Precise spindle five-degree-of-freedom kinematic error measurement method based on single grayscale image
Through a single grayscale image-based method, combined with a white light interference lens and an industrial camera, high-precision measurement of the motion error of five degrees of freedom of the ultra-precision air-floating spindle is achieved, and the complex and cost-effective measurement problems in the prior art are solved. The error resolution of nano-level accuracy and the accuracy measurement of micro-radian level are achieved.
Patent Information
- Application Number
- CN202310599665.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-24
- Publication Date
- 2024-12-03
AI Technical Summary
The prior art is difficult to accurately measure the five degrees of freedom motion errors of the ultra-precision air-floating spindle at the same time, and the traditional methods are complex and costly, making it difficult to avoid the impact of machining errors of standard test objects.
Using a single grayscale image measurement method, a white light interference lens and industrial camera combined with grayscale image analysis can realize high-precision measurement of radial, axial and slanting motion errors of ultra-precision air-floating spindles, simplifying measurements Devices and steps.
Achieve high-precision measurement of the five degrees of freedom motion errors of the ultra-precision air-floating spindle, simplifies the measurement steps and devices, reduces costs, and can distinguish in-plane jump errors and out-of-plane swings at micro-radian levels in nano-level accuracy. error.
Smart Images

Figure CN119063623A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of ultra-precision engineering and relates to a measurement method of a precision spindle five-degree-of-freedom motion error measurement system based on a single grayscale image. Background Art
[0002] With the development of emerging technologies such as microelectronics, aerospace, and bioengineering, the requirements for part machining accuracy are becoming increasingly stringent, gradually evolving to above the micron level and moving towards the nanometer and sub-nanometer levels. Ultra-precision machine tools are essential to meeting the precision requirements of ultra-precision machining. As a core component of machine tools, the motion error of the ultra-precision air-bearing spindle is one of the most important factors affecting machining accuracy. Therefore, measuring the motion error of the ultra-precision air-bearing spindle is crucial for improving machining accuracy.
[0003] Any rotating object has six degrees of freedom in space: three linear motions and three rotational motions. For a rotating spindle, except for the established motion of the spindle itself rotating around the axis mean line, the motions in the remaining five directions are error motions, including:
[0004] Radial error motion: The runout of the spindle in the direction perpendicular to the axis mean line, which includes two degrees of freedom and is usually called runout in the x and y directions;
[0005] Axial error motion: Runout coaxial with the axis mean line, including one degree of freedom, usually called runout in the z direction;
[0006] Yaw error motion: The yaw motion that makes an angle with the axis mean line contains two degrees of freedom. These two error motions are usually called yaw around the x and y directions.
[0007] With the continuous development of sensor technology and the use of sensors such as capacitive displacement sensors and laser displacement sensors, the measurement accuracy of spindle motion errors is constantly improving. However, usually, the measurement of spindle motion errors can only measure a certain error motion at a time. It is more difficult to accurately measure five error motions at the same time. It often requires multiple standard test objects and multiple sensors to work together to complete the measurement.
[0008] Commonly used standard test objects include test balls and test cylinders. However, when using standard test objects, there is an inherent problem in that the machining errors of the test objects can be coupled with the rotational errors of the spindle. Only when the roundness error of the test ball is very small relative to the spindle motion error, usually at least an order of magnitude smaller, can it be ignored, and the measured displacement data is assumed to be derived entirely from the spindle motion error. However, the motion errors of ultra-precision air-bearing spindles are at the nanometer level, and standard test objects with machining errors far smaller than this level are extremely difficult to obtain. Therefore, the machining errors cannot be ignored in this case, and error separation technology is required to separate the actual ultra-precision air-bearing spindle motion errors.
[0009] Error separation can usually be accomplished through multiple detectors or multiple positioning. Multiple probes are used to simultaneously measure and separate motion errors. In 1966, Yasuo Aoki and Shigeo Ohen of Japan proposed a three-sensor method (i.e., three-point method) for measuring the roundness error of parts. This method relies on the relative position accuracy of multiple sensors, otherwise it will cause harmonic suppression problems and result in poor measurement accuracy. The reversal method is commonly used for multiple positioning. In 1972, Donaldson of the United States used a sensor to measure a standard ball in two steps. After collecting one set of data, the standard ball and sensor were simultaneously rotated 180 degrees around their axis and installed. Another set of data was collected. The two sets of displacement data were combined and processed to obtain the spindle motion error and the test ball roundness error. Similarly, the position accuracy of each rotation must be ensured.
