Ultrasonic machining surface microstructure multi-scale analysis method

Abstract

The invention belongs to the technical field of ultrasonic machining, and discloses an ultrasonic machining surface micro-texture multi-scale analysis method, which comprises the following steps: acquiring three-dimensional morphology point cloud data of a machining surface; converting the point cloud data into a two-dimensional matrix and preprocessing the two-dimensional matrix to remove noise and holes; performing multi-scale decomposition on the preprocessed matrix by adopting wavelet transform to obtain low-frequency and high-frequency signals of different levels; reconstructing the low-frequency and high-frequency three-dimensional morphology of each level based on the decomposition result; calculating the three-dimensional surface roughness Sa of each reconstructed morphology; and in combination with the spatial frequency range and the Sa value corresponding to each level, contributions of microtexture components with different spatial frequencies to the surface roughness are analyzed, so that dominant frequency components influencing the surface roughness are determined. According to the method, multi-scale quantitative characterization of the surface microtexture is realized, key scale components which play a leading role in roughness can be accurately identified, and a reliable basis is provided for process optimization.

Description

Technical Field

[0001] This application belongs to the field of ultrasonic processing technology, specifically relating to a multi-scale analysis method for the microtexture of ultrasonically processed surfaces. Background Technology

[0002] Ultrasonic machining technology, with its advantages of high cutting efficiency and excellent surface integrity, is widely used in high-performance machining. This technology transforms continuous cutting in traditional machining into periodic intermittent cutting by causing the tool to vibrate at high frequencies. This results in tool trajectories exhibiting complex features such as spatial waves or ellipses, thereby forming microtexture morphologies with specific directions and periods on the workpiece surface.

[0003] Currently, the evaluation of surface quality in ultrasonic machining still primarily relies on macroscopic morphological parameters such as surface roughness. However, surface roughness, as a highly generalized statistical quantity, represents the comprehensive result of the superposition of all scale characteristics of the surface. In ultrasonic machining, the presence of surface microtextures often leads to an increase in overall roughness, which, according to traditional quality evaluation standards, would result in a decline in machining quality. However, practice shows that microtextured surfaces with specific geometric characteristics can actually improve the tribological properties, lubrication retention, and fatigue life of parts under many working conditions. Therefore, relying solely on surface roughness as a macroscopic parameter to evaluate the surface performance of ultrasonically machined parts has significant limitations and inadequacies.

[0004] Existing research on surface microtexture analysis mostly focuses on qualitative descriptions of its overall morphology, lacking in-depth analysis of its intrinsic multi-scale structure. A processed surface is typically composed of superimposed morphological components of different spatial frequencies, encompassing low-frequency components reflecting macroscopic geometric contours, mid-frequency components reflecting cutting marks and main vibration characteristics, and high-frequency components reflecting the material's microscopic removal mechanisms. Current techniques struggle to effectively separate and extract these morphological features at different scales, let alone quantify the specific contributions of different scale components to surface functional properties (such as roughness). Therefore, this field requires a method capable of multi-scale quantitative characterization and analysis of surface microtexture to address how to decompose surface morphology at different scales and establish the correlation between morphological features at specific scales and surface performance parameters. Summary of the Invention

[0005] To address the shortcomings of existing technologies in analyzing the contributions of components at various scales, this application provides a multi-scale analysis method for microtexture on ultrasonically processed surfaces. This method can separate and extract the surface microtexture morphology in a multi-scale space and accurately quantify the distribution and evolution of morphological features at different scales, thereby revealing the contribution of components at each scale to surface roughness and providing a technical basis for accurately identifying key scale features that have a dominant influence on surface properties.

[0006] To achieve the above technical objectives, this application specifically adopts the following technical solution: In one aspect of this application, a multi-scale analysis method for the microtexture of ultrasonically processed surfaces is provided, comprising the following steps: S1. Obtain the three-dimensional topographic point cloud data of the ultrasonically processed surface; S2. Convert the three-dimensional topographic point cloud data into a two-dimensional matrix, and preprocess the two-dimensional matrix to remove noise and holes; S3. Wavelet transform is used to decompose the preprocessed two-dimensional matrix into multi-scale signals to obtain low-frequency and high-frequency signals at different decomposition levels. S4. Based on the results of the multi-scale decomposition, reconstruct the low-frequency and high-frequency three-dimensional morphology of each decomposition level; S5. Calculate the three-dimensional surface roughness Sa of the three-dimensional morphology reconstructed at each decomposition level; S6. Combining the spatial frequency range corresponding to each decomposition level with the three-dimensional surface roughness Sa, analyze the contribution of micro-texture components with different spatial frequencies to the surface roughness, so as to determine the dominant frequency components affecting the surface roughness.

