time measurement image is calculated by combining the angular position of the real-time measurement image and the global position of the reference image.

7. The method for high-precision measurement of rotation angle of an object surface according to claim 6, characterized in that: The coarse registration of the position of the real-time measurement image on the reference image also includes: An oblique line or a curved line is added to the surface of the information ring, and the global position of the real-time measurement image on the reference 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.

8. The method for high-precision measurement of rotation angle of an object surface according to claim 1, characterized in that: The step of obtaining the final measurement image and accurately registering it with the reference image, and calculating the global position of the final measurement image in combination with the global position of the real-time measurement image specifically includes: Obtain the final measurement image; Constructing the geometric relationship between the final measured image and the reference image based on the image transformation model; Calculating the displacement parameters of the final measurement image and the reference image based on the geometric relationship between the final measurement image and the reference image based on the image matching algorithm; The global position of the real-time measurement image is combined with the displacement parameter calculation of the final measurement image and the reference image to obtain the global position of the final measurement image.

9. The method for high-precision measurement of rotation angle of an object surface according to claim 1, characterized in that: It also includes performing polar coordinate conversion according to a lookup table, wherein the formula of the lookup table is as follows: In the above formula, (u,v) is the new labeled image point, (x,y) is the original labeled image point, s is the decimal of x, t is the decimal of y, P m (s) is the interpolation function of s, m is the interpolation node of s, m = -1, 0, 1, 2, 3, P n (t) is the t interpolation function, n is the interpolation node of t, n = -1, 0, 1, 2, 3.

10. A high-precision measurement system for the rotation angle of an object surface, characterized in that: include: A numbering module is used to make information rings on the surface of the target to be measured and number them; An acquisition module is used to acquire a sequence of images, perform polar coordinate conversion on them, and splice the converted sequence of images to obtain a reference image; A calibration module is used to calibrate the reference image to obtain the global position of the reference image and the circumference of the target to be measured; A coarse registration module is used to obtain a real-time measurement image during the measurement process, and to perform coarse registration on the position of the real-time measurement image on the reference image, and to calculate the global position of the real-time measurement image in combination with the global position of the reference image; A precise registration module is used to obtain the final measurement image and perform precise registration with the reference image, and calculate the global position of the final measurement image in combination with the global position of the real-time measurement image; The calculation module is used to calculate the angle of the final measurement image according to the global position of the final measurement image and the circumference of the target to be measured.

Images