[0010] In the above methods, the measurement accuracy is not high because the relative position accuracy and sensor consistency of the sensors are difficult to guarantee. In order to directly avoid the influence of the processing error of the standard measurement object, the optical measurement method is used to measure the spindle motion. Peizhi Jia used laser alignment technology to measure the radial error and runout error, and its axial error needed to be measured with the help of heterodyne interference; HeNi also used laser alignment technology to measure the runout error, but needed to use a confocal microscope to measure the axial error, and then measure the mirror error by micro-magnifying the laser spot. Although the optical measurement method can avoid the problem of processing errors of standard measurement objects, it requires the use of different measurement methods and different optical paths to complete the measurement of motion errors of multiple degrees of freedom. As a result, the measurement optical path contains a large number of optical components and sensors, and the measurement system is complex. It is necessary to carefully calibrate the relevant components and equipment to reduce measurement errors.
[0011] With the development of computer vision technology, visual measurement has enabled the measurement of spindle motion errors. Compared to other optical measurement methods, visual measurement requires simpler equipment and is less expensive. Jin A measured the spindle's radial error by recording the motion trajectory of a luminous dot with an industrial camera. Fu Pengqiang processed interference fringes and determined the spindle's runout error based on their inclination and distance. Because visual processing only uses two-dimensional images, which contain limited spatial motion information, it is difficult to simultaneously measure all five degrees of freedom (DOF) of motion errors. Summary of the Invention
[0012] The purpose of the present invention is to provide a method for measuring the five-degree-of-freedom motion error of a precision spindle based on a single grayscale image, which simplifies the measuring device and steps while ensuring that the five-degree-of-freedom motion error of an ultra-precision air-floating spindle can be measured with high precision.
[0013] The technical solution adopted in the present invention is:
[0014] A precision spindle five-degree-of-freedom motion error measurement system based on a single grayscale image, the measurement system includes an ultra-precision air-floating spindle, a test workpiece, a white light interference lens, an adapter, an industrial camera, a computer, and a fiber optic cold light source. The test workpiece is fixed to the ultra-precision air-floating spindle, the white light interference lens is arranged on one end of the adapter and is arranged opposite to the test workpiece, the fiber optic cold light source is connected to the white light interference lens, and the industrial camera is located on the other end of the adapter.
[0015] An annular machining groove is provided on the end surface of the test workpiece opposite to the white light interference lens, and the inner diameter of the annular machining groove is smaller than the field of view of the industrial camera.
[0016] According to the above-mentioned measurement method of the precision spindle five-degree-of-freedom motion error measurement system based on a single grayscale image, the method includes the following steps:
[0017] Step 1: Fine-tune the relative position between the test workpiece and the industrial camera to minimize the angle between the rotation axis of the ultra-precision air bearing spindle and the optical axis, so that the white light interference fringes produced by the white light interference lens appear only one level on the entire test workpiece surface. After completing the fine-tuning, in order to obtain the relationship between the grayscale and height of the pixel points, control the white light interference lens to move at a constant speed from near to far, and simultaneously capture images at a fixed frame rate. The relationship between grayscale and height is established based on the change in grayscale in each frame and the height position corresponding to each frame.
[0018] Step 2: After all the preparations are completed, fix the position of the industrial camera lens and start the ultra-precision air-floating spindle to rotate at a constant speed. The speed of the ultra-precision air-floating spindle is known. When the speed of the ultra-precision air-floating spindle remains stable, the computer starts the acquisition program of the industrial camera. The acquired image includes the acquisition time of the current image. The accurate position of the ultra-precision air-floating spindle at this time can be obtained through this time and the speed of the ultra-precision air-floating spindle.
[0019] After the ultra-precision air-bearing spindle rotates for several cycles, the industrial camera acquisition program is stopped and the acquired images are pre-processed with Gaussian filters to reduce the impact of image noise;
[0020] Step 3: After image preprocessing, the Canny edge detection algorithm is used to extract edges. The horizontal and vertical gradient values and directions of the image are calculated. Strong and weak edges are determined based on the set gradient threshold. Non-maximum suppression is then used to retain the required true edge points.
