A machine tool track microscopic vision measurement method based on small circle coding within a large circle

Abstract

This invention relates to a microscopic visual measurement method for machine tool trajectories, and more specifically, a microscopic visual measurement method for machine tool trajectories based on small circle codes scattered within a large circle. The method includes the following steps: S1, a measuring head is fixedly attached to the first moving end of the machine tool, and a measuring target is fixedly attached to the second moving end of the machine tool; the measuring head is connected to a host computer; S2, the field of view of the measuring head is adjusted to the center of the scale range of the measuring target; S3, the image from the host computer is observed to complete the focusing operation, and the coded pattern within the field of view of the measuring head on the measuring target is acquired to determine whether the optical axis of the measuring head is orthogonal to the target plane; S4, the second moving end is moved, with the movement range not exceeding the effective measurement range of the target, to measure the actual movement trajectory of the spindle relative to the worktable. This non-contact microscopic visual measurement method improves the accuracy, efficiency, and safety of positioning.

Description

Technical Field

[0001] This invention relates to a microscopic visual measurement method for machine tool trajectories, and more specifically, to a microscopic visual measurement method for machine tool trajectories based on the coding of small circles scattered within a large circle. Background Technology

[0002] Machine tool error parameters are a prerequisite for machine tool error compensation and fault diagnosis. Accurately measuring the machine tool's motion trajectory using instruments is the primary method for quickly obtaining these parameters. Traditional machine tool trajectory measuring instruments generally suffer from drawbacks such as high cost, low degree of freedom, contact operation, and inconvenient installation and debugging. To address these issues, a machine tool trajectory microscopic vision measurement method can be employed. This method utilizes microscopic vision technology to measure the machine tool trajectory in real time. By installing a microscope and camera on the machine tool, the trajectory is recorded. Then, image processing and error analysis methods are used to measure and analyze the trajectory to obtain the machine tool's error parameters. Summary of the Invention

[0003] This invention provides a microscopic visual measurement method for machine tool trajectories based on small circle codes distributed within a large circle. The aim is to improve the accuracy, efficiency, and safety of positioning by using a non-contact microscopic visual measurement method.

[0004] The above objectives are achieved through the following technical solutions:

[0005] A microscopic visual measurement method for machine tool trajectories based on the coding of small circles scattered within a large circle includes the following steps:

[0006] S1. A measuring head is fixedly connected to the first moving end of the machine tool, and a measuring target is fixedly connected to the second moving end of the machine tool. The measuring head is connected to the host computer.

[0007] S2. Adjust the field of view of the measuring head to the center of the scale range of the measuring target;

[0008] S3. Observe the image of the host computer to complete the focusing operation, collect the coded pattern in the target and within the field of view of the measuring head, and determine whether the optical axis of the measuring head is orthogonal to the target plane by detecting the coded pattern.

[0009] S4. Move the second moving end, ensuring the range of motion does not exceed the effective measurement range of the target, in order to measure the true motion trajectory of the spindle relative to the worktable.

[0010] The first mobile terminal has at least a longitudinal direction as its degree of freedom, and the second mobile terminal has at least a horizontal direction as its degree of freedom.

[0011] In step three, the measuring head is roughly adjusted to the center of the target by feeding the machine tool in the X and Y directions;

[0012] Observe the image captured by the camera in the host computer, adjust the Z-axis of the machine tool, set the Z-axis to the minimum feed, move the Z-axis of the machine tool, and when the image clarity of the camera remains unchanged, the focusing operation is completed. After that, the Z-axis position of the machine tool will no longer change.

[0013] The coded pattern within the field of view includes at least one complete sub-coded pattern;

[0014] The complete sub-encoding pattern includes: circle I, within which two circles II and III of different diameters are displayed. Circle II and circle III are located on the horizontal and vertical lines of symmetry of circle I, respectively. The distance between the center of circle II and the center of circle I is spacing I, and the distance between the center of circle III and the center of circle I is spacing II. Spacing I and spacing II are equal.

[0015] There is a small circular encoding area between circle I, circle II and circle III for storing binary bit information.

[0016] The array in the small circular encoding area has a circle IV, which represents two numbers in binary when it is displayed and when it is not displayed.

