A method for measuring inner diameter and roundness based on pneumatic and visual splicing technology

Abstract

The application provides a kind of inner diameter and roundness measurement method based on pneumatic and visual splicing technology, belongs to part inner diameter measurement technology.First, local image of part is photographed by microscope multi-point and spliced, and complete image of part is obtained;initial center is obtained by edge detection and circle fitting, and part outer contour coordinates are translated to absolute coordinate system by taking it as a reference.Pneumatic method is used to measure inner diameter of part in multiple directions, and inner diameter measurement point coordinates are generated.Inner diameter measurement point and image outer contour coordinates are converted into polar coordinates respectively, corresponding contour curves are constructed, and curvature is calculated point by point, and two curve positions are registered relying on pneumatic measurement point.Combining with accuracy distribution and fusion weight of visual detection and pneumatic measurement, two contours are fused, and finally actual inner diameter and roundness parameters of part are calculated based on fused contour.The application solves the problem that discrete data of pneumatic measurement cannot cover contour details, and solves the step error problem existing in visual splicing method.

Description

Technical Field

[0001] This invention relates to the field of part inner diameter measurement technology. Specifically, this invention relates to a method for measuring inner diameter and roundness based on pneumatic and visual splicing technology. Background Technology

[0002] Shaft and hole components are core basic parts in precision machinery, aerospace, precision instruments, and automobile manufacturing. The inner diameter and roundness accuracy of their holes directly determine the assembly accuracy, motion stability, and service life of the entire machine, making them key indicators for component machining quality inspection. With the continuous iteration of precision manufacturing technology, the industry has placed higher demands on the dimensional inspection accuracy, comprehensive inspection completeness, and data stability of shaft and hole components. Traditional manual inspection and single-instrument inspection methods are no longer sufficient to meet the needs of high-precision and high-reliability batch inspection.

[0003] Chinese Patent 121520950A provides a high-precision detection device for the inner diameter of a ring-shaped part and its usage method. The high-precision detection device for the inner diameter of a ring-shaped part includes: an inner diameter measuring instrument with a measuring rod placed laterally for contacting the inner wall of the workpiece; two spaced-apart height measuring instruments for measuring their own height position, fixed to a fixed part parallel to the measuring rod; a bracket with the fixed part at its upper end and the inner diameter measuring instrument fixed at its lower end; and a detection module electrically connected to the inner diameter measuring instrument and the height measuring instruments, which calculates the inclination of the measuring rod using the height value obtained by the height measuring instrument and the distance between the two height measuring instruments, and calculates the actual inner diameter value using the preliminary inner diameter value obtained by the inner diameter measuring instrument and the inclination value.

[0004] When measuring the inner diameter using existing technology, the measurement data consists of diameters at different locations, which is a discrete measurement method. It is difficult to cover the entire circumference and cannot take all details into account. It will ignore local defect information and affect the measurement accuracy. Summary of the Invention

[0005] The present invention aims to provide a method for measuring the inner diameter and roundness based on pneumatic and visual splicing technology, so as to improve the technical objective of measuring the inner diameter of parts.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] This invention provides a method for measuring inner diameter and roundness based on pneumatic and visual stitching technology, comprising the following steps:

[0008] Step 1: Place the part to be tested under a microscope, take pictures of its multiple local features, and stitch the images of the local features together to obtain a complete image of the part to be tested;

[0009] Step 2: Perform edge detection on the image of the part to be tested to obtain the edge of the circular contour, perform circle fitting to obtain the coordinates of the first circle center, translate the coordinates of the first circle center to the origin of the absolute coordinate system, and translate all coordinates of the circular contour as a whole according to the difference between the coordinates of the first circle center and the origin of the absolute coordinate system.

[0010] Step 3: Measure the inner diameter of the part under test at multiple angles using a pneumatic measurement method, and establish the contour point coordinates of the inner diameter of the part under test on the absolute coordinate system based on the measured inner diameter at multiple angles.

[0011] Step 4: Convert the contour point coordinates of the inner diameter of the part to be measured into polar coordinates and radial relationship, construct the contour curve, and calculate the curvature at each point of the contour curve.