[0007] In one implementation, the preprocessing in step S2 includes: performing region segmentation on the two-dimensional matrix, and setting differentiated binarization thresholds based on the grayscale histograms of each region to generate a mask matrix, thereby removing outliers and empty regions.

[0008] In one implementation, the wavelet transform in step S3 employs a bioorthogonal wavelet basis.

[0009] In one implementation, the biorthogonal wavelet basis is the bior6.8 wavelet basis.

[0010] In one implementation, the wavelet transform decomposition process is as follows: for each decomposition level, through the wavelet function... and scaling function With two-dimensional matrix respectively Perform convolution operations to calculate the high-frequency detail coefficients G and low-frequency approximation coefficients D for this layer:

[0011] in, a and b The dimension of the two-dimensional matrix. coordinates Topographic data of the location.

[0012] In one implementation, during the wavelet decomposition process in step S3, the low-frequency signal at any decomposition level... With high frequency signals Satisfy: Original signal It can be obtained from the deepest low-frequency signal and the first to the second layer. The high-frequency signals of the layer are completely reconstructed, that is: .

[0013] In one implementation, the number of wavelet decomposition levels in step S3 Determined by the following formula:

[0014] in, The center frequency of the wavelet. The sampling interval is... This is the cutoff wavelength for high-frequency information.

[0015] In one implementation, the high-frequency signal in step S3 includes high-frequency sub-signals in the horizontal, vertical, and diagonal directions, as follows:

[0016] in, In the horizontal direction, Vertical direction It is the diagonal direction.

[0017] In one implementation, step S4 involves weighting and superimposing the high-frequency sub-signals in the horizontal, vertical, and diagonal directions to reconstruct the comprehensive high-frequency three-dimensional morphology of each level.

[0018] In one implementation, during the reconstruction of the three-dimensional topography in step S4, a grid interpolation method is used to resample the coefficients after wavelet decomposition to generate a feature matrix with the same size as the original data.

[0019] In one implementation, the formula for calculating the three-dimensional surface roughness Sa in step S5 is:

[0020] in, These are the sampling lengths in the x and y directions, respectively. This represents the distance from the coordinate point to the reference plane.

[0021] In one implementation, step S6, analyzing the contribution of microtextile components with different spatial frequencies to surface roughness includes: comparing the variation amplitude of Sa values ​​of adjacent decomposition levels, and determining the spatial frequency range corresponding to the level where the Sa value changes significantly as the frequency range of microtextile components that have a dominant contribution to surface roughness.

[0022] The beneficial effects of this application are as follows: This application utilizes wavelet multi-scale decomposition technology to effectively separate the microtexture morphology of ultrasonically processed surfaces across different spatial frequency dimensions. This method can accurately quantify the contribution of each scale component to surface roughness, revealing the dominant influence of mid-to-high frequency microtexture components on surface morphology characteristics. This analysis provides a clear direction for optimizing ultrasonic processing parameters, making it possible to control the surface microtexture for specific functional requirements, overcoming the limitations of traditional roughness evaluation methods in characterizing functional surfaces. Furthermore, the automatic identification and removal of void regions effectively improves the stability and reliability of the analysis results. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an embodiment of this application; Figure 2 This is a comparison image (a) of the original morphology of the ultrasonically processed microtextured surface under a laser confocal microscope and (b) of the reconstructed two-dimensional matrix image. Figure 3 This is a grayscale histogram of a certain partition in an embodiment of this application; Figure 4 This is a binary image of a certain partition in an embodiment of this application; Figure 5 This is the low-frequency three-dimensional morphology obtained after five decompositions of the original morphology in the embodiments of this application; Figure 6 This is the high-frequency three-dimensional morphology obtained after five decompositions of the original morphology in the embodiments of this application; Figure 7 The results show the low-frequency three-dimensional topography roughness after five decompositions in the embodiments of this application. Figure 8 The results show the high-frequency three-dimensional morphological roughness after five decompositions in the embodiments of this application. Detailed Implementation

[0024] The technical solution of this application will be clearly and completely described below with reference to specific embodiments. However, those skilled in the art will understand that the embodiments described below are only some embodiments of this application, not all embodiments, and are only used to illustrate this application, and should not be regarded as limiting the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] In one specific embodiment, a multi-scale analysis method for the microtexture of ultrasonically processed surfaces is provided, comprising the following steps: S1. Obtain the three-dimensional topographic point cloud data of the ultrasonically processed surface.

