Method, device and equipment for evaluating surface texture uniformity of curved surface workpiece

Abstract

This application provides a method, apparatus, and device for evaluating the uniformity of surface texture on curved workpieces, relating to the field of surface quality inspection technology. The method includes: acquiring a reference texture image of a standard curved workpiece and a measured texture image of the workpiece to be tested; the reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece's surface. A difference operation is performed between the measured texture image and the reference texture image to extract a pure texture information layer image; multidimensional texture features are extracted from the pure texture information layer image. The measured texture image is compared with the theoretical reflective gradient distribution to obtain a brightness deviation; based on the brightness deviation and multidimensional texture features, the defect classification result of the measured texture image is determined. Based on the multidimensional texture features and the defect classification result, a uniformity evaluation result is output. This method aims to improve the detection accuracy of surface textures.

Description

Technical Field

[0001] This application relates to the field of surface quality testing technology, and more specifically, to a method, apparatus and equipment for evaluating the uniformity of surface texture of curved workpieces. Background Technology

[0002] Currently, with the increasing demands for product appearance quality in the manufacturing industry, the uniformity of surface texture on curved surfaces has become a crucial indicator for evaluating the surface quality of high-end products. In consumer electronics, automotive parts, precision molds, and decorative hardware, processes such as laser etching, chemical etching, and mechanical wire drawing are widely used to create decorative or functional micro / nano textures on curved metal or plastic surfaces. The uniformity of these textures directly affects the visual consistency, tactile experience, and optical performance of products; therefore, reliable testing methods are needed for rapid and accurate evaluation of surface texture uniformity.

[0003] However, in existing technologies, surface texture detection methods are mainly divided into two categories: contact and non-contact. Contact detection methods obtain the contour height information of the surface by scanning along the surface with a probe; among non-contact detection methods, confocal microscopes and white light interferometers are commonly used precision measurement devices.

[0004] However, in existing technologies, the accuracy of surface texture detection is low because the curvature varies at different locations on the surface. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus, and device for evaluating the uniformity of surface texture of curved workpieces, so as to solve the above-mentioned problems existing in the prior art and improve the detection accuracy of surface texture.

[0006] Firstly, a method for evaluating the uniformity of surface texture of a curved workpiece is provided, the method including: Acquire a reference texture image of a standard curved surface workpiece and a measured texture image of the workpiece to be tested; the reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece surface; The measured texture image and the reference texture image are compared by performing a difference operation to extract a pure texture information layer image; multidimensional texture features are then extracted from the pure texture information layer image. The measured texture image is compared with the theoretical reflective gradient distribution to obtain the brightness deviation; based on the brightness deviation and the multidimensional texture features, the defect classification result of the measured texture image is determined. Based on the multidimensional texture features and the defect classification results, output the uniformity evaluation results.

[0007] Secondly, a device for evaluating the uniformity of surface texture of a curved workpiece is provided, the device comprising: The acquisition module is used to acquire a reference texture image of a standard curved surface workpiece and a measured texture image of the curved surface workpiece to be tested; the reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece surface; The extraction module is used to perform a difference operation between the measured texture image and the reference texture image to extract a pure texture information layer image; and to extract multi-dimensional texture features from the pure texture information layer image. The classification module is used to compare the measured texture image with the theoretical reflective gradient distribution to obtain the brightness deviation; and to determine the defect classification result of the measured texture image based on the brightness deviation and the multidimensional texture features. The output module is used to output a uniformity evaluation result based on the multidimensional texture features and the defect classification result.

[0008] Thirdly, an electronic device is provided, which includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When a processor executes a program stored in memory, it implements any of the steps described in the first aspect above.

[0009] Fourthly, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when executed by a processor, the computer program implements the steps of any of the methods described in the first aspect above.

[0010] This application provides a method, apparatus, and device for evaluating the uniformity of surface texture of curved workpieces. The method acquires a reference texture image of a standard curved workpiece and a measured texture image of the workpiece to be tested. The reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece's surface. A difference operation is performed between the measured texture image and the reference texture image to extract a pure texture information layer image. Multidimensional texture features are extracted from the pure texture information layer image. The measured texture image is compared with the theoretical reflective gradient distribution to obtain a brightness deviation. Based on the brightness deviation and multidimensional texture features, the defect classification result of the measured texture image is determined. Based on the multidimensional texture features and the defect classification result, a uniformity evaluation result is output. In this scheme, the theoretical reflective gradient distribution of a smooth, standard curved workpiece is used as a reference. The curvature interference is effectively eliminated through difference operations. Texture uniformity evaluation can be achieved without high-precision three-dimensional topography measurement, significantly reducing detection costs, increasing detection speed, eliminating curvature interference, and achieving rapid, non-destructive, and quantitative evaluation of surface texture uniformity, thereby improving the detection accuracy of surface textures. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating a method for evaluating the uniformity of surface texture of a curved workpiece, provided in an embodiment of this application; Figure 2 A flowchart illustrating another method for evaluating the uniformity of surface texture of a curved workpiece provided in this application embodiment; Figure 3 This application provides an embodiment of a device for evaluating the uniformity of surface texture of a curved workpiece. Figure 4 A schematic diagram of the structure of another surface texture uniformity evaluation device for curved workpieces provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by those skilled in the art. The words "first," "second," and similar terms used in this application do not indicate any order, quantity, or importance, but are only used to distinguish different components. The words "comprising" or "including," etc., mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but do not exclude other elements or objects. The words "connected," "coupled," or "connected," etc., are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up," "down," "left," "right," etc., are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0014] Currently, with the increasing demands for product appearance quality in the manufacturing industry, the uniformity of surface texture on curved surfaces has become a crucial indicator for evaluating the surface quality of high-end products. In consumer electronics, automotive parts, precision molds, and decorative hardware, processes such as laser etching, chemical etching, and mechanical wire drawing are widely used to create decorative or functional micro / nano textures on curved metal or plastic surfaces. The uniformity of these textures directly affects the visual consistency, tactile experience, and optical performance of products; therefore, reliable testing methods are needed for rapid and accurate evaluation of surface texture uniformity.