[0012] Step 5: Convert the coordinates of the circular profile in Step 2 to polar coordinates and radial relationship to obtain the profile curve, and calculate the curvature at each point of the profile curve; compare the curvature values ​​of the two profile curves in Step 4 and Step 5, and use the profile points obtained by pneumatic measurement as reference points to unify the positions of the two profile curves.

[0013] Step 6: Calculate the weights of the visual stitching method and the aerodynamic measurement method, and fuse the two contour curves;

[0014] Step 7: Calculate the inner diameter and roundness of the two measured pieces based on the fused contour curve.

[0015] In step one, the method for image stitching of local feature images is as follows: preprocess multiple local feature images, extract features using the Sift operator, and perform image stitching by matching features using the ratio of nearest neighbor Euclidean distance to second nearest neighbor Euclidean distance.

[0016] In step three, the method for measuring the inner diameter of the part under test at multiple angles using a pneumatic measurement method is as follows: fix the pneumatic measuring instrument probe to the shaft hole of the part under test, set the position of the first measurement to 0°, measure the corresponding diameter value, and rotate the shaft hole at a fixed angle to measure the diameter value corresponding to different positions.

[0017] In step three, the method for establishing the contour point coordinates of the inner diameter of the measured part is as follows: set the midpoint of the diameter line of the shaft hole obtained in the first measurement as the origin of the absolute coordinate system, rotate the shaft hole at a fixed angle to measure the diameter value corresponding to different positions, and set the midpoint of all diameter lines at the origin of the absolute coordinate system.

[0018] In step four, the contour points of the inner diameter of the measured part are discrete points in the polar coordinate and radial relationship diagram. The contour curve is constructed by cubic spline interpolation.

[0019] In steps four and five, the curvature at each point of the contour curve is calculated as follows: for each curve, it is described in vector equation form, and then the curvature is calculated.

[0020] In step six, the fusion method of the two contour curves is as follows: Take two adjacent reference points on the two contour lines respectively to obtain aerodynamic measurement points A and B on the contour line obtained by aerodynamic measurement, and visual measurement points C and D on the contour line obtained by visual measurement. Repeatedly measure the radius of aerodynamic measurement points A, B, C, and D to obtain the mean μ1 and standard deviation of the radius corresponding to aerodynamic measurement point A. 1. The mean μ2 and standard deviation of the radius corresponding to visual measurement point C. 2. The mean μ3 and standard deviation of the radius corresponding to aerodynamic measuring point B. 3. The mean μ4 and standard deviation of the radius corresponding to visual measurement point D. 4. First, determine the weights of the aerodynamic measurement method and the visual stitching method at the two reference points:

[0021] The weight of the aerodynamic measurement method at the first reference point is:

[0022] ;

[0023] The weights of the visual method corresponding to the first reference point are:

[0024] ;

[0025] The weight of the aerodynamic measurement method at the second reference point is:

[0026] ;

[0027] The weights of the visual method at the second reference point are:

[0028] ;

[0029] Then determine the weights of the aerodynamic measurement method and the visual stitching method at the midpoint between the two reference points;

[0030] The weight of the aerodynamic measurement method at the midpoint between the two reference points is:

[0031] ;

[0032] The weights of the visual stitching method at the midpoint between the two reference points are:

[0033] ;

[0034] The radius of the midpoint between the two reference points is calculated based on the weights of the aerodynamic measurement method and the visual stitching method.

[0035] ;

[0036] Where R j R is the radius corresponding to the midpoint measured by a visual stitching method. p This is the radius corresponding to the midpoint measured by a pneumatic measurement method.

[0037] Repeated measurements of the radius at pneumatic measuring point A, pneumatic measuring point B, visual measuring point C, and visual measuring point D yield the mean and standard deviation of the radius. The mean and standard deviation of the radius can also be recalculated using the Monte Carlo method, based on the mean and standard deviation of the radius, by setting the number of iterations and generating a random number distribution.

[0038] The radius corresponding to the midpoint measured by the pneumatic measurement method is calculated using an interpolation method.

[0039] The technical effects of this invention are as follows:

[0040] (1) Based on the discrete points of pneumatic measurement, the present invention uses cubic spline interpolation to construct the overall profile curve, which can realize the measurement value at the interpolation point. The first and second derivatives of the measurement value are continuous, ensuring that the constructed profile curve is smooth as a whole and will not have abrupt changes, which conforms to the long period characteristics of roundness error.