[0026] In some embodiments, a laser confocal microscope is used to perform high-precision three-dimensional topographic measurements on the ultrasonically processed surface to obtain its three-dimensional point cloud data. Ultrasonic processing includes, for example, ultrasonic turning, milling, and grinding. In practice, the ultrasonically processed sample is first stably fixed on the stage of the laser confocal microscope. By adjusting the Z-axis height, the sample surface is positioned near the focal plane of the objective lens, thereby ensuring image clarity and measurement accuracy.

[0027] S2. Convert the three-dimensional topographic point cloud data into a two-dimensional matrix, and preprocess the two-dimensional matrix to remove noise and holes.

[0028] In some embodiments, MATLAB software is used to read the three-dimensional topographic point cloud data obtained in step S1.

[0029] In some embodiments, the obtained point cloud data is converted into a two-dimensional matrix, where the rows and columns of the matrix correspond to sampling points on the measured surface in the X and Y directions, respectively, and the value of each element in the matrix corresponds to the Z-axis height information of that point. Therefore, the two-dimensional matrix completely preserves the three-dimensional topographic features of the original surface.

[0030] In some embodiments, the preprocessing process specifically includes region segmentation and masking: the two-dimensional matrix is ​​segmented into regions, and a differentiated binarization threshold is set according to the grayscale histogram of each region to generate a mask matrix, thereby removing outliers and empty regions.

[0031] First, the entire image represented by the two-dimensional matrix is ​​divided into multiple uniformly sized sub-regions using a region segmentation method. For each sub-region, its gray-level histogram is extracted; the gray-level histogram represents the distribution of all pixels within that region at different gray levels.

[0032] Next, based on the distribution characteristics presented by the gray-level histograms of each sub-region, a differentiated binarization threshold is set for each sub-region. Using the specific threshold, each sub-region is binarized to extract the effective data area in each region.

[0033] Finally, the binarized results of all sub-regions are integrated to generate a mask matrix with the same size as the original two-dimensional matrix. This mask matrix is ​​used to identify valid and invalid regions in the original data, such as noise, outliers, and data holes. By applying this mask matrix, outliers introduced during the measurement process and voids caused by missing data can be effectively removed from the original data, resulting in purified, high-quality surface topography data.

[0034] S3. Wavelet transform is used to decompose the preprocessed two-dimensional matrix into multi-scale signals to obtain low-frequency and high-frequency signals at different decomposition levels.

[0035] In some embodiments, the wavelet transform employs a biororthogonal wavelet basis, preferably the bior6.8 wavelet basis. The bior6.8 wavelet basis possesses symmetry and finite support, effectively preserving image edge features while avoiding artifact generation.

[0036] In some embodiments, the specific calculation process of wavelet decomposition is as follows: for each decomposition level, the wavelet function is used... and scaling function With two-dimensional matrix respectively Perform convolution operations to calculate the high-frequency detail coefficients G and low-frequency approximation coefficients D for this layer. The calculation formula is:

[0037] in, a and b The dimension of the two-dimensional matrix. coordinates Topographic data at the location, and The first Wavelet functions and scaling functions of the layers.

[0038] In some embodiments, the maximum number of wavelet decomposition levels Determined by the matrix dimensions of the image: .

[0039] However, to preserve surface features of analytical value, it is necessary to consider an appropriate number of decomposition layers. The appropriate number of decomposition layers is calculated using the following formula:

[0040] in, The center frequency of the wavelet. The sampling interval is... This is the cutoff wavelength for high-frequency information.

[0041] In some embodiments, during wavelet decomposition, the high-frequency signal G at each level contains components in three directions: horizontal direction... Vertical direction and diagonal direction : .

[0042] In some embodiments, the signals at each decomposition level satisfy a complete reconstruction relationship: original signal S It can be derived from the deepest layer (the first) Low-frequency signals (layer) With the first floor to the second floor All high-frequency signals in the layer The sum of all elements is completely reconstructed, that is: .

[0043] S4. Based on the results of the multi-scale decomposition, reconstruct the low-frequency and high-frequency three-dimensional morphology of each decomposition level.