[0015] In one example, surface texture detection methods are mainly divided into two categories: contact and non-contact. Contact detection methods obtain the contour height information of the surface by scanning along the surface with a probe. Although this method has high measurement accuracy, it has the following drawbacks: slow detection speed, only applicable to small local areas, and unable to evaluate the overall uniformity of large-format surfaces; the probe may scratch soft surfaces, thus it is not a non-destructive testing method; for complex surfaces, a large number of scanning paths are required, resulting in extremely low efficiency. In non-contact detection methods, confocal microscopy and white light interferometry are commonly used precision measurement devices. Confocal microscopy obtains the three-dimensional morphology of the surface by scanning point by point, with high axial resolution, but a small field of view. For large-format surfaces, multiple stitching is required, and the measurement time can reach several hours. White light interferometry uses the interference principle to obtain surface height information with nanometer-level accuracy, but it is also limited by a very small measurement field of view and is sensitive to changes in surface curvature, easily generating measurement errors in edge areas. However, in existing technologies, the different curvatures at different locations on the surface lead to low accuracy in surface texture detection.

[0016] The method for evaluating the surface texture uniformity of curved workpieces provided in this application can be applied to electronic devices, terminal devices, devices or apparatuses for evaluating the surface texture uniformity of curved workpieces, or other devices or apparatuses capable of executing this embodiment, without limitation. This embodiment describes the execution subject as an electronic device. This application is applicable to the surface texture uniformity detection of curved workpieces in industries such as consumer electronics, automotive parts, and precision molds.

[0017] The terminal can be a user equipment (UE) such as a mobile phone, smartphone, laptop computer, digital broadcast receiver, personal digital assistant (PDA), or tablet computer (PAD), handheld device, in-vehicle device, wearable device, computing device, or other processing device connected to a wireless modem, mobile station (MS), or mobile terminal. This terminal has the ability to communicate with one or more core networks via a radio access network (RAN).

[0018] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application. Furthermore, the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0019] Figure 1 This is a flowchart illustrating a method for evaluating the uniformity of surface texture in a curved workpiece, as provided in an embodiment of this application. Figure 1 As shown, the method may include: Step S101: Obtain the reference texture image of the standard curved surface workpiece and the measured texture image of the curved surface workpiece to be tested; the reference texture image includes the theoretical reflective gradient distribution generated by the curvature of the surface of the workpiece.

[0020] For example, a standard curved surface workpiece without surface texture defects is placed at the inspection station, and images are acquired under a preset light source angle and camera viewpoint to obtain a reference texture image that reflects the theoretical reflective gradient distribution inherent in the workpiece due to the curvature variation of its surface. The theoretical reflective gradient distribution includes: higher brightness in the area directly facing the light source and lower brightness in the area away from the light source, forming a gradient related to curvature. Then, under the same light source angle and camera viewing angle, measured texture images of the workpiece surface under test are acquired.

[0021] Optionally, geometric correction is performed on the acquired measured texture image, that is, pixel coordinates are established using the Zhang calibration method. The mapping to the UV parameter domain (u,v) is achieved by establishing a mapping relationship between pixel coordinates and the cylindrical parameter domain through ray tracing. Bicubic interpolation resampling is then used to obtain a measured texture image uniformly distributed in the (u,v) parameter domain. Specifically, the geometric correction process includes: First calibration method: Using a camera calibration algorithm, multiple calibration board images in different poses are used to solve for the intrinsic and extrinsic parameter matrices of the industrial camera and the lens distortion coefficients. Based on the distortion coefficients, the acquired original surface texture image is subjected to distortion correction processing to obtain a corrected image after eliminating radial and tangential lens distortion. Second correction step, mapping from pixel coordinates to the surface parameter domain: A three-dimensional geometric model of the measured surface is acquired. The flattening parameter domain of the measured surface is established with the rotation axis of the turntable as the axial direction and the angular direction around the rotation axis as the viewing direction. Using a ray tracing algorithm, starting from the camera's optical center position, the first intersection point between the camera's optical center and the three-dimensional geometric model is traced along the viewing direction of each pixel, establishing a one-to-one mapping relationship between pixel coordinates and flattening parameter domain coordinates. The third correction step, uniform resampling: Based on the mapping relationship between pixel coordinates and flattening parameter domain coordinates, uniform grid sampling points are set in the flattening parameter domain. Through the interpolation resampling algorithm, the gray values ​​of multiple neighboring pixels around each sampling point are weighted and interpolated to calculate the gray value of each sampling point, generating an unfolded texture image with uniform spatial resolution in the flattening parameter domain, which serves as the input data for subsequent steps.

[0022] Step S102: Perform a difference operation between the measured texture image and the reference texture image to extract the pure texture information layer image; extract the multidimensional texture features from the pure texture information layer image.