[0021] Step 4: By performing least squares circle fitting on the edge points detected by the Canny edge detection algorithm, the change in the center position of the fitted circle reflects the radial motion error of the ultra-precision air-bearing spindle. However, due to the existence of the circular contour machining error, when the ultra-precision air-bearing spindle rotates, the circular motion of the least squares fitting circle will add a fixed radius of rotation around the center of rotation of the ultra-precision air-bearing spindle to the radial motion error of the ultra-precision air-bearing spindle. At this time, the movement of the center of the circle in the x and y directions is as follows:
[0022] x=δx+ρ·cos(ωt+φ)+x m , y=δy+ρ·sin(ωt+φ)+y m ,
[0023] Where δx and δy are the radial motion errors in the plane of the ultra-precision air-floating spindle, ρ is the radius of the rotation of the fitting circle around the rotation center of the ultra-precision air-floating spindle, ω is the speed of the ultra-precision air-floating spindle, φ is the initial phase, and x is the rotation speed of the ultra-precision air-floating spindle. m and y m The position of the rotation center of the ultra-precision air bearing spindle in the image coordinate system is obtained by fitting the measured data points using sine and cosine fitting to obtain ρ, φ, and x. m and y m The sine and cosine components of the equations are removed from the center of the circle to obtain the radial motion errors δx and δy of the ultra-precision air-bearing spindle.
[0024] Step 5: After measuring the radial motion error, obtain the position x of the rotation center of the ultra-precision air-bearing spindle in each grayscale image. m and y mSince the workpiece clamped on the ultra-precision air-bearing spindle can be considered to be rigidly connected to the ultra-precision air-bearing spindle, the yaw motion error of the ultra-precision air-bearing spindle is equal to the inclination change of the workpiece end face. Because the white light interference lens produces white light interference fringes on the workpiece end face that are related to height, the height information of each point on the workpiece surface can be directly obtained from the grayscale information of the image. The plane equation obtained by fitting the plane where each point lies can be used to obtain the yaw motion error of the ultra-precision air-bearing spindle through the change of the plane normal vector.
[0025] In order to simplify the calculation, only a circle of pixel points with the rotation center of the ultra-precision air bearing spindle as the center is selected as the data point fitting plane. The plane equation is as follows:
[0026] z=Ax+By+C
[0027] At this time, the plane normal vector is (A, B, -1). Since the yaw motion error of the ultra-precision air-floating spindle is at the micro-radian level, the components δα and δβ in the x and y directions are approximately equal to A and B. Similarly, due to the influence of surface processing errors, when the ultra-precision air-floating spindle rotates, the change in the normal vector inclination angle obtained by fitting the plane will add a fixed-size rotational motion around the ultra-precision air-floating spindle's rotation axis to the yaw motion error of the ultra-precision air-floating spindle. At this time, the normal vector inclination angle changes α and β are as follows:
[0028] α=δα+ρ θ ·cos(ωt+φ)+α m , β=δβ+ρ θ sin(ωt+φ)+β m
[0029] Where, ρ θ is the magnitude of the rotational motion of the normal vector of the fitting plane around the axis of rotation of the ultra-precision air bearing spindle, ω is the speed of the ultra-precision air bearing spindle, φ is the initial phase, α m and β m The fixed deviation is caused by the misalignment between the camera coordinate system and the ultra-precision air-floating spindle coordinate system. Sine and cosine fitting is performed on the data of α and β respectively to fit ρ θ ,φ,α m and β m The sine and cosine components composed of the components are removed from the yaw motion of the normal vector to obtain the yaw motion errors δα and δβ of the ultra-precision air-bearing spindle;
[0030] The axial motion error of the ultra-precision air-floating spindle is obtained through the height change of the center of rotation. The coordinates of the center of rotation are directly substituted into the fitted plane equation. The obtained z is the height change of the center of rotation, which is also the axial motion error of the ultra-precision air-floating spindle.