[0044] Due to the downsampling operation during wavelet transform, the size of the coefficient matrix at each decomposition level is reduced. To obtain a 3D topography with the same spatial coordinate frame as the original data, resampling and reconstruction processing is required.

[0045] In some embodiments, when reconstructing the 3D topography, a grid interpolation method is used to resample the coefficients after wavelet decomposition.

[0046] Specifically, the coordinate grid corresponding to the wavelet coefficients of each decomposition level is first extracted. Then, based on the complete spatial coordinate system of the original two-dimensional matrix, the griddata interpolation algorithm is used to interpolate the low-frequency approximation coefficients of each level and the high-frequency detail coefficients in the horizontal, vertical and diagonal directions, respectively, to generate a two-dimensional feature matrix with the same size as the original image, effectively ensuring the spatial consistency between different decomposition levels and between the levels and the original data.

[0047] In some embodiments, the comprehensive high-frequency three-dimensional topography of each level is reconstructed by weighted superposition of the high-frequency sub-signals in the horizontal, vertical, and diagonal directions. Specifically, for the first... The hierarchy is decomposed, and its high-frequency sub-signals in the horizontal direction, vertical direction, and diagonal direction are synthesized, that is, by... The calculation yields a comprehensive high-frequency signal that represents all the detailed features of this level.

[0048] Finally, all the interpolated and reconstructed low-frequency feature matrices and the integrated high-frequency feature matrices are organized according to their corresponding levels and directions. These feature matrices are then fused with the original measured spatial coordinates to output a standard three-column (x, y, z) data format.

[0049] S5. Calculate the three-dimensional surface roughness Sa of the reconstructed three-dimensional morphology at each decomposition level.

[0050] To quantitatively characterize the surface morphology at different scales, three-dimensional surface roughness parameters are calculated for each decomposition level of the reconstructed three-dimensional morphology.

[0051] In some embodiments, the three-dimensional arithmetic mean height Sa is used as the evaluation index. The formula for calculating the three-dimensional surface roughness Sa is:

[0052] in, These represent the sampling lengths of the measured surface in the x and y directions, respectively. To reconstruct the distance from coordinate points on the 3D topography to the reference plane, the reference plane is typically determined by fitting the surface data using the least squares method to eliminate the influence of the macroscopic surface shape on roughness calculation.

[0053] S6. Combining the spatial frequency range corresponding to each decomposition level with the three-dimensional surface roughness Sa, analyze the contribution of micro-texture components with different spatial frequencies to the surface roughness, so as to determine the dominant frequency components affecting the surface roughness.

[0054] In some embodiments, the spatial frequency range corresponding to the decomposition level is determined by the following steps: Determine the sampling frequency based on the sampling interval. f s : ; Corresponding Nyquist frequency f n That is, the highest resolution frequency is:

[0055] Minimum resolution frequency f o Represented as:

[0056] in, L This represents the number of sampling points; Then, by combining the number of wavelet decomposition levels, the specific spatial frequency range corresponding to each decomposition level can be determined.

[0057] In some embodiments, analyzing the contribution of microtextile components at different spatial frequencies to surface roughness includes: comparing the variation amplitude of Sa values ​​between adjacent decomposition levels, and determining the spatial frequency range corresponding to the level where the Sa value changes significantly as the frequency range of the microtextile component that has a dominant contribution to surface roughness. For example, comparing the relative variation amplitude of Sa values ​​between adjacent decomposition levels, when the variation amplitude of Sa between adjacent levels exceeds a set threshold, the spatial frequency range corresponding to that level is determined as the frequency range of the microtextile component that has a dominant contribution to surface roughness.

[0058] Example A multi-scale analysis method for microtexture of ultrasonically processed surfaces, the process of which is as follows: Figure 1 As shown.

[0059] I. High-resolution three-dimensional morphology and point cloud data of the ultrasonically processed surface under test are obtained using a laser confocal microscope. The ultrasonically processed surface under test is fixed on the stage of the confocal microscope, and the Z-axis is adjusted to be near the focal plane of the objective lens. Relevant sampling parameters are set. In this example, the sampling length is 640 μm, the sampling interval is 1.25 μm, the number of sampling points in the X and Y directions is 512, and the number of sampling points in the Z direction is 262,144. During the test, the laser beam scans the sample surface point by point through the scanning galvanometer. The reflected signal is received by the detector and converted into an electrical signal. After image reconstruction, the grayscale image or height information of the processed surface can be obtained, realizing high-precision measurement of the microtexture morphology.