[0023] For example, each pixel in both the measured texture image and the reference texture image has a grayscale value. The grayscale values ​​of each pixel in the measured texture image are subtracted from the corresponding pixel values ​​in the reference texture image to eliminate interference from reflective gradients caused by surface curvature, extracting a pure texture information layer containing only surface texture information. Then, multidimensional texture features are extracted from the pure texture information layer image, including grayscale uniformity features, orientation consistency features, and periodicity features.

[0024] Among these features, grayscale uniformity reflects the overall consistency of the surface texture's depth. If the texture etching is uniform, the grayscale values ​​across the entire surface should be concentrated around a certain average value; otherwise, the grayscale values ​​will fluctuate significantly. This metric is extracted by calculating the average grayscale value and standard deviation for all pixels in the clean texture information layer image, and using the ratio of the standard deviation to the average value to measure the degree of fluctuation. The smaller the fluctuation relative to the average level, the better the grayscale uniformity.

[0025] The orientation consistency feature reflects the regularity of the orientation of brushed textures. If the texture orientation is consistent (e.g., parallel brushing), the dominant texture orientations of each pixel should be concentrated around the same angle; if the orientations are disordered, the angles will be scattered. This indicator is extracted by using a Gabor filter bank to extract the dominant texture orientation of each pixel, and then calculating the dispersion of all pixel orientations (i.e., orientation variance DOV). The orientation variance DOV is the orientation consistency feature. The smaller the orientation variance, the better the orientation consistency.

[0026] Periodicity is used to reflect the periodic repetition stability of textures like mesh / frosted surfaces. If the texture period is consistent, the spacing between adjacent peaks should be uniform, and the amplitude of each period should be stable. This indicator is extracted by: extracting a one-dimensional grayscale profile sequence along the dominant texture direction, calculating the autocorrelation function, detecting peak regularity, and comprehensively scoring it from three dimensions: spacing uniformity, amplitude stability, and peak significance. The higher the score, the better the periodicity.

[0027] Therefore, by using the inherent theoretical reflective gradient distribution of an ideal smooth surface as an evaluation benchmark, and eliminating curvature interference through the benchmark texture image difference method, quantitative evaluation of texture uniformity can be achieved. This technical approach breaks through the traditional mindset of "eliminating reflective interference" and transforms reflective gradient from "noise" into a usable "signal".

[0028] Step S103: Compare the measured texture image with the theoretical reflective gradient distribution to obtain the brightness deviation; determine the defect classification result of the measured texture image based on the brightness deviation and multidimensional texture features.

[0029] For example, the brightness value of the measured texture image is subtracted from the theoretical value at the corresponding position in the theoretical reflective gradient distribution, pixel by pixel, to obtain the brightness deviation of each pixel. Then, based on the brightness deviation and multidimensional texture features, the defect classification result of the measured texture image is determined comprehensively.

[0030] Step S104: Output uniformity evaluation results based on multidimensional texture features and defect classification results.

[0031] For example, the grayscale uniformity feature, orientation consistency feature, and periodicity feature are normalized to obtain normalized grayscale uniformity feature, orientation consistency feature, and periodicity feature. Then, the normalized grayscale uniformity feature, orientation consistency feature, and periodicity feature are weighted and summed according to preset weights to obtain a comprehensive score. The comprehensive score is then compared with multiple preset grading thresholds to obtain the uniformity evaluation result of the workpiece surface under test.

[0032] The method provided in this application acquires a reference texture image of a standard curved surface workpiece and a measured texture image of the workpiece to be tested. The reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece's surface. A difference operation is performed between the measured texture image and the reference texture image to extract a pure texture information layer image. Multidimensional texture features are extracted from the pure texture information layer image. The measured texture image is compared with the theoretical reflective gradient distribution to obtain a brightness deviation. Based on the brightness deviation and multidimensional texture features, the defect classification result of the measured texture image is determined. Based on the multidimensional texture features and the defect classification result, a uniformity evaluation result is output. In this scheme, the theoretical reflective gradient distribution of a smooth standard curved surface workpiece is used as a reference. Curvature interference is effectively eliminated through difference operations. Texture uniformity evaluation can be achieved without high-precision three-dimensional topography measurement, significantly reducing detection costs, increasing detection speed, eliminating curvature interference, and realizing rapid, non-destructive, and quantitative evaluation of surface texture uniformity, thereby improving the detection accuracy of surface texture.

[0033] Figure 2 A flowchart illustrating another method for evaluating the uniformity of surface texture in curved workpieces provided in this application is shown below. Figure 2 As shown, in this embodiment... Figure 1 Based on the embodiments, the method is described in detail below, and the method includes: Step S201: Obtain the reference texture image of the standard curved surface workpiece and the measured texture image of the curved surface workpiece to be tested; the reference texture image includes the theoretical reflective gradient distribution generated by the curvature of the surface of the workpiece.

[0034] For example, this step is described in step S101, and will not be repeated here.

[0035] Step S202: Perform a difference operation between the measured texture image and the reference texture image to extract the pure texture information layer image; In one example, S202 includes: obtaining the gray level of a specified reference region in a reference texture image, obtaining the gray level of the same spatial location in a measured texture image corresponding to the specified reference region, and calculating the proportional relationship between the two as a reflectance calibration coefficient; generating a reference texture image scaled by the reflectance calibration coefficient; performing a pixel-by-pixel difference operation on the measured texture image and the reference texture image scaled by the reflectance calibration coefficient to obtain a difference image; the difference image includes texture information; dividing the difference image into multiple local sub-blocks, performing histogram stretching on each local sub-block, then stitching the local sub-blocks together and smoothing the transition boundaries between the local sub-blocks to obtain a contrast-enhanced difference image; and sequentially performing coarse-grained filtering and fine-grained filtering on the contrast-enhanced difference image to obtain a clean texture information layer image.