[0017] The lowest bit of the small circle code is located at the bottom right corner of the small circle code area. The number of bits in the small circle code positions gradually increases in the upward winding direction until the top left corner becomes the highest bit of the small circle code position.

[0018] The effective measurement range of the measurement target includes a coded array pattern;

[0019] The array coding pattern includes generating 8720 sub-coding patterns within a 150mm×150mm area. The intervals between each sub-coding pattern are equal, with 80 codes per row and 109 codes per column.

[0020] The design size of the four sub-code patterns is no larger than the camera's field of view.

[0021] The measuring head and the measuring target move relative to each other. The measuring head acquires each frame of image during the movement, and each frame of image is processed as follows:

[0022] Step 1: Basic Image Processing: Obtain clear boundaries of the array-coded pattern, suppress noise interference with image quality, and improve the algorithm's robustness to lighting conditions.

[0023] Step 2: Solve for the scaling factor, the rotation angle of the encoded pattern, and the pixel size. This step is only performed in the first frame.

[0024] Step 3: Solve for the position of pixel I in circle 1;

[0025] Step 4: Set the selected circle I rectangle boundary as the ROI region to improve decoding processing speed;

[0026] Step 5: Decode the circle to determine that circle I is located in the nth row and mth column of the initial encoded circle;

[0027] Step 6: Solve the absolute position of the image center frame by frame to obtain the machine tool planar motion measurement trajectory.

[0028] S5. Compare the measured planar motion trajectory with the theoretical planar motion trajectory.

[0029] The beneficial effects of the microscopic visual measurement method for machine tool trajectory based on the coding of small circles scattered within a large circle, as described in this invention, are as follows:

[0030] Using computer vision methods, motion trajectory measurement of CNC milling machines / machining centers can be achieved. Compared with traditional measurement methods, it has the advantages of low cost, non-contact, simple installation and debugging, absolute measurement, arbitrary trajectory measurement, large range and low measurement resolution, and real-time measurement.

[0031] It uses only classic visual algorithms such as edge detection, least squares fitting, and grayscale determination, which improves the efficiency of decoding computation.

[0032] The accuracy of positioning depends only on the outline of the large circle's edge and its relative position, thus improving positioning precision.

[0033] By introducing scaling and rotation factors, successful decoding is possible even under scaling and rotation conditions, thus improving the robustness of decoding.

[0034] This encoding method has a large number of bits and can achieve 16-bit binary encoding, which means it can generate a maximum of 65,536 different codes. Attached Figure Description

[0035] Figure 1 A flowchart of a machine tool trajectory microscopic visual measurement method based on small circle coding distributed within a large circle;

[0036] Figure 2 A diagram illustrating a sub-encoding;

[0037] Figure 3 This is a schematic diagram showing the positions of each circle in the sub-encoding;

[0038] Figure 4 A schematic diagram of the sub-encoding pattern corresponding to 1111011110111101;

[0039] Figure 5 This is a schematic diagram of the coded array pattern;

[0040] Figure 6 A schematic diagram showing four sub-coded patterns within the camera's field of view;

[0041] Figure 7 This is a flowchart for solving machine tool trajectories based on an array of measurement-coded patterns. Detailed Implementation

[0042] A microscopic visual measurement method for machine tool trajectories based on the coding of small circles scattered within a large circle includes the following steps:

[0043] S1. The first moving end of the machine tool, such as the spindle, is fixedly connected to the measuring head; the second moving end of the machine tool, such as the worktable, is fixedly connected to the measuring target; and the measuring head is connected to the host computer.

[0044] The degrees of freedom of the first moving end include at least the longitudinal direction, i.e., the Z-axis direction, and the degrees of freedom of the second moving end include at least the horizontal direction, such as the X and Y axes.

[0045] The industrial camera in the measuring head uses the Medvision SUM133GC board, with a resolution of 1280×1024, a frame rate of 220FPS, and a pixel size of 4.0μm; the industrial lens uses the Moritex MML2-ST65S telecentric lens with a magnification of 2x.

[0046] The measurement target is formed by integrating at least a highly uniform planar light source and a photomask, and the effective measurement range of the measurement target is approximately 150mm × 150mm.