[0060] Second, the point cloud data of the processed morphology was read using MATLAB software and converted into a two-dimensional matrix. Finally, it was compared with the actual morphology. Figure 2 As shown; where the number of rows and columns of the matrix is ​​the same as the number of sampling points of the dot matrix, the example is 512×512.

[0061] Third, the entire image is divided into multiple sub-regions using the region segmentation method, and the gray-level histogram of each region is extracted. For example, the gray-level histogram of the first region is shown below. Figure 3 As shown; based on its grayscale distribution characteristics, differentiated binarization thresholds are set to extract the effective data areas in each region, such as... Figure 4 As shown; finally, a mask matrix is ​​formed for the entire image, and outliers and holes are removed.

[0062] IV. Considering the complex structures of ultrasonically processed surfaces, such as abrupt contour changes, periodic perturbations, tool marks, and processing defects, which exhibit significant locality and non-stationarity, a biorthogonal wavelet basis (bior6.8) is selected for multi-scale decomposition of the microtextured surface. This wavelet basis possesses good symmetry, finite support, and orthogonality, and can simultaneously preserve image edge features and overall contour information, effectively avoiding artifacts and information loss, making it suitable for the local analysis of non-stationary signals.

[0063] V. The number of wavelet decomposition levels determines the number of scales at which information is decomposed. Typically, the maximum number of wavelet decomposition levels is determined by the matrix dimension of the image. .

[0064] The image is transformed into a 512×512 two-dimensional matrix, therefore the theoretical decomposition level is obtained as follows: j max =9. However, after decomposing to 9 layers, the image will be overcompressed, and the image information will consist of only one data point, lacking analytical value. To ensure the data is meaningful, an appropriate number of decomposition layers should be considered:

[0065] Among them, the center frequency =0.7649 (bior 6.8 wavelet); sampling interval =1.25 μm; The cutoff wavelength for high-frequency information is obtained; finally, the cutoff wavelength is calculated. =4.94. To completely extract high-frequency information with wavelengths less than 50 nm, the decomposition level is set to 5. VI. The image is decomposed into five levels using the Bior 6.8 wavelet basis. Due to the downsampling operation in wavelet transform, the image size after each decomposition is half that of the previous level, and the information gradually concentrates in the lower-frequency morphological structures. Each decomposition layer can extract low-frequency approximation components and high-frequency detail components, including horizontal, vertical, and diagonal features. By extracting the coordinate grid of each scale coefficient and resampling the coefficients using the griddata interpolation algorithm based on the spatial coordinate system of the original image, a two-dimensional feature matrix with the same size as the original image is generated. The two-dimensional wavelet decomposition function is as follows: .

[0066] In wavelet decomposition, the relationship between adjacent levels can be represented as follows: .

[0067] VII. The multi-scale wavelet coefficient matrix reconstructed by the above interpolation is organized and output according to level and direction, including the low-frequency components (D1 to D5) of each level and the corresponding high-frequency components in three directions. Furthermore, the comprehensive high-frequency features of each level are extracted by weighted superposition of the high-frequency directional coefficients to fully characterize the detailed features of surface microtexture at different scales. Finally, all feature matrices are fused with the original coordinates according to a unified format, and the output is a three-column data set (…). x , y , z The structure provides fundamental data support for subsequent quantitative analysis of surface features, statistical modeling, and process evaluation. Finally, the low-frequency and high-frequency information after five-level decomposition are respectively as follows: Figure 5 and Figure 6 As shown.

[0068] 8. Determine the frequency range of each decomposition layer; based on the set sampling parameters, the sampling frequency can be obtained. f s Highest resolution frequency f n and minimum resolution frequency f o The relationship is as follows:

[0069]

[0070]

[0071] in, L This represents the number of sampling points. The spatial frequency distribution range of these five layers is shown in Table 1. Table 1. Spatial frequencies corresponding to the five-level wavelet decomposition

[0072] 9. Output the three-dimensional data of low-frequency and high-frequency signals of each decomposition layer and reconstruct the three-dimensional morphology; perform three-dimensional surface roughness calculation and analysis; three-dimensional arithmetic mean. Sa Represented as: .

[0073] 10. Combining the 3D topographic images at different decomposition levels, their corresponding roughness Sa (or other evaluation criteria), and spatial frequency, the results at different scales are analyzed. Sa The causes of the changes were analyzed.