[0036] For example, the measured texture image Compared with the baseline texture image Perform the difference operation to obtain the difference image. ,in The reflectance calibration coefficients are used, and adaptive histogram equalization and cascaded filtering are performed sequentially to obtain a clean texture information layer image.

[0037] Specifically, the grayscale level of a specified reference region in the benchmark texture image is obtained, and the grayscale level of the corresponding spatial location in the measured texture image is obtained. The grayscale level of the measured texture image is divided by the grayscale level of the benchmark texture image to obtain the proportional relationship, which is used as the reflectance calibration coefficient. The specified reference region refers to one or more representative local regions in the benchmark texture image, used to calculate the grayscale ratio between the benchmark and measured texture images. This region is not a fixed preset location but is selected based on the actual situation. For example, the specified reference region can be the area directly facing the light source, a multi-region average, or the area with the flattest texture. The area directly facing the light source refers to the central region with the highest image brightness, which has the highest signal-to-noise ratio in the benchmark texture image. Multi-region averaging refers to uniformly selecting multiple regions in the image and taking the average grayscale value to improve the robustness of the calibration coefficient. Therefore, the average grayscale value of all pixels within the specified reference region of the benchmark texture image and the average grayscale value of all pixels within the same region in the measured texture image can be obtained.

[0038] A baseline texture image scaled by reflectivity calibration coefficients is generated. A pixel-by-pixel difference operation is performed between the measured texture image and the scaled baseline texture image to obtain a difference image; the difference image includes texture information. The difference image is divided into multiple local sub-blocks, and histogram stretching is performed on each local sub-block. The local sub-blocks are then stitched together, and the transition boundaries between them are smoothed to obtain a contrast-enhanced difference image. Finally, coarse-grained filtering and fine-grained filtering are sequentially performed on the contrast-enhanced difference image to obtain a clean texture information layer image.

[0039] Step S203: Extract multidimensional texture features from the pure texture information layer image.

[0040] In one example, the multidimensional texture features include grayscale uniformity features, orientation consistency features, and periodicity features; the orientation consistency features are used to indicate the overall dominant texture direction in the clean texture information layer image; extracting the periodicity features in the clean texture information layer image includes: extracting the grayscale value of each pixel in the difference image along the overall dominant texture direction indicated by the orientation consistency features to generate a one-dimensional grayscale profile sequence; determining the normalized autocorrelation function of the one-dimensional grayscale profile sequence; determining local maxima greater than a preset significance threshold within the positive displacement range of the autocorrelation function, and taking the local maxima as effective peaks; recording the displacement position and amplitude of each effective peak; determining the interval regularity coefficient, amplitude stability coefficient, and peak significance coefficient based on the autocorrelation function, the displacement position, and amplitude of each effective peak; weighting and summing the interval regularity coefficient, the amplitude stability coefficient, and the peak significance coefficient according to preset weights to obtain a periodic score; the periodic score is a periodic feature.

[0041] For example, the multidimensional texture features include grayscale uniformity features, orientation consistency features, and periodicity features. Among them, the orientation consistency feature is used to indicate the overall dominant texture direction in the clean texture information layer image.

[0042] Among them, gray-level uniformity is defined as ,in The standard deviation of grayscale The UI value represents the grayscale mean. The closer the UI value is to 1, the more uniform the grayscale distribution.

[0043] Directional consistency feature is based on The filter bank extracts the dominant texture direction at each pixel location and calculates the direction variance for the dominant texture direction at each pixel location. , This represents the dominant texture direction at each pixel location, that is, the angle of the dominant direction of texture extension at each pixel location. A smaller value indicates a more consistent texture direction. The specific calculation method is as follows: A Gabor filter bank containing multiple direction parameters is constructed. Each filter in the Gabor filter bank corresponds to a different direction angle; for example, the different direction angles include 0°, 30°, 60°, 90°, 120°, and 150°, used to detect texture lines with different directions in the image. The clean texture information layer image is convolved with the Gabor filter in each direction to obtain the response energy value of each pixel in each direction. For each pixel location, the response energy values ​​of the pixel location in different directions are compared, and the direction with the largest response energy value is taken as the dominant texture direction of that pixel. Record the angle value of the direction; finally, count the angle value of the texture dominant direction of each pixel, and calculate the direction variance based on the angle value of the texture dominant direction of each pixel. The direction variance is the value of the direction consistency feature; the direction consistency feature is used to indicate the overall texture dominant direction.

[0044] For periodic features, this is achieved by calculating the autocorrelation function. The peak regularity is used to evaluate the periodicity consistency of the texture; textures with high periodicity consistency... The peak values ​​are regularly distributed and evenly spaced. The specific calculation method for the periodicity is as follows: 1. One-dimensional autocorrelation function extraction: along the dominant direction of the overall texture, in the difference image Extract the grayscale value of each pixel to generate a one-dimensional grayscale profile sequence, and calculate the normalized autocorrelation function of the one-dimensional grayscale profile sequence. , τ represents the displacement, i.e., the pixel distance translated along the dominant direction of the overall texture; τ = 0 indicates that the grayscale profile sequence is most similar to itself (without translation); τ = 1 indicates that the grayscale profile sequence is compared with the sequence after being translated forward by 1 pixel.