[0047] The host computer is a computer, and the industrial camera's USB data cable is connected to the computer.

[0048] S2. Turn on the high uniformity planar light source of the measuring target and adjust the field of view of the measuring head to the center of the scale range of the measuring target.

[0049] S3. Observe the image from the host computer to complete the focusing operation. By detecting the measurement code pattern in the target, determine whether the optical axis of the measuring head is orthogonal to the target plane. Specifically, by feeding the machine tool in the X and Y directions, roughly adjust the measuring head to the center of the target.

[0050] Observe the camera's image on the computer, adjust the machine tool's Z-axis, set the Z-axis to minimum feed, move the machine tool's Z-axis, and when the camera's image clarity remains unchanged, the focusing operation is complete. After that, the position of the machine tool's Z-axis will no longer change.

[0051] The distance between the lens tip and the target is approximately 65mm.

[0052] The measurement coding pattern includes at least one complete sub-coding pattern, and the internal composition of a complete sub-coding pattern is as follows:

[0053] refer to Figure 2 and 3The system includes the largest circle, denoted as Circle I, used for positioning. Circle I has a diameter of φ0.6mm and a black outline, with a white background area inside. It also includes a large black reference circle, denoted as Circle II, located to the left of the white area inside Circle I. Circle II has a diameter of φ0.08mm, and its center is 0.24mm from the center of Circle I. The center of Circle II lies on the horizontal line of symmetry of Circle I. Inside Circle I, on the upper side, is a smaller black reference circle, denoted as Circle III, with a diameter of φ0.06mm. The center of Circle III is 0.24mm from the center of Circle I, and its center lies on the vertical line of symmetry of Circle I. The relative center positions and dimensions of these two reference circles are consistent within each Circle I code; that is, in any sub-code pattern, the dimensions of Circle II and Circle III, as well as their positions within Circle I, are the same.

[0054] The line connecting the center of circle I and the center of circle II is denoted as line I, and the line connecting the center of circle I and the center of circle III is denoted as line II. Line I and line II are perpendicular to each other. The diameters of circles II and III are different and are used to identify the rotation direction of the pattern code.

[0055] The small circular coding area inside circle I is used to store binary bit information, and the relative position of this coding area in each sub-coding pattern remains consistent.

[0056] Within the small circle encoding area: the small circle is denoted as circle IV. If the color of circle IV is white, it represents binary "0"; if the color is black, it represents binary "1". This small circle encoding area can represent a maximum of 2. 16 The bit encoding allows for the insertion of more reserved positions for small circle codes to further increase the maximum number of bits; the bottom right corner is the position of the lowest bit small circle code, and the top left corner is the position of the highest bit small circle code. The number of bits increases sequentially along the arrow direction; the position of circle IV is determined based on the vertical distance to the virtual positioning X-axis generated by circle II and the virtual positioning Y-axis generated by circle III. The vertical distances of the nearest small circle code center from the positioning X-axis and positioning Y-axis are 0.16mm and 0.11mm, respectively.

[0057] Encoding dimensions for microscopic visual positioning, such as Figure 3 As shown, the unit is mm; adjacent circles IV are evenly spaced, with the horizontal arrangement direction parallel to the horizontal diameter and the vertical arrangement direction parallel to the vertical diameter. The horizontal spacing of each circle IV is 0.08 mm and the vertical spacing is 0.09 mm. Multiple sub-code patterns are generated according to this size and the rule of carrying over 1 bit in sequence.

[0058] refer to Figure 4 , is the sub-encoding pattern corresponding to 11110111101111101;

[0059] Using computer programming, sub-coded pattern arrays are generated to form a coded array pattern, refer to... Figure 5This involves generating 8720 sub-coded patterns within an area of ​​approximately 150mm × 150mm. Each sub-coded pattern is evenly spaced, with 80 codes per row and 109 codes per column, forming a coded array pattern. The starting point of the coded array pattern, the sub-coded pattern with code number 0, is located at the lower left corner. The code numbers of the sub-coded patterns increase by 1 from left to right. A global absolute position rectangular coordinate system XOY is established with the center of sub-coded pattern 0. Once the number of sub-coded patterns in that row is full, the row is moved up one row, and the numbers continue to increase by 1 from left to right.