[0041] (2) The Monte Carlo method was used to sample the measurement samples of the two methods at the reference point. The maximum sample size was used to simulate the distribution to replace the mean and standard deviation of the finite sample, which improved the reliability of the results.

[0042] (3) The present invention calculates the weights for data fusion based on the measurement standard deviation at the reference point, which fully considers the measurement stability of the two methods and is more reliable.

[0043] (4) The present invention integrates the visual stitching method and the pneumatic measurement method, which solves the problem that the discrete data of pneumatic measurement cannot cover the contour details, and at the same time solves the step error problem of the visual stitching method. Attached Figure Description

[0044] This manual includes the following figures, which illustrate the following:

[0045] Figure 1 This is a flowchart of a method for measuring inner diameter and roundness based on pneumatic and visual stitching technology according to the present invention.

[0046] Figure 2 This is a schematic diagram of the coordinate translation of the inner diameter circle contour in the visual stitching method of the present invention, which is a method for measuring inner diameter and roundness based on pneumatic and visual stitching technologies.

[0047] Figure 3 This is a schematic diagram of a pneumatic measurement method for measuring inner diameter and roundness based on pneumatic and visual stitching technology according to the present invention.

[0048] Figure 4 This invention provides a method for measuring the inner diameter and roundness of a part based on pneumatic and visual stitching technology, which converts the coordinates of the circumferential contour points of the measured part into polar coordinates and radial relationships.

[0049] Figure 5 This is a schematic diagram showing the relationship between two contour curves for a method of measuring inner diameter and roundness based on pneumatic and visual stitching technology according to the present invention. Detailed Implementation

[0050] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, in order to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention, and to facilitate its implementation.

[0051] This invention provides a method for measuring inner diameter and roundness based on pneumatic and visual stitching technology, comprising the following steps:

[0052] Step 1: Place the part to be tested under a microscope, take pictures of its multiple local features, and stitch the images of the local features together to obtain a complete image of the part to be tested;

[0053] Step 2: Perform edge detection on the image of the part to be tested to obtain the edge of the circular contour, perform circle fitting to obtain the coordinates of the first circle center, translate the coordinates of the first circle center to the origin of the absolute coordinate system, and translate all coordinates of the circular contour as a whole according to the difference between the coordinates of the first circle center and the origin of the absolute coordinate system.

[0054] Step 3: Measure the inner diameter of the part under test at multiple angles using a pneumatic measurement method, and establish the contour point coordinates of the inner diameter of the part under test on the absolute coordinate system based on the measured inner diameter at multiple angles.

[0055] Step 4: Convert the contour point coordinates of the inner diameter of the part to be measured into polar coordinates and radial relationship, construct the contour curve, and calculate the curvature at each point of the contour curve.

[0056] Step 5: Convert the coordinates of the circular profile in Step 2 to polar coordinates and radial relationship to obtain the profile curve, and calculate the curvature at each point of the profile curve; compare the curvature values ​​of the two profile curves in Step 4 and Step 5, and use the profile points obtained by pneumatic measurement as reference points to unify the positions of the two profile curves.

[0057] Step 6: Calculate the weights of the visual stitching method and the aerodynamic measurement method, and fuse the two contour curves;

[0058] Step 7: Calculate the inner diameter and roundness of the two measured pieces based on the fused contour curve.

[0059] In step one, the method for image stitching of local feature images is as follows: preprocess multiple local feature images, extract features using the Sift operator, and perform image stitching by matching features using the ratio of nearest neighbor Euclidean distance to second nearest neighbor Euclidean distance.

[0060] In step three, the method for measuring the inner diameter of the part under test at multiple angles using a pneumatic measurement method is as follows: fix the pneumatic measuring instrument probe to the shaft hole of the part under test, set the position of the first measurement to 0°, measure the corresponding diameter value, and rotate the shaft hole at a fixed angle to measure the diameter value corresponding to different positions.

[0061] In step three, the method for establishing the contour point coordinates of the inner diameter of the measured part is as follows: set the midpoint of the diameter line of the shaft hole obtained in the first measurement as the origin of the absolute coordinate system, rotate the shaft hole at a fixed angle to measure the diameter value corresponding to different positions, and set the midpoint of all diameter lines at the origin of the absolute coordinate system.