[0045] 2. Significant Peak Detection: Within the positive shift interval where τ>0 of the autocorrelation function, a preset significance threshold is used to detect peak values. Local maxima are selected as valid peak values ​​to obtain M valid peak values. The displacement positions of the valid peak values ​​are recorded. and amplitude Where i = 1, 2, ..., M; if the number of effective peak values If the texture does not exhibit recognizable periodicity, then a periodicity score will be applied. .

[0046] 3. Interval regularity coefficient: Calculates the interval sequence between adjacent peaks. ,beg mean and standard deviation Define the interval regularity coefficient :

[0047] Should The range of values ​​is The closer to 1, the more uniform the peak interval and the higher the consistency of the texture period.

[0048] 4. Amplitude stability coefficient: Calculated by the mean of the peak amplitude. and standard deviation Define the amplitude stability coefficient :

[0049] Should The range of values ​​is The closer the value is to 1, the better the repeatability of the texture in each cycle.

[0050] 5. Peak Significance Coefficient: Defined as the peak significance coefficient to reflect the strength of a periodic signal relative to the DC component. :

[0051] Should The range of values ​​is (0,1], and the larger the amplitude, the more significant the periodicity.

[0052] 6. Comprehensive calculation of periodic scores: The periodicity score is obtained by weighting and summing the interval regularity coefficient, amplitude stability coefficient, and peak significance coefficient according to preset weights; the periodicity score represents the periodic characteristic. The calculation formula is as follows:

[0053] Wherein, the weight coefficients satisfy The default is acceptable. , , . The value ranges from [0,1], with larger values ​​indicating better texture periodicity and higher periodicity consistency. It should be noted that the data used in this application are merely examples and are not intended to limit the scope of the application.

[0054] Step S204: Compare the measured texture image with the theoretical reflective gradient distribution to obtain the brightness deviation; based on the brightness deviation and multidimensional texture features, determine the defect classification result of the measured texture image.

[0055] In one example, S204 includes: subtracting the brightness value of the measured texture image from the theoretical value at the corresponding position in the theoretical reflective gradient distribution pixel by pixel to obtain the brightness deviation at each pixel position; marking the effective non-uniform regions based on the brightness deviation and taking the effective non-uniform regions as the target local regions; marking the regions in the measured texture image other than the target local regions as normal texture regions; and determining the defect classification result of the target local regions based on the brightness deviation and multi-dimensional texture features.

[0056] In one example, marking effective non-uniform regions based on the brightness deviation amount and using the effective non-uniform regions as target local regions includes: generating a deviation distribution map based on the brightness deviation amount at each pixel position; determining the dispersion of all brightness deviation amounts in the deviation distribution map, using a preset multiple of dispersion as a screening threshold, and marking pixels with absolute deviation values ​​greater than the screening threshold as candidate abnormal pixels; performing spatial neighborhood analysis on the candidate abnormal pixels, merging spatially adjacent candidate abnormal pixels into the same pixel set; marking the pixel set as an effective non-uniform region, and using the effective non-uniform region as the target local region.

[0057] In one example, the multidimensional texture features include orientation consistency features and periodicity features; the periodicity features are the overall periodicity of the clean texture information layer image; the orientation consistency features are obtained based on the response energy values ​​of each pixel; based on the brightness deviation and multidimensional texture features, the defect classification result of the target local region is determined, including: for each target local region, determining the average deviation of the brightness deviation at all pixel positions; determining the response energy value of each pixel in the normal texture region; calculating the average response energy value of each pixel in the normal texture region as the response energy level of the normal texture region; for each target local region, if the average deviation of the target local region is positive, the defect classification result of the target local region is determined to be a depth-deficient defect; if the average deviation of the target local region is negative, the defect classification result of the target local region is determined to be a depth-over-depth defect; determining the average response energy value of the target local region; if the average response energy value of the target local region is less than the response energy level of the normal texture region, the defect classification result of the target local region is determined to be a texture-deficient defect; determining the regional periodicity score of the target local region; if the regional periodicity score is less than the periodicity feature, it is determined to be a periodic anomaly defect.

[0058] For example, the brightness value of the measured texture image is subtracted from the theoretical value at the corresponding position in the theoretical reflective gradient distribution, pixel by pixel, to obtain the brightness deviation at each pixel position. A deviation distribution map is generated based on the brightness deviation at each pixel position. The dispersion of all brightness deviations in the deviation distribution map is calculated; for example, the dispersion is the standard deviation. A preset multiple of the dispersion is used as a screening threshold (e.g., 3 times the standard deviation). Pixels with an absolute value of deviation greater than this screening threshold are marked as candidate abnormal pixels. Spatial neighborhood analysis is performed on the candidate abnormal pixels, merging spatially adjacent candidate abnormal pixels into the same pixel set. The region corresponding to the same pixel set is the effective non-uniform region, and the effective non-uniform region is used as the target local region in the measured texture image. Regions in the measured texture image other than the target local region are marked as normal texture regions. Based on the brightness deviation and multi-dimensional texture features, the defect classification result of the target local region is determined.

[0059] Based on the changes in brightness deviation and multidimensional texture features, the defect classification results of the target local region are determined, including: insufficient depth type, characterized by an overall bright texture region (texture depth less than the design value); excessive depth type, characterized by an overall dark texture region (texture depth greater than the design value); texture missing type, characterized by loss of texture directionality in a local region (significantly reduced response energy value); and periodic anomaly type, characterized by abnormal periodic features in a local region (irregular autocorrelation function peaks). Specifically, for each target local region, the average deviation of brightness deviation at all pixel positions is calculated; the response energy value of each pixel within the normal texture region is calculated, where the response energy value is the maximum response energy value of each pixel. The average of the maximum response energy values ​​of all pixels within the normal texture region is then used as the response energy level of the normal texture region.