[0060] In a single image captured by the camera, the pixel coordinate system is uov, with the origin located at the top-left vertex pixel of the image. To ensure that positional information is not lost during visual localization, a complete sub-encoding pattern must always exist within the camera's field of view. The design size of the four sub-encoding patterns should be smaller than the camera's field of view, such as... Figure 6 The actual width of the camera's field of view is w, and the height is h; the horizontal and vertical contour distances of the sub-coded pattern are Dx and Dy, respectively.

[0061] S4. After orthogonalization, input the program into the CNC system to perform planar motion, such as XOY plane motion. The motion range does not exceed the effective measurement range of the target. Start using this device to measure the actual motion trajectory of the spindle relative to the worktable, such as straight lines, circles, free curves, etc.

[0062] Among them, the machine tool trajectory is solved based on the array of measurement coding patterns:

[0063] The measuring head moves with the machine tool spindle, and the measuring target moves with the machine tool table, thus causing relative motion between the measuring head and the measuring target. The industrial camera in the measuring head captures each frame of the image during the motion, and each frame of the image is processed as follows:

[0064] Step 1: Basic Image Processing

[0065] For each input image frame, grayscale conversion is performed sequentially, transforming the RGB image into a grayscale image so that each pixel has a grayscale value between 0 and 255 for subsequent image processing operations. Gaussian filtering is then applied to suppress noise interference with image quality and preserve sharp boundaries as much as possible. Finally, binarization is performed to improve the algorithm's robustness to lighting conditions.

[0066]

[0067] Among them, I new I is the new grayscale value to replace the pixel coordinates (u,v). origin Here is the old grayscale value at pixel coordinates (u,v), and threshold is the segmentation threshold, which is around 200.

[0068] Step 2: Solve for the scaling factor, rotation angle, and pixel size:

[0069] Furthermore, the first frame image must contain two horizontally arranged complete circles; Canny outer contour edge detection is performed to obtain the boundary contours of the two circles I. Least squares circle fitting is then performed to solve for the center pixel coordinates (u) of the two circles I. r ,v r ), (u l ,v l ) and the average radius of circle I R b The decoding scaling ratio can be obtained by comparing the ratio of the designed radius of circle I to the actual calculated radius of circle I. The scaling ratio coefficient is used to calculate the pixel size of the size to be determined within the algorithm.

[0070] By using the pixel coordinates of the center of the circle, the angle between the line connecting the two centers relative to the horizontal direction can be obtained, which is the rotation angle of the entire coded pattern. The sign of the angle is determined by the relative position of the two centers.

[0071]

[0072] Using the design distance d real By comparing the distance to the actual pixel, the actual size (unit) of a single pixel can be obtained.

[0073]

[0074] This step is performed only in the first frame.

[0075] Step 3: Determine the position of pixel I in circle 1:

[0076] Incomplete circle I boundaries are filtered out by area and edge length, retaining only a complete circle I edge contour; least squares circle fitting is performed on the circle I edge to obtain the pixel coordinates (u) of the circle I center. b ,v b );

[0077] Step 4: ROI Region Selection:

[0078] Set the selected circle I rectangle boundary as the ROI region to improve decoding processing speed;

[0079] The pixel coordinates of the two diagonal points of the rectangular region are: (int(u b -(R b +1)),int(v b -(R b +1))),(int(u b +(R b +1)),int(v b +(R b+1))), where: int() is the integer function;

[0080] Step 5, Decoding Operation:

[0081] Canny edge detection and least squares circle fitting are performed within the ROI region. The large reference circle and the small reference circle are distinguished based on the circle radius, and the pixel coordinates of the center of the large reference circle and the center of the small reference circle are calculated as (u1,v1) and (u2,v2) respectively.

[0082] Based on the center coordinates of circle II, circle III, circle I, and the perpendicular relationship between the lines, the equations of the lines locating the X-axis (a1u + b1v + c1 = 0) and the Y-axis (a2u + b2v + c2 = 0) are solved. The coefficients a1, b1, c1, a2, b2, and c2 of the line equations are determined using the center coordinates of circle II, circle III, circle I, and their perpendicular relationships.