[0031] The beneficial effects of the present invention are as follows: the present invention can obtain the five-degree-of-freedom motion error of the ultra-precision air-floating spindle at that position by analyzing and processing a single grayscale image information of the workpiece surface; the measurement steps are simple, and the five-degree-of-freedom motion error of the ultra-precision air-floating spindle movement can be obtained by analyzing and processing the collected grayscale image, without the need for additional operations to separate processing errors; the measuring device is simple and low-cost, and through a workpiece that has undergone simple surface processing and an industrial camera with a white light interference lens, there is no need to use expensive high-precision standard parts or complex measuring devices; the measurement accuracy is high, and the in-plane runout error with nanometer-level accuracy can be obtained through the edge features of the workpiece surface on the grayscale image, and the out-of-plane runout error at the nanometer level and the out-of-plane yaw error at the micro-radian level can be obtained through the distribution of light intensity information.
[0032] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 It is a structural schematic diagram of the ultra-precision air-bearing spindle five-degree-of-freedom motion error synchronous measurement system based on a single grayscale image of the present invention;
[0034] Figure 2 This is a flow chart of a method for synchronously measuring the five-degree-of-freedom motion error of an ultra-precision air-bearing spindle based on a single grayscale image of the present invention;
[0035] Figure 3 It is a schematic diagram of testing the surface features of a workpiece in the method for synchronously measuring the five-degree-of-freedom motion error of an ultra-precision air-bearing spindle based on a single grayscale image of the present invention.
[0036] Figure 4 It is a schematic diagram of a grayscale image collected in the method for synchronously measuring the five-degree-of-freedom motion error of an ultra-precision air-bearing spindle based on a single grayscale image of the present invention.
[0037] Figure 5 It is a schematic diagram of the coordinate system involved in the method for synchronously measuring the five-degree-of-freedom motion error of an ultra-precision air-bearing spindle based on a single grayscale image of the present invention.
[0038] In the figure, 1. Ultra-precision air-bearing spindle, 2. Test workpiece, 3. White light interference lens, 4. Adapter, 5. Industrial camera, 6. Computer, 7. Fiber optic cold light source, 8. Processing groove, 9. Actual shooting field of view. DETAILED DESCRIPTION
[0039] The preferred embodiment of the present invention is shown in Figure 1 and Figure 2 First, a relatively simple special surface needs to be processed as the object to be measured. The surface features can be seen Figure 3 The typical processing steps for this surface feature are to first flatten the end of a round copper rod and then cut a groove 8 into the surface. This process is relatively easy. The inner diameter of the circular groove 8 needs to be slightly smaller than the field of view of the industrial camera 5 to ensure that the circular edge remains within the camera's field of view during measurement. Before measurement begins, the test workpiece 2 is secured to the ultra-precision air bearing spindle 1 via vacuum suction.
[0040] Because the ultra-precision air-bearing spindle's runout error and axial runout error in this method require spatial height information for each point in the grayscale image, and grayscale images only contain the two-dimensional position information of each pixel, we use a white-light interferometer lens to establish a relationship between grayscale and height. When the stable white light generated by the fiber-optic cold light source 7 passes through the white-light interferometer lens 3, regular white-light interference fringes are produced at the point where the optical path difference is zero. The grayscale variation of these fringes is directly related to the height between the white-light interferometer lens 3 and the workpiece 2. This allows us to establish a relationship between the grayscale and height of each point in the image.
[0041] In order to improve the measurement accuracy of axial motion error and yaw motion error, we need to fine-tune the relative position between the workpiece 2 and the camera 5 so that the angle between the rotation axis of the ultra-precision air-floating spindle 1 and the optical axis is as small as possible, so that the white light interference fringes have only one level on the entire workpiece surface. Because the height corresponding to the width of each level of the white light interference fringes is only a few hundred nanometers, when the axial motion error of the ultra-precision air-floating spindle is at the nanometer level and the yaw motion error is at the micro-radian level, smaller motion errors will bring about larger changes in the surface grayscale, greatly improving the measurement resolution. After fine-tuning, the grayscale image captured by the industrial camera is as follows: Figure 4 As shown, the white part is the surface of the processed frustum, the black part is the processed groove part, and the edge at the junction of black and white is sharp and clear.
[0042] After fine-tuning, in order to obtain the relationship between the grayscale and height of the pixel points, we can control the white light interferometer lens 3 to move at a constant speed from near to far, while capturing images at a fixed frame rate. The relationship between grayscale and height can be established based on the grayscale changes in each frame and the corresponding height position of each frame.