[0060] Then, for each target local region, if the average deviation of the target local region is positive, the defect classification result of the target local region is determined to be a depth-deficient defect; if the average deviation of the target local region is negative, the defect classification result of the target local region is determined to be a depth-over-depth defect. The average response energy value of the response energy values ​​of all pixels in the target local region is calculated. If the average response energy value of the target local region is less than the response energy level of the normal texture region, the defect classification result of the target local region is determined to be a texture-deficient defect. The regional periodicity score of the target local region is calculated. If the regional periodicity score is less than the periodicity feature, it is determined to be a periodic anomaly defect. The calculation range of the regional periodicity score is the local image within the target local region of the measured texture image, while the calculation range of the periodicity feature is the entire image of the clean texture information layer. Although their calculation ranges are different, the calculation methods are the same. The calculation process of the regional periodicity score is the same as that of the periodicity feature, and will not be repeated here.

[0061] Step S205: Output uniformity evaluation results based on multidimensional texture features and defect classification results.

[0062] In one example, S205 includes: normalizing the grayscale uniformity feature, orientation consistency feature, and periodicity feature to obtain normalized grayscale uniformity feature, orientation consistency feature, and periodicity feature; weighting and summing the normalized grayscale uniformity feature, orientation consistency feature, and periodicity feature according to preset weights to obtain a comprehensive score; and comparing the comprehensive score with multiple preset grading thresholds to determine the uniformity evaluation result of the workpiece to be tested.

[0063] For example, based on multidimensional texture features and defect classification results, uniformity quantification scoring is performed, and the uniformity evaluation result is output. Specifically, grayscale uniformity features, orientation consistency features, and periodicity features are normalized to obtain normalized grayscale uniformity features, orientation consistency features, and periodicity features. Then, according to preset weights, the normalized grayscale uniformity features, orientation consistency features, and periodicity features are weighted and summed to calculate the comprehensive score. , in , , For the preset weights and The overall score is compared with multiple preset grading thresholds to achieve grading: Excellent (excellent texture uniformity, no visible defects). Good (good texture uniformity, with very slight unevenness). The result is acceptable (the texture is basically uniform, with some local unevenness but it does not affect the overall appearance). This is considered substandard (the texture is obviously uneven and needs to be reworked or scrapped).

[0064] Among them, UI, Both `Periodicity_score` and `Periodicity_score` are dimensionless indices within the range [0,1], ensuring dimensional consistency in the weighted fusion. The DOV value is normalized to obtain... , This represents the theoretical maximum directional variance.

[0065] Therefore, the comprehensive evaluation method based on multidimensional texture features also considers gray-level uniformity features. The system uses weighted fusion of directional consistency and periodicity characteristics to obtain a comprehensive score, avoiding the one-sidedness of single-dimensional evaluation and enabling the identification of comprehensive defects such as uneven depth and inconsistent direction that are difficult to detect by traditional methods. Intelligent defect classification categorizes texture unevenness defects into four types: insufficient depth, excessive depth, missing texture, and periodic anomaly, facilitating guidance for process optimization.

[0066] The method provided in this application acquires a reference texture image of a standard curved surface workpiece and a measured texture image of the workpiece to be tested. The reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece's surface. A difference operation is performed between the measured texture image and the reference texture image to extract a pure texture information layer image. Multidimensional texture features are extracted from the pure texture information layer image. The measured texture image is compared with the theoretical reflective gradient distribution to obtain a brightness deviation. Based on the brightness deviation and multidimensional texture features, the defect classification result of the measured texture image is determined. Based on the multidimensional texture features and the defect classification result, a uniformity evaluation result is output. In this scheme, the theoretical reflective gradient distribution of a smooth standard curved surface workpiece is used as a reference. Curvature interference is effectively eliminated through difference operations. Texture uniformity evaluation can be achieved without high-precision three-dimensional topography measurement, significantly reducing detection costs, increasing detection speed, eliminating curvature interference, and realizing rapid, non-destructive, and quantitative evaluation of surface texture uniformity. Furthermore, the non-contact inspection of this application will not cause scratches or damage to the surface being tested, making it suitable for quality control of precision parts; the inspection speed is fast, the single image area is large, and the processing time is less than 1 second, which can meet the cycle time requirements of online or near-line inspection on the production line.

[0067] In one embodiment, this application also provides a device for evaluating the uniformity of surface texture of a curved workpiece, such as... Figure 3 As shown, it includes: a base system (1) for providing necessary support and positioning for the part inspection device, including a movable bracket (11) that allows the light source and image acquisition system to move as a whole; a turntable (2) for fixing the workpiece to be measured and for being rotatable and / or tiltable; a light source system (3) including a ring LED array light source and a light source controller (which may also include a light source angle adjustment mechanism) for providing stable and angle-adjustable illumination; an image acquisition system (4) including an industrial camera, lens and polarizing filter (which may also include a camera angle adjustment mechanism) for acquiring texture images; and the part to be measured (21).