[0062] In step four, the contour points of the inner diameter of the measured part are discrete points in the polar coordinate and radial relationship diagram. The contour curve is constructed by cubic spline interpolation.

[0063] In steps four and five, the curvature at each point of the contour curve is calculated as follows: for each curve, it is described in vector equation form, and then the curvature is calculated.

[0064] In step six, the fusion method of the two contour curves is as follows: Take two adjacent reference points on the two contour lines respectively to obtain aerodynamic measurement points A and B on the contour line obtained by aerodynamic measurement, and visual measurement points C and D on the contour line obtained by visual measurement. Repeatedly measure the radius of aerodynamic measurement points A, B, C, and D to obtain the mean μ1 and standard deviation of the radius corresponding to aerodynamic measurement point A. 1. The mean μ2 and standard deviation of the radius corresponding to visual measurement point C. 2. The mean μ3 and standard deviation of the radius corresponding to aerodynamic measuring point B. 3. The mean μ4 and standard deviation of the radius corresponding to visual measurement point D. 4. First, determine the weights of the aerodynamic measurement method and the visual stitching method at the two reference points:

[0065] The weight of the aerodynamic measurement method at the first reference point is:

[0066] ;

[0067] The weights of the visual method corresponding to the first reference point are:

[0068] ;

[0069] The weight of the aerodynamic measurement method at the second reference point is:

[0070] ;

[0071] The weights of the visual method at the second reference point are:

[0072] ;

[0073] Then determine the weights of the aerodynamic measurement method and the visual stitching method at the midpoint between the two reference points;

[0074] The weight of the aerodynamic measurement method at the midpoint between the two reference points is:

[0075] ;

[0076] The weights of the visual stitching method at the midpoint between the two reference points are:

[0077] ;

[0078] The radius of the midpoint between the two reference points is calculated based on the weights of the aerodynamic measurement method and the visual stitching method.

[0079] ;

[0080] Where R j R is the radius corresponding to the midpoint measured by a visual stitching method. p This is the radius corresponding to the midpoint measured by a pneumatic measurement method.

[0081] Repeated measurements of the radius at pneumatic measuring point A, pneumatic measuring point B, visual measuring point C, and visual measuring point D yield the mean and standard deviation of the radius. The mean and standard deviation of the radius can also be recalculated using the Monte Carlo method, based on the mean and standard deviation of the radius, by setting the number of iterations and generating a random number distribution.

[0082] The radius corresponding to the midpoint measured by the pneumatic measurement method is calculated using an interpolation method.

[0083] The following is a detailed description of a method for measuring the inner diameter and roundness based on pneumatic and visual stitching technology according to the present invention.

[0084] The part under test is placed under an electron microscope, and multiple images of its local features are taken from different angles and positions. This overcomes the limitations of the small field of view of electron microscopes, without reducing magnification, filling in blind spots of the part, balancing detail and complete imaging, and saving on the cost of large field-of-view lenses. After the images are acquired by computer, they are preprocessed to remove electron microscope noise, correct brightness and shading distortion, reduce invalid features, and improve the efficiency of subsequent matching.

[0085] Preprocessing can involve grayscale conversion, filtering and noise reduction, contrast correction, and distortion correction. Grayscale conversion involves removing redundant color channels, simplifying image data, and retaining only grayscale gradient information. Filtering and noise reduction involves filtering out Gaussian noise and salt-and-pepper noise, eliminating false edges caused by electron beam interference, and avoiding the generation of invalid false feature points. Contrast correction involves balancing the brightness difference between different image segments, unifying the brightness benchmark at different shooting angles, and reducing grayscale shifts caused by variations in electron beam intensity. Distortion correction involves minor correction of geometric distortion in microscopic imaging to ensure uniform geometric scale at the same physical location in adjacent images.