[0083] Next, the algorithm automatically calculates the pixel coordinates of the center of each circle IV in the small circular coding region sequentially. The method for calculating the center position of a single circle IV is as follows:

[0084] Assume the theoretical vertical pixel distances from the center of the first circle IV to the positioning X-axis and positioning Y-axis are d1 and d2, respectively, and these two distances are known in the design; let p1 and p2 represent the pixel coordinates (u) of the center of circle I. b ,v b Positional relationship with the positioning X-axis line and positioning Y-axis line:

[0085] p1 = a2 × u b +b2×u b +c2

[0086] p2=a1×u b +b1×u b +c1

[0087] Based on the expression for the distance from a point to two lines, an equation is established to obtain the pixel position (u) of the center of the first circle IV. s1 ,v s1 ):

[0088]

[0089]

[0090] By sequentially determining the center positions of the small circular coding circles at specific distances along the arrow directions, the pixel coordinates of the centers of the 16 circles IV are calculated. si ,v si ), i = 1, 2, 3…16;

[0091] The average grayscale value of the 16 coded circles IV, which are 1 pixel away from the center pixel coordinates, is determined along the direction of the arrow. If the average grayscale value is greater than a certain threshold, then the area is white and there are no black circles, representing binary "0". If the grayscale value of the position of circle IV is less than a certain threshold, then the area is black circle IV, representing binary "1".

[0092] By combining "0" and "1" in sequence according to the number of bits, the binary code of circle I can be obtained. Then, the binary code is converted into a decimal number. The remainder of number divided by the number of codes in a row (80) is n, and the remainder is m. This gives us the position of circle I in the nth row and mth column of the initial coded circle.

[0093] Step 6: Solve the absolute position of the image center frame by frame:

[0094] The actual horizontal center-to-center spacing d is designed using coding. x The longitudinal center-to-center spacing dimension d y This allows us to obtain the actual coordinates of the center of circle I relative to the first coded point:

[0095] x1=m×d x

[0096] y1=n×d y

[0097] Taking rotation angle into account, the coordinates are corrected as follows:

[0098] x2=x1cosα-y1sinα

[0099] y2=x1sinα+y1cosα

[0100] Assuming the image width is W pixels and the height is H pixels, and the pixel coordinate system is transformed into a Cartesian coordinate system, then the absolute position of the center of each frame of the image in a Cartesian coordinate system relative to the origin of the initial encoding point 0 is:

[0101]

[0102] Y = (-H / unit + v) b )×unit+y2

[0103] By continuously capturing n frames of video from the camera and performing the above processing on each frame, the set of continuous motion coordinate points (X) of the machine tool can be obtained. j ,Y j (j = 1, 2, 3...n) By connecting these coordinate points in sequence, the machine tool planar motion measurement trajectory can be obtained;

[0104] By comparing the measured planar motion trajectory with the theoretical planar motion trajectory, the error parameters of the machine tool can be analyzed, thereby enabling accuracy compensation and improving the accuracy level of the machine tool.

[0105] The above method can also perform trajectory measurement in vertical planes (YOZ plane, XOZ plane). That is, a 90-degree deflection lens is added to the camera lens to change the lens's field of view from the XOY plane to the YOZ plane or XOZ plane to achieve YOZ plane or XOZ plane imaging. Then, the plane light source of the measurement target is aligned with the lens's field of view. For example, if the lens's field of view is in the YOZ plane, the measurement target is changed from horizontal to vertical, and the plane light source is placed in the YOZ plane. After focusing the lens, the major and minor axes of the ellipse of the large circle pattern in the target are detected, and the difference between the major and minor axes of the ellipse is calculated. When this difference is less than a certain threshold, it can be determined that the optical axis of the deflection lens is orthogonal to the target plane, and subsequent trajectory measurement can begin. The subsequent operation is the same as the XOY plane trajectory measurement operation principle.