[0068] Among them, such as Figure 3 As shown, the light source system (3) and the image acquisition system (4) are connected to the movable bracket (11) via a hinge assembly. The light source system (3) and the image acquisition system (4) can directly share a single hinge assembly connected to the movable bracket (11). In this case, the light source system (3) does not include a light source angle adjustment mechanism, and the image acquisition system (4) does not include a camera angle adjustment mechanism. Alternatively, the light source system (3) includes a light source angle adjustment mechanism, and the image acquisition system (4) includes a camera angle adjustment mechanism. In this case, the light source angle adjustment mechanism of the light source system (3) and the camera angle adjustment mechanism of the image acquisition system (4) are connected to the movable bracket (11) via a hinge assembly.

[0069] The geometric constraints include: the rotation axis of the hinge assembly passes through the preset observation area of ​​the measured part (21), and the optical axis of the light source system (3) intersects the optical axis of the image acquisition system (4) at the same point on the rotation axis. When adjusting the rotation angle of the hinge assembly, the light source system (3) and the image acquisition system (4) swing around the rotation axis as a whole, so that the illumination center of the light source and the imaging center of the camera always coincide.

[0070] The device also includes an image processing and evaluation unit for performing geometric correction, reflection gradient removal, texture feature extraction, and uniformity quantization scoring; and a multi-view acquisition control module for controlling the linkage between the stage, light source, and camera.

[0071] The light source angle adjustment mechanism and the camera angle adjustment mechanism allow for separate adjustment of the angles of the light source and the camera relative to the measured surface; a polarizing filter is installed at the front of the lens to suppress glare caused by specular reflection.

[0072] Therefore, the light source angle adjustment mechanism and the camera angle adjustment mechanism maintain the camera imaging center and the light source illumination center at all times through geometric constraints, while allowing the light source and camera to adjust their relative angles while maintaining this coincidence constraint, and providing real-time angle data feedback. Through the optimized configuration of the light source-camera angle, differentiated acquisition strategies are adopted for different surface types (cylindrical surfaces, spherical surfaces, freeform surfaces), significantly improving the signal-to-noise ratio and contrast of texture imaging.

[0073] Corresponding to the above method, embodiments of this application also provide another device for evaluating the uniformity of surface texture of curved workpieces, such as... Figure 4 As shown, the device includes: The acquisition module 41 is used to acquire the reference texture image of the standard curved surface workpiece and the measured texture image of the curved surface workpiece to be tested; the reference texture image includes the theoretical reflective gradient distribution generated by the curvature of the surface of the workpiece; Extraction module 42 is used to perform a difference operation between the measured texture image and the reference texture image to extract the pure texture information layer image; and to extract the multi-dimensional texture features in the pure texture information layer image. The classification module 43 is used to compare the measured texture image with the theoretical reflective gradient distribution to obtain the brightness deviation; based on the brightness deviation and multidimensional texture features, the defect classification result of the measured texture image is determined. Output module 44 is used to output uniformity evaluation results based on multidimensional texture features and defect classification results.

[0074] The functions of each functional unit of the surface texture uniformity evaluation device for curved workpieces provided in the above embodiments of this application can be realized through the above method steps. Therefore, the specific working process and beneficial effects of each unit in the surface texture uniformity evaluation device for curved workpieces provided in the embodiments of this application will not be repeated here.

[0075] This application also provides an electronic device, such as... Figure 5 As shown, it includes a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540.

[0076] Memory 530 is used to store computer programs; The processor 510 performs the above steps when executing the program stored in the memory 530.

[0077] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0078] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0079] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0080] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0081] The implementation methods and beneficial effects of the various components of the electronic device in the above embodiments for solving the problem can be found in [reference needed]. Figure 1 The steps in the illustrated embodiments are used to implement the electronic device. Therefore, the specific working process and beneficial effects of the electronic device provided in this application will not be repeated here.

[0082] In another embodiment provided in this application, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the method for evaluating the uniformity of surface texture of a curved workpiece as described in any of the above embodiments.

[0083] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the method for evaluating the uniformity of surface texture of curved workpieces as described in any of the above embodiments.

[0084] Those skilled in the art will understand that the embodiments in this application can be provided as methods, systems, or computer program products. Therefore, the embodiments in this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments in this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] This application describes embodiments of methods, apparatus (systems), and computer program products according to embodiments of this application with reference to flowchart illustrations and / or block diagrams. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] Although preferred embodiments have been described in this application, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of this application.

[0089] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims in this application and their equivalents, then this application also intends to include these modifications and variations.

Claims

1. A method for evaluating the uniformity of surface texture on a curved workpiece, characterized in that, include: Acquire the reference texture image of the standard curved surface workpiece and the measured texture image of the curved surface workpiece under test; The reference texture image includes a theoretical reflective gradient distribution generated by the surface curvature of the workpiece; The measured texture image and the reference texture image are compared by performing a difference operation to extract a pure texture information layer image; multidimensional texture features are then extracted from the pure texture information layer image. The measured texture image is compared with the theoretical reflective gradient distribution to obtain the brightness deviation; based on the brightness deviation and the multidimensional texture features, the defect classification result of the measured texture image is determined. Based on the multidimensional texture features and the defect classification results, output the uniformity evaluation results.

2. The method as described in claim 1, characterized in that, The measured texture image is compared with the theoretical reflective gradient distribution to obtain the brightness deviation; based on the brightness deviation and the multidimensional texture features, the defect classification result of the measured texture image is determined, including: The brightness value of the measured texture image is subtracted from the theoretical value at the corresponding position in the theoretical reflective gradient distribution by pixel to obtain the brightness deviation at each pixel position. Based on the brightness deviation, mark the effective non-uniform regions and use the effective non-uniform regions as target local regions; mark the regions in the measured texture image other than the target local regions as normal texture regions. Based on the brightness deviation and the multidimensional texture features, the defect classification result of the target local region is determined.