[0086] Feature extraction is performed using the SIFT operator. SIFT feature extraction is scale-, rotation-, and illumination-invariant, adapting to shooting offsets and brightness differences. Key points are concentrated at the edges and corners of parts, and defects, exhibiting good feature stability. Image stitching is completed by using the ratio of nearest neighbor Euclidean distance to second nearest neighbor Euclidean distance, resulting in a complete image of the tested part. Nearest neighbor distance ratio matching can quickly eliminate mismatches, is unaffected by image scaling and brightness, has good parameter versatility, and is computationally efficient. The principle of image stitching is as follows: using the selected precise matching feature pairs, the RANSAC algorithm is used to solve the homography transformation matrix between adjacent images, obtaining the relative translation, rotation, and scaling relationships of the segmented images; based on the transformation matrix, all images to be stitched are mapped to the reference image coordinate system, achieving image spatial alignment and precise overlap of overlapping areas; brightness smoothing and feathering weighted fusion are performed on overlapping pixels to eliminate stitching gaps and brightness discontinuities, finally synthesizing a complete full-domain image of the shaft hole.

[0087] Edge detection is performed on the stitched image to obtain the circular outline edge. The least squares method is used to fit the circle to obtain the first circle center coordinates (x0, y0). The circle center coordinates are then translated to the origin of the absolute coordinate system, and the coordinate differences between the two directions are obtained as Δx = x0 - 0, Δy = y0 - 0. All coordinates of the circular outline (x0, y0) are then calculated. i ,y i Based on the difference between the fitted center coordinates and the origin, the entire system is translated to obtain the new coordinates x. j= x i -Δx, y j= y i -Δy, such as Figure 2 As shown, the measurement and processing of the visual part are completed.

[0088] The pneumatic measurement method is used to measure the inner diameter of the part at multiple angles and calculate its coordinates, such as... Figure 3 To illustrate, set the first diameter position to zero degrees, coinciding with the x-axis, and measure the diameter value as D1. Take half of this value as the radius R1. The coordinates of the first contour point are then (x1, y1) = (R1, 0). After rotating it by Δθ, continue to measure the diameter using a pneumatic measurement method to obtain the radius R2 at the second position. The coordinates of the second contour point are then x2 = R2 × cos(Δθ), y2 = R2 × sin(Δθ). Fix the rotation angle at Δθ each time, and the coordinates of the kth point on the contour are then x1, y1, and y2. k =R k ×cos(kΔθ), y k =R k× sin(kΔθ). This invention uses fixed-angle rotation sampling, ensuring uniform angle rules and a fixed coordinate calculation formula, simplifying calculations and facilitating batch processing. It directly converts the measured radius and rotation angle into contour points, eliminating the need for complex calibration and quickly restoring the actual contour of the inner hole.

[0089] The coordinates of the circumferential contour points of the measured part obtained from pneumatic measurement are converted into polar coordinates and radial relationships, such as... Figure 4 The horizontal axis represents the measurement position θ, and the vertical axis represents the radius R corresponding to the measurement position. After obtaining a series of discrete distribution points, a contour curve is constructed using cubic spline interpolation, and the curvature at each point on the contour curve is calculated. Cubic spline interpolation generates a smooth, continuous contour curve from finite discrete measurement points, fills in the gap data between measurement points, and restores the true inner wall contour. Simultaneously, the interpolation yields a piecewise differentiable analytical expression, allowing for precise point-by-point calculation of curvature. The curvature values ​​are then used to determine burrs, dents, and roundness distortion.

[0090] The coordinates of the circular contour detected by the visual method are converted into polar coordinates and radial coordinates to obtain the contour curve, and the curvature is calculated. Specifically, for each curve, it is described in vector equation form, i.e., r = r(t) = (x(t), y(t), z(t)), according to the formula... The curvature q at each point of the two curves is calculated, and the curvature values ​​of the two contour curves corresponding to the aerodynamic measurement method and the visual stitching method are compared. Using the discrete points obtained by the aerodynamic measurement method as reference points, the positions of the two contour curves are unified, such as... Figure 5 As shown. This invention uses spatial vector equations to represent contours, adapting to contours from both visual and aerodynamic data sources. The mathematical expression is standardized, facilitating unified curvature calculation. Registration is based on aerodynamic data: aerodynamic measurements offer high accuracy, enabling spatial alignment of two curves, eliminating positional offset errors caused by installation and photography, and achieving point-to-point comparison of contours from the same source.