[0043] After all preparatory steps are complete, the camera lens is fixed in position and the ultra-precision air bearing spindle 1 is started to rotate at a constant speed. Once the speed of the ultra-precision air bearing spindle 1 remains stable, computer 6 initiates the acquisition process for industrial camera 5. The captured grayscale image includes the acquisition time of the current image. This time, combined with the ultra-precision air bearing spindle speed, allows the precise position of the ultra-precision air bearing spindle to be determined.
[0044] After the ultra-precision air-bearing spindle 1 rotates for multiple cycles, the acquisition program of the industrial camera 5 is stopped, and the acquired image is pre-processed by Gaussian filtering to reduce the influence of image noise;
[0045] After image preprocessing is completed, the Canny edge detection algorithm is used to extract edges, calculate the gradient value and gradient direction in the horizontal and vertical directions of the image, determine strong edges and weak edges based on the set gradient threshold, and then retain the required real edge points based on non-maximum suppression.
[0046] By performing least squares circle fitting on the edge points detected by the Canny edge detection algorithm, the change in the center position of the fitted circle reflects the radial motion error of the ultra-precision air-bearing spindle. However, due to the existence of the machining error of the circular contour, when the ultra-precision air-bearing spindle rotates, the circular motion of the least squares fitting circle will add a fixed radius of rotation around the center of rotation of the ultra-precision air-bearing spindle to the radial motion error of the ultra-precision air-bearing spindle. At this time, the movement of the center of the circle in the x and y directions is as follows:
[0047] x=δx+p·cos(ωt+φ)+x m , y=δy+ρ·sin(ωt+φ)+y m ,
[0048] Where δx and δy are the radial motion errors in the plane of the ultra-precision air-floating spindle, ρ is the radius of the rotation of the fitting circle around the rotation center of the ultra-precision air-floating spindle, ω is the speed of the ultra-precision air-floating spindle, φ is the initial phase, and x is the rotation speed of the ultra-precision air-floating spindle. m and y m The position of the rotation center of the ultra-precision air bearing spindle in the image coordinate system is obtained by fitting the measured data points using sine and cosine fitting to obtain ρ, φ, and x. m and y m The sine and cosine components of the equations are removed from the center of the circle to obtain the radial motion errors δx and δy of the ultra-precision air-bearing spindle.
[0049] It is worth noting that the radial motion error detected by this method is the projection of the actual radial motion error of the ultra-precision air-bearing spindle in the camera coordinate system. This method involves three coordinate systems, namely the ultra-precision air-bearing spindle coordinate system, the workpiece coordinate system, and the camera coordinate system. The coordinate system diagram is shown in Figure 5 Due to the fine-tuning before measurement, the deflection angles between the three coordinate systems are all below 100 microradians. At this time, the projection error caused by the misalignment of the coordinate systems is below 0.001 nanometers, which is much smaller than the radial motion error of the ultra-precision air-bearing spindle. Therefore, the influence of the misalignment of the coordinate systems can be ignored.
[0050] After measuring the radial motion error, the position x of the rotation center of the ultra-precision air-bearing spindle in each grayscale image is obtained. m and y mSince the workpiece is clamped on the ultra-precision air-bearing spindle and can be regarded as rigidly connected to the ultra-precision air-bearing spindle, the yaw motion error of the ultra-precision air-bearing spindle is equal to the inclination change of the workpiece end face. Because the white light interference lens produces white light interference fringes related to the height on the workpiece end face, the height information of each point on the workpiece surface can be directly obtained through the grayscale information of the image. The plane equation obtained by fitting the plane where each point is located can be used to obtain the yaw motion error of the ultra-precision air-bearing spindle through the change of the plane normal vector.