3. The method as described in claim 2, characterized in that, Based on the brightness deviation, effective non-uniform regions are marked, and these effective non-uniform regions are used as target local regions, including: A deviation distribution map is generated based on the brightness deviation at each pixel location; Determine the degree of dispersion of all brightness deviations in the deviation distribution map, use the degree of dispersion of a preset multiple as a screening threshold, and mark pixels whose absolute value of deviation is greater than the screening threshold as candidate abnormal pixels; Spatial neighborhood analysis is performed on the candidate abnormal pixels, and spatially adjacent candidate abnormal pixels are merged into the same pixel set; the pixel set is marked as an effective non-uniform region, and the effective non-uniform region is used as the target local region.

4. The method as described in claim 2, characterized in that, The multidimensional texture features include directional consistency features and periodicity features; the periodicity features are the overall periodicity features of the pure texture information layer image; the directional consistency features are obtained based on the response energy values ​​of each pixel. Based on the brightness deviation and the multidimensional texture features, the defect classification result of the target local region is determined, including: For each target local region, determine the average deviation of the brightness deviation at all pixel positions; Determine the response energy value of each pixel within the normal texture region; calculate the average response energy value of each pixel within the normal texture region as the response energy level of the normal texture region; For each target local area, if the average deviation of the target local area is positive, then the defect classification result of the target local area is determined to be a depth-insufficient defect. If the average deviation of the target local area is negative, then the defect classification result of the target local area is determined to be an excessively deep defect. Determine the average response energy value of the target local region; if the average response energy value of the target local region is less than the response energy level of the normal texture region, then determine the defect classification result of the target local region as a texture missing type defect; Determine the regional periodicity score of the local area of ​​the target; if the regional periodicity score is less than the periodicity feature, it is determined to be a periodic anomaly defect.

5. The method as described in claim 1, characterized in that, The measured texture image and the reference texture image are compared by performing a difference operation to extract a clean texture information layer image, including: Obtain the gray level of a specified reference region in the baseline texture image, obtain the gray level of the same spatial location in the measured texture image corresponding to the specified reference region, and calculate the ratio between the two as the reflectance calibration coefficient; Generate a reference texture image scaled by the reflectivity calibration coefficient; A pixel-by-pixel difference operation is performed between the measured texture image and the reference texture image scaled by the reflectivity calibration coefficient to obtain a difference image; the difference image includes texture information. The difference image is divided into multiple local sub-blocks, and histogram stretching is performed on each local sub-block. Then, the local sub-blocks are stitched together and the transition boundaries between the local sub-blocks are smoothed to obtain a difference image with enhanced contrast. The contrast-enhanced differential image is then subjected to coarse-grained filtering and fine-grained filtering sequentially to obtain a clean texture information layer image.

6. The method as described in claim 5, characterized in that, The multidimensional texture features include grayscale uniformity features, directional consistency features, and periodicity features; the directional consistency features are used to indicate the overall dominant texture direction in the clean texture information layer image; extracting the periodicity features from the clean texture information layer image includes: Based on the overall texture dominance direction indicated by the directional consistency feature, the gray value of each pixel is extracted in the difference image along the overall texture dominance direction to generate a one-dimensional gray-scale profile sequence. Determine the normalized autocorrelation function of the one-dimensional grayscale profile sequence; Within the positive displacement range of the autocorrelation function, local maxima greater than a preset significance threshold are identified, and these local maxima are taken as effective peak values; the displacement position and amplitude of each effective peak value are recorded. Based on the autocorrelation function, the displacement position and amplitude of each effective peak value, determine the interval regularity coefficient, amplitude stability coefficient and peak value significance coefficient; The interval regularity coefficient, the amplitude stability coefficient, and the peak significance coefficient are weighted and summed according to preset weights to obtain a periodic score; the periodic score is a periodic feature.

7. The method as described in claim 1, characterized in that, The multidimensional texture features include grayscale uniformity features, directional consistency features, and periodic features; Based on the multidimensional texture features and the defect classification results, a uniformity evaluation result is output, including: The gray-level uniformity feature, the orientation consistency feature, and the periodicity feature are normalized to obtain normalized gray-level uniformity feature, orientation consistency feature, and periodicity feature. The normalized gray-scale uniformity feature, directional consistency feature, and periodicity feature are weighted and summed according to preset weights to obtain a comprehensive score. The comprehensive score is compared with multiple preset grading thresholds to determine the uniformity evaluation result of the workpiece under test.

8. A device for evaluating the uniformity of surface texture of a curved workpiece, characterized in that, The device includes: The acquisition module is used to acquire a reference texture image of a standard curved surface workpiece and a measured texture image of the curved surface workpiece to be tested; the reference texture image includes a theoretical reflective gradient distribution generated by the curvature of the workpiece surface; The extraction module is used to perform a difference operation between the measured texture image and the reference texture image to extract a pure texture information layer image; and to extract multi-dimensional texture features from the pure texture information layer image. The classification module is used to compare the measured texture image with the theoretical reflective gradient distribution to obtain the brightness deviation; and to determine the defect classification result of the measured texture image based on the brightness deviation and the multidimensional texture features. The output module is used to output a uniformity evaluation result based on the multidimensional texture features and the defect classification result.

9. An electronic device, characterized in that, The electronic device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-7.

Images