[0091] exist Figure 5In the diagram, positions 1 and 2 represent two adjacent reference points. Points A and B were measured using aerodynamic measurement methods, while points C and D were obtained through edge detection after visual stitching. Five repeated measurements were performed at point A, and the mean μ and standard deviation of the five sample radii were calculated. Using the Monte Carlo method with μ and Based on this, set the number of loops to 10. 6 Generate a random number distribution and recalculate the radius, mean μ1, and standard deviation. 1 represents the mean and standard deviation of the radius obtained from multiple repeated measurements by the pneumatic gauge at this point. For point C, five repeated measurements were also performed, and the mean radius μ2 and standard deviation were obtained based on the sample using Monte Carlo simulation. 2 represents the mean and standard deviation of the radius obtained from multiple repeated measurements using the visual stitching method. Using the same method, the mean radii at points B and D are obtained as μ3 and μ4, respectively, with standard deviations of [missing values]. 3 and 4. By combining five single-point field measurements with large-sample Monte Carlo simulation, the limitations of a small number of field measurements are overcome. Massive simulated data is generated based on the measured mean and standard deviation, yielding μ... It can accurately reflect the inherent random errors of both pneumatic and visual detection equipment, and precisely quantify the measurement stability of each.

[0092] For the midpoint between positions 1 and 2, the coordinates measured by the visual stitching method are (x... i ,y i The corresponding position radius is:

[0093]

[0094] The coordinates of the midpoint between two adjacent contour points measured by the pneumatic measurement method are used to obtain the coordinate point (x) using an interpolation method. p ,y p The corresponding position radius is:

[0095]

[0096] The pneumatic measurement method only has discrete measurement points. By interpolating, the aerodynamic equivalent radius at the midpoint between two reference points is calculated, realizing a one-to-one spatial correspondence between aerodynamic and visual measurement points, and solving the problem that the sampling points of the two detection methods are inconsistent and cannot be directly compared and fused.

[0097] Standard deviation measures the stability of the measurement; therefore, weights are calculated based on the standard deviations of the radii at positions 1 and 2 obtained using the visual stitching method and the pneumatic measurement method. At position 1, the weight for the pneumatic measurement method is... The weights of the visual stitching method are In position 2, the weight of the aerodynamic measurement method is The weights of the visual stitching method are The weight of the aerodynamic measurement method is obtained by taking the average of the two points. The weight of the visual stitching method is .

[0098] The radius of the measurement point at the middle position is:

[0099]

[0100] This invention integrates high-precision pneumatic measurement and large-scale visual stitching measurement data. Pneumatics suppresses system errors, and vision completes the full-domain contour. By combining the strengths of both, the fusion radius at the intermediate position has smaller errors and the contour restoration is more in line with the actual shape of the workpiece compared to a single detection result.

[0101] R l By decomposing the corresponding angles, the coordinates (x, y) of all points can be obtained. l ,y l Then, the least squares method is used for fitting to obtain the diameter and roundness of the hole, which is the result of the fusion of the two methods.

[0102] The beneficial effects of the present invention are described in detail below.

[0103] This invention integrates visual stitching and pneumatic measurement technologies, breaking through the accuracy bottleneck of single measurement methods and significantly improving the overall accuracy of inner diameter and roundness measurements. Single visual measurement is susceptible to image noise, edge blurring, and shooting angle deviations, leading to deviations in contour extraction and circle fitting, making it difficult to accurately reconstruct the actual hole contour of the part. Single pneumatic measurement can only obtain inner diameter data at discrete angular positions, failing to achieve continuous contour depiction across the entire area, resulting in data discontinuity errors in roundness calculations. This invention uses the SIFT operator to accurately extract local image features, combined with Euclidean distance matching to complete high-precision image stitching, obtaining a complete contour image of the part and ensuring the integrity of the entire visual measurement. Simultaneously, leveraging the high-precision point measurement advantages of pneumatic measurement, it obtains accurate inner diameter data from multiple angles. Through the complementary use of these two technologies, the shortcomings of each are compensated, improving measurement accuracy from the data source.