[0051] In order to simplify the calculation, only a circle of pixel points with the rotation center of the ultra-precision air bearing spindle as the center is selected as the data point fitting plane. The plane equation is as follows:
[0052] z=Ax+By+C
[0053] At this time, the plane normal vector is (A, B, -1). Since the yaw motion error of the ultra-precision air-floating spindle is at the micro-radian level, the components δα and δβ in the x and y directions are approximately equal to A and B. Similarly, due to the influence of surface processing errors, when the ultra-precision air-floating spindle rotates, the change in the normal vector inclination angle obtained by fitting the plane will add a fixed-size rotational motion around the ultra-precision air-floating spindle's rotation axis to the yaw motion error of the ultra-precision air-floating spindle. At this time, the normal vector inclination angle changes α and β are as follows:
[0054] α=δα+ρ θ ·cos(ωt+φ)+α m , β=δβ+ρ θ sin(ωt+φ)+β m
[0055] Where, ρ θ is the magnitude of the rotational motion of the normal vector of the fitting plane around the axis of rotation of the ultra-precision air bearing spindle, ω is the speed of the ultra-precision air bearing spindle, φ is the initial phase, α m and β m The fixed deviation is caused by the misalignment between the camera coordinate system and the ultra-precision air-floating spindle coordinate system. Sine and cosine fitting is performed on the data of α and β respectively to fit ρ θ ,φ,α m and β m The sine and cosine components composed of the two components are removed from the yaw motion of the normal vector to obtain the yaw motion errors δα and δβ of the ultra-precision air-bearing spindle.
[0056] The axial motion error of the ultra-precision air-bearing spindle is determined by the height change of the center of rotation. Simply substituting the coordinates of the center of rotation into the fitted plane equation yields z, the resulting height change and, therefore, the axial motion error of the ultra-precision air-bearing spindle. Similarly, this height change is the projection of the ultra-precision air-bearing spindle's axial motion into the camera coordinate system. However, because the angle between the ultra-precision air-bearing spindle coordinate system and the camera coordinate system is only a few hundred microradians, the resulting error is less than 0.001 nanometers, far smaller than the axial motion error and can be ignored.
[0057] Through the above method, ultra-precision five-degree-of-freedom motion error measurement can be completed simultaneously through only a single grayscale image, and nanometer-level resolution of axial motion error and radial motion error, and micro-radian-level resolution of tilt motion error can be achieved. The device is simple and easy to operate, which greatly reduces the cost of ultra-precision air-floating spindle motion error measurement.
[0058] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.
Claims
1. A precision spindle five-degree-of-freedom motion error measurement system based on a single grayscale image, characterized in that: The measuring system comprises an ultra-precision air-floating spindle (1), a test workpiece (2), a white light interference lens (3), an adapter (4), an industrial camera (5), a computer (6), and an optical fiber cold light source (7); the test workpiece (2) is fixed on the ultra-precision air-floating spindle (1); the white light interference lens (3) is arranged on one end of the adapter (4) and is arranged opposite to the test workpiece (2); the optical fiber cold light source (7) is connected to the white light interference lens (3); and the industrial camera (5) is located on the other end of the adapter (4). An annular machining groove (8) is provided on the end surface of the test workpiece (2) opposite to the white light interference lens (3), and the inner diameter of the annular machining groove (8) is smaller than the field of view of the industrial camera (5).
2. The measurement method of the precision spindle five-degree-of-freedom motion error measurement system based on a single grayscale image according to claim 1 is characterized in that: The method comprises the following steps: Step 1: fine-tune the relative position between the test workpiece (2) and the industrial camera (5) so that the angle between the rotation axis of the ultra-precision air-floating spindle (1) and the optical axis is as small as possible, so that the white light interference fringes generated by the white light interference lens appear only once on the entire surface of the test workpiece (2). After the fine-tune is completed, in order to obtain the relationship between the grayscale and the height of the pixel point, the white light interference lens (3) is controlled to move at a uniform speed from near to far, and pictures are collected at a fixed frame rate at the same time, and the relationship between the grayscale and the height is established according to the change of the grayscale in each frame of the photo and the height position corresponding to each frame of the photo; Step 2: After all the preparations are completed, fix the lens position of the industrial camera (5), start the ultra-precision air-floating spindle (1) to rotate at a constant speed, the speed of the ultra-precision air-floating spindle (1) is known, and when the speed of the ultra-precision air-floating spindle (1) remains stable, the computer (6) starts the acquisition program of the industrial camera (5), and the acquired image includes the acquisition time of the current image. The accurate position of the ultra-precision air-floating spindle at this time is obtained through the time and the speed of the ultra-precision air-floating spindle. After the ultra-precision air-floating spindle (1) rotates for multiple cycles, the industrial camera (5) acquisition program is stopped, and the acquired image is pre-processed by Gaussian filtering to reduce the influence of image noise; Step 3: After the image is preprocessed, the Canny edge detection algorithm is used to extract the edge. The gradient value and gradient direction of the image in the horizontal and vertical directions are calculated. The strong edge and weak edge are determined according to the set gradient threshold, and the required real edge points are retained according to non-maximum suppression. Step 4: By performing least squares circle fitting on the edge points detected by the canny edge detection algorithm, the change in the center position of the fitted circle reflects the radial motion error of the ultra-precision air-floating spindle. Due to the existence of the machining error of the circular contour, when the ultra-precision air-floating spindle rotates, the circular motion of the least squares fitting circle will add a fixed radius of rotation around the rotation center of the ultra-precision air-floating spindle to the radial motion error of the ultra-precision air-floating spindle. At this time, the movement of the center of the circle in the x and y directions is as follows: x=δx+ρ·cos(ωt+φ)+x m, y=δy+ρ·sin(ωt+φ)+y m , Where δx and δy are the radial motion errors in the plane of the ultra-precision air-floating spindle, ρ is the radius of the rotation of the fitting circle around the rotation center of the ultra-precision air-floating spindle, ω is the speed of the ultra-precision air-floating spindle, φ is the initial phase, and x is the rotation speed of the ultra-precision air-floating spindle. m and m is the position of the rotation center of the ultra-precision air bearing spindle in the image coordinate system. By performing sine and cosine fitting on the measured data points, ρ, φ, and x are fitted. m and m The sine and cosine components composed of the two components are removed from the motion of the center of the circle to obtain the radial motion errors δx and δy of the ultra-precision air-bearing spindle. Step 5: After measuring the radial motion error, obtain the position x of the rotation center of the ultra-precision air-floating spindle in each grayscale image. m and m , since the workpiece is clamped on the ultra-precision air-floating spindle and can be regarded as a rigid connection with the ultra-precision air-floating spindle, the yaw motion error of the ultra-precision air-floating spindle is equal to the inclination change of the workpiece end face. Because the white light interference lens produces white light interference fringes related to the height on the workpiece end face, the height information of each point on the workpiece surface is directly obtained through the grayscale information of the image, and the plane equation obtained by fitting the plane where each point is located is obtained. The yaw motion error of the ultra-precision air-floating spindle is obtained through the change of the plane normal vector; A circle of pixel points with the rotation center of the ultra-precision air bearing spindle as the center is selected as the data point fitting plane. The plane equation is as follows: z=Ax+By+C At this time, the plane normal vector is (A, B, -1). Since the yaw motion error of the ultra-precision air-floating spindle is at the micro-radian level, the components δα and δβ in the x and y directions are approximately equal to A and β. Similarly, due to the influence of surface machining errors, when the ultra-precision air-floating spindle rotates, the change in the normal vector inclination obtained by the fitting plane will add a fixed size of rotational motion around the ultra-precision air-floating spindle rotation axis to the yaw motion error of the ultra-precision air-floating spindle. At this time, the normal vector inclination changes α and β are as follows: a=da+r θ ·cos(ωt+φ)+α m ,β=δβ+ρ θ ·sin(ωt+φ)+β m In the formula, ρ θ is the magnitude of the rotational motion of the normal vector of the fitting plane around the axis of rotation of the ultra-precision air bearing spindle, ω is the speed of the ultra-precision air bearing spindle, φ is the initial phase, α m and β m The fixed deviation is caused by the misalignment between the camera coordinate system and the ultra-precision air-floating spindle coordinate system. Sine and cosine fitting is performed on the data of α and β respectively to fit ρ θ , φ, α m and β m The sine and cosine components composed of the yaw motion of the normal vector are removed to obtain the yaw motion errors δα and δβ of the ultra-precision air-bearing spindle. The axial motion error of the ultra-precision air-floating spindle is obtained through the height change of the center of rotation. The coordinates of the center of rotation are directly substituted into the fitted plane equation. The obtained z is the height change of the center of rotation, which is also the axial motion error of the ultra-precision air-floating spindle.