[0074] The three-dimensional surface roughness results for low and high frequency sub-levels are as follows: Figure 7 and Figure 8 As shown. Compared to layer D3, the surface roughness parameters of layer D4 are... Sa The increase reached 15.36%, indicating that after the high-frequency details in G4 were stripped away, the lost information had a significant impact on the roughness characteristics of the original surface morphology, showing a dominant contribution. In the D5 layer, only the overall trend of the surface contour was retained, while the detailed structure was basically filtered out. Sa The variation was only 2.87%, indicating that the roughness response tended to stabilize. This result suggests that the macrostructure formed during ultrasonic milling has a relatively weak impact on roughness, and the surface roughness characteristics are mainly determined by the mid-to-high frequency microtexture. Therefore, to control the roughness of ultrasonically machined surfaces, it is crucial to focus on regulating key process parameters that affect the formation of mid-to-high frequency microtextures, such as ultrasonic amplitude, feed rate, and spindle speed.

[0075] Although the embodiments of this application have been described above in conjunction with the accompanying drawings, this application is not limited to the specific embodiments and application fields described above. The specific embodiments described above are merely illustrative and instructive, not restrictive. Those skilled in the art can make many other forms based on the guidance of this specification and without departing from the scope of protection of the claims of this application, and these are all within the scope of protection of this application.

Claims

1. A multi-scale analysis method for microtexture of ultrasonically processed surfaces, characterized in that, Includes the following steps: S1. Obtain the three-dimensional topographic point cloud data of the ultrasonically processed surface; S2. Convert the three-dimensional topographic point cloud data into a two-dimensional matrix, and preprocess the two-dimensional matrix to remove noise and holes; S3. Wavelet transform is used to decompose the preprocessed two-dimensional matrix into multi-scale signals to obtain low-frequency and high-frequency signals at different decomposition levels. S4. Based on the results of the multi-scale decomposition, reconstruct the low-frequency and high-frequency three-dimensional morphology of each decomposition level; S5. Calculate the three-dimensional surface roughness Sa of the three-dimensional morphology reconstructed at each decomposition level; S6. Combining the spatial frequency range corresponding to each decomposition level with the three-dimensional surface roughness Sa, analyze the contribution of micro-texture components with different spatial frequencies to the surface roughness, so as to determine the dominant frequency components affecting the surface roughness.

2. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 1, characterized in that, The preprocessing includes: segmenting the two-dimensional matrix into regions, and setting differentiated binarization thresholds based on the grayscale histograms of each region to generate a mask matrix, thereby removing outliers and empty regions.

3. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 1, characterized in that, The wavelet transform uses a bioorthogonal wavelet basis.

4. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 3, characterized in that, The biorthogonal wavelet basis is the bior6.8 wavelet basis.

5. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 1, characterized in that, The wavelet transform decomposition process is as follows: for each decomposition level, the wavelet function is used... and scaling function With two-dimensional matrix respectively Perform convolution operations to calculate the high-frequency detail coefficients G and low-frequency approximation coefficients D for this layer: in, a and b The dimension of the two-dimensional matrix. coordinates Topographic data of the location.

6. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 5, characterized in that, During wavelet decomposition, the low-frequency signal at any decomposition level With high frequency signals Satisfy: Original signal It can be obtained from the deepest low-frequency signal and the first to the second layer. The high-frequency signals of the layer are completely reconstructed, that is: 。 7. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 1, characterized in that, The number of wavelet decomposition levels in step S3 Determined by the following formula: in, The center frequency of the wavelet. The sampling interval is... This is the cutoff wavelength for high-frequency information.

8. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 1, characterized in that, The high-frequency signal mentioned in step S3 includes high-frequency sub-signals in the horizontal, vertical, and diagonal directions, represented as follows: in, In the horizontal direction, Vertical direction It is the diagonal direction.

9. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 8, characterized in that, In step S4, the high-frequency sub-signals in the horizontal, vertical and diagonal directions are weighted and superimposed to reconstruct the comprehensive high-frequency three-dimensional morphology of each level.

10. The multi-scale analysis method for microtexture of ultrasonically processed surfaces according to claim 1, characterized in that, The formula for calculating the three-dimensional surface roughness Sa is as follows: in, These are the sampling lengths in the x and y directions, respectively. This represents the distance from the coordinate point to the reference plane.

Images