[0104] This invention employs curvature matching calibration and coordinate system unification technology to effectively eliminate system bias and ensure the alignment consistency of dual measurement data. By unifying the absolute coordinate system through center translation, this invention solves the problems of visual imaging center offset and inconsistent aerodynamic measurement references. Simultaneously, it converts the two profile curves into a polar coordinate radial relationship, calculates the global curvature using vector equations, and completes curvature comparison calibration. Using high-precision aerodynamic profile points as a reference, it unifies the hyperbola position, completely avoiding fusion errors caused by reference misalignment and profile offset between visual and aerodynamic measurement data, making the basis for subsequent data fusion more accurate and reliable.

[0105] This invention introduces a dynamic weight allocation mechanism based on the measurement standard deviation, combined with Monte Carlo simulation optimization data, to significantly improve the stability and anti-interference capability of measurement data. This invention abandons the traditional fixed-weight fusion method, allocating the weights of pneumatic and visual measurements based on the standard deviation of repeated measurements at each measuring point. The smaller the standard deviation and the higher the measurement stability, the greater the corresponding weight, accurately reflecting the actual measurement accuracy characteristics of the two measurement technologies. Simultaneously, by combining multiple repeated measurements with Monte Carlo large-cycle random number distribution simulation, the mean radius and standard deviation data are optimized, effectively reducing the random errors of single measurements, equipment fluctuation errors, and environmental interference errors, making the weight allocation more scientific and the data fitting results more closely reflect the actual size of the parts.

[0106] This invention achieves continuous and refined processing of discrete measurement data by combining cubic spline interpolation and point interpolation, solving the problem of discontinuous measurement data. Pneumatic measurement can only acquire discrete contour points at fixed angles, and while visually stitched contours are continuous, their local accuracy is insufficient. This invention utilizes cubic spline interpolation to construct smooth and continuous contour curves for discrete pneumatic measurement points. Simultaneously, it uses an interpolation algorithm to accurately solve for the pneumatic measurement radius data at the midpoint of the two reference points, filling in the gaps in the overall contour data. This enables 360° full-area contour depiction of the inner hole of a part without blind spots, freeing roundness calculations from the influence of sparse data points and incomplete contours, resulting in more comprehensive and accurate roundness detection results.

[0107] This invention features a standardized and highly adaptable measurement process, meeting the inner diameter and roundness inspection needs of various shaft and hole parts. Through fixed-angle rotation measurement, a unified coordinate system benchmark, and standardized feature extraction and matching processes, the measurement process is unaffected by part placement or local imaging deviations. The operation process is standardized and controllable, adaptable to the inspection of circular inner hole parts of different specifications and precision requirements. Simultaneously, the entire process integrates automated computer image processing, data calculation, and intelligent fusion, eliminating the need for manual intervention in the calculation process, effectively reducing human error caused by manual operation, improving measurement efficiency and inspection consistency, and making it suitable for industrial batch inspection and high-precision part sampling inspection scenarios.

[0108] The data fusion fitting algorithm of this invention further optimizes the final measurement results. It solves the weight of the intermediate point by averaging the weights of the two reference points, accurately transitions the fusion ratio of the two technologies across all positions in the entire domain, and takes into account the high-precision advantages of aerodynamic measurement and the advantages of the full-domain continuity of the contour in visual measurement. Finally, it fits the accurate contour coordinates after fusion using the least squares method. The resulting inner diameter and roundness parameters have significantly reduced errors compared to a single measurement method, and the reliability and accuracy of the measurement results are greatly improved.

[0109] The present invention has been described above by way of example with reference to the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any non-substantial improvements made using the inventive concept and technical solution; or the direct application of the inventive concept and technical solution to other situations without modification, are all within the protection scope of the present invention.

Claims

1. A method for measuring inner diameter and roundness based on pneumatic and visual stitching technology, characterized in that, Includes the following steps: Step 1: Place the part to be tested under a microscope, take pictures of its multiple local features, and stitch the images of the local features together to obtain a complete image of the part to be tested; Step 2: Perform edge detection on the image of the part to be tested to obtain the edge of the circular contour, perform circle fitting to obtain the coordinates of the first circle center, translate the coordinates of the first circle center to the origin of the absolute coordinate system, and translate all coordinates of the circular contour as a whole according to the difference between the coordinates of the first circle center and the origin of the absolute coordinate system. Step 3: Measure the inner diameter of the part under test at multiple angles using a pneumatic measurement method, and establish the contour point coordinates of the inner diameter of the part under test on the absolute coordinate system based on the measured inner diameter at multiple angles. Step 4: Convert the contour point coordinates of the inner diameter of the part to be measured into polar coordinates and radial relationship, construct the contour curve, and calculate the curvature at each point of the contour curve. Step 5: Convert the coordinates of the circular profile in Step 2 to polar coordinates and radial relationship to obtain the profile curve, and calculate the curvature at each point of the profile curve; compare the curvature values ​​of the two profile curves in Step 4 and Step 5, and use the profile points obtained by pneumatic measurement as reference points to unify the positions of the two profile curves. Step 6: Calculate the weights of the visual stitching method and the aerodynamic measurement method, and fuse the two contour curves; Step 7: Calculate the inner diameter and roundness of the two measured pieces based on the fused contour curve.

2. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 1, characterized in that: In step one, the method for image stitching of local feature images is as follows: preprocess multiple local feature images, extract features using the Sift operator, and perform image stitching by matching features using the ratio of nearest neighbor Euclidean distance to second nearest neighbor Euclidean distance.

3. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 1, characterized in that: In step three, the method for measuring the inner diameter of the part under test at multiple angles using a pneumatic measurement method is as follows: fix the pneumatic measuring instrument probe to the shaft hole of the part under test, set the position of the first measurement to 0°, measure the corresponding diameter value, and rotate the shaft hole at a fixed angle to measure the diameter value corresponding to different positions.

4. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 1, characterized in that: In step three, the method for establishing the contour point coordinates of the inner diameter of the measured part is as follows: set the midpoint of the diameter line of the shaft hole obtained in the first measurement as the origin of the absolute coordinate system, rotate the shaft hole at a fixed angle to measure the diameter value corresponding to different positions, and set the midpoint of all diameter lines at the origin of the absolute coordinate system.

5. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 1, characterized in that: In step four, the contour points of the inner diameter of the measured part are discrete points in the polar coordinate and radial relationship diagram. The contour curve is constructed by cubic spline interpolation.

6. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 1, characterized in that: In steps four and five, the curvature at each point of the contour curve is calculated as follows: for each curve, it is described in vector equation form, and then the curvature is calculated.

7. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 1, characterized in that: In step six, the fusion method of the two contour curves is as follows: Take two adjacent reference points on the two contour lines respectively to obtain aerodynamic measurement points A and B on the contour line obtained by aerodynamic measurement, and visual measurement points C and D on the contour line obtained by visual measurement. Repeatedly measure the radius of aerodynamic measurement points A, B, C, and D to obtain the mean μ1 and standard deviation of the radius corresponding to aerodynamic measurement point A.

1. The mean μ2 and standard deviation of the radius corresponding to visual measurement point C.

2. The mean μ3 and standard deviation of the radius corresponding to aerodynamic measuring point B.

3. The mean μ4 and standard deviation of the radius corresponding to visual measurement point D.

4. First, determine the weights of the aerodynamic measurement method and the visual stitching method at the two reference points: The weight of the aerodynamic measurement method at the first reference point is: ； The weights of the visual method corresponding to the first reference point are: ； The weight of the aerodynamic measurement method at the second reference point is: ； The weights of the visual method at the second reference point are: ； Then determine the weights of the aerodynamic measurement method and the visual stitching method at the midpoint between the two reference points; The weight of the aerodynamic measurement method at the midpoint between the two reference points is: ； The weights of the visual stitching method at the midpoint between the two reference points are: ； The radius of the midpoint between the two reference points is calculated based on the weights of the aerodynamic measurement method and the visual stitching method. ； Where R j R is the radius corresponding to the midpoint measured by a visual stitching method. p This is the radius corresponding to the midpoint measured by a pneumatic measurement method.

8. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 7, characterized in that: Repeated measurements of the radius at pneumatic measuring point A, pneumatic measuring point B, visual measuring point C, and visual measuring point D yield the mean and standard deviation of the radius. The mean and standard deviation of the radius can also be recalculated using the Monte Carlo method, based on the mean and standard deviation of the radius, by setting the number of iterations and generating a random number distribution.

9. The method for measuring inner diameter and roundness based on pneumatic and visual stitching technology as described in claim 7, characterized in that: The radius corresponding to the midpoint measured by the pneumatic measurement method is calculated using an interpolation method.

Images