Claims

1. A micro-vision measurement method of machine tool trajectory based on encoding of small circles scattered within a large circle, characterized in that, The method comprises the following steps: S1, a measuring head is fixed on a first moving end of a machine tool, a measuring target is fixed on a second moving end of the machine tool, and the measuring head is connected to an upper computer; S2, the field of view of the measuring head is adjusted to the center of the scale range of the measuring target; S3, imaging of the upper computer is observed to complete focusing operation, an encoding pattern in the measuring target and within the field of view of the measuring head is collected, and whether the optical axis of the measuring head is orthogonal to the plane of the target is determined through detection of the encoding pattern; S4, the second moving end is moved, and the movement range does not exceed the effective measurement range of the target, so as to measure the real movement trajectory of the spindle relative to the workbench; The degrees of freedom of the first moving end at least include a longitudinal direction, and the degrees of freedom of the second moving end at least include a horizontal plane direction; the first moving end is a spindle, and the second moving end is a workbench.

2. The machine tool trajectory microscopic vision measurement method based on the small circle coding within the great circle scatter according to claim 1, characterized in that: In the step S3, the measuring head is roughly adjusted to the center of the target through X-direction and Y-direction feeding of the machine tool; Imaging of the camera in the upper computer is observed, the Z-axis of the machine tool is adjusted, the Z-axis is adjusted to minimum feeding, the Z-axis of the machine tool is moved, and when the imaging definition of the camera is unchanged, the focusing operation is completed, and then the position of the Z-axis of the machine tool is no longer changed.

3. The machine tool trajectory microscopic vision measurement method based on the small circle coding within the great circle scatter according to claim 2, characterized in that: The encoding pattern within the field of view includes at least one complete sub-encoding pattern; The one complete sub-encoding pattern includes: a circle I, two circles II and III with different diameters are arranged in the circle I, the circles II and III are respectively located on the horizontal and vertical symmetry lines of the circle I, the center distance of the circle II and the center of the circle I is a distance I, the center distance of the circle III and the center of the circle I is a distance II, and the distance I is equal to the distance II; The circle I, the circle II and the circle III are provided with a small circle encoding area for storing binary bit information.

4. The machine tool track microscopic vision measurement method based on the small circle coding within the large circle distribution according to claim 3, characterized in that: The small circle encoding area is provided with a circle IV, and the circle IV appears and does not appear to represent two numbers in binary.

5. The machine tool track microscopic vision measurement method based on the small circle coding within the large circle distribution according to claim 4, characterized in that: The lower right corner of the small circle encoding area is the lowest bit small circle encoding position, the bit number of the small circle encoding position gradually increases in the upward serpentine direction until the upper left corner is the highest bit small circle encoding position.

6. The machine tool track microscopic vision measurement method based on the small circle coding within the great circle scatter according to claim 5, characterized in that: The effective measurement range in the measuring target includes an encoding array pattern; The array encoding pattern includes 8720 sub-encoding patterns generated within a range of 150 mm*150 mm, the intervals between each sub-encoding pattern are equal, the number of encodings in a row is 80, and the number of encodings in a column is 109.

7. The machine tool track microscopic vision measurement method based on the small circle coding within the large circle distribution according to claim 3, characterized in that: The design size of the four sub-encoding patterns is not greater than the field of view size of the camera.

8. The machine tool track microscopic vision measurement method based on the small circle coding within the large circle distribution according to claim 3, characterized in that: The measuring head and the measuring target move relatively, the measuring head collects each frame of image in the movement process, and each frame of image is processed as follows: Step one, basic image processing: clear boundaries of the array encoding pattern are obtained, noise interference on image quality is suppressed, and robustness of the algorithm to illumination conditions is improved; Step two, solving scaling coefficients, rotation angles of the encoding pattern and pixel sizes, and this step is only performed at the first frame; Step three, solving the pixel position of the circle I; Step four, the selected circle I rectangular boundary is set as an ROI area to improve decoding processing speed; Step five, decoding operation, the circle I is located in n rows and m columns of the initial encoding circle; Step six, absolute position of the picture center is solved frame by frame to obtain the machine tool plane movement measurement trajectory.

9. The machine tool track microscopic vision measurement method based on the small circle coding within the large circle distribution according to claim 8, characterized in that: The step S5 of comparing the measured planar motion trajectory with the theoretical planar motion trajectory is also included.

Images