A method for evaluating vulcanization uniformity of an outer jacket of a rubber cable

Abstract

The application belongs to the technical field of image processing, and particularly relates to a vulcanization uniformity evaluation method for an outer sheath of a rubber-sheathed cable, which comprises the following steps: performing polar coordinate transformation on a CT cross-sectional image of the rubber-sheathed cable to obtain a two-dimensional developed image; performing radial statistics on the developed image, extracting a radial gray sequence, fitting a process reference curve, and calculating a radial residual sequence; calculating an average correlation length based on an autocorrelation function of the radial residual sequence, combining residual variance to construct a radial defect comprehensive index; extracting a circumferential gray sequence at different depths of the developed image, calculating a circumferential uniformity basic index, using spectrum analysis to obtain a main frequency amplitude, and constructing a circumferential defect intensity index; and comprehensively determining vulcanization uniformity according to the radial defect comprehensive index and the circumferential defect intensity index. The application can effectively separate a process trend from real defects, evaluate radial clusters and circumferential periodic unevenness, and improve the accuracy of cable quality evaluation.

Description

Technical Field

[0001] This invention relates to the field of image processing technology. More specifically, this invention relates to a method for evaluating the uniformity of vulcanization of the outer sheath of rubber-sheathed cables. Background Technology

[0002] Rubber-sheathed cables are widely used in power transmission, mining machinery, and industrial automation. Their outer sheath plays a crucial role in insulation protection, tensile strength, abrasion resistance, and environmental isolation. The outer sheath material is typically a rubber-based composite, which undergoes a continuous vulcanization process under high temperature and pressure during production to achieve cross-linking reactions between molecular chains, forming a stable three-dimensional network structure, thereby obtaining excellent elasticity, strength, and aging resistance.

[0003] The uniformity of the vulcanization process directly determines the consistency of the internal density distribution of the material. If there are local areas of under-sulfurization or over-sulfurization, it will lead to stress concentration, which can easily cause safety hazards such as cracking and brittle fracture, seriously affecting the product life and operational reliability.

[0004] Existing methods for assessing vulcanization quality primarily rely on offline sampling and physical property testing, such as tensile strength and hardness tests, or cross-sectional microscopic observation. These methods are typically destructive, making it impossible to perform full inspection of finished cables, and they only reflect the performance of local samples, failing to comprehensively characterize the vulcanization distribution across the entire cable cross-section. Furthermore, traditional visual inspection methods can only observe surface defects and are ineffective in detecting internal density differences caused by uneven raw material mixing or temperature gradients.

[0005] In recent years, industrial CT technology has been introduced into the cable inspection field due to its non-destructive testing advantages. However, direct observation of CT grayscale images is often limited by subjective human judgment and lacks quantitative evaluation standards. Furthermore, during the manufacturing process, the cable outer sheath exhibits a normal radial density gradient trend due to heat conduction mechanisms. This normal process trend often masks actual minute defect fluctuations. In addition, different types of sulfurization unevenness are difficult to distinguish under conventional statistical indicators; for example, clustered impurities and periodic mechanical fluctuations are difficult to differentiate using conventional statistical indicators, leading to difficulties in fault tracing.

[0006] Therefore, there is an urgent need for an evaluation method that can effectively eliminate normal process trends and distinguish different types of non-uniformity. Summary of the Invention

[0007] To address the technical problems of existing methods failing to effectively eliminate interference from normal process trends and distinguishing between different types of uneven vulcanization, this invention provides a method for evaluating the uniformity of vulcanization of the outer sheath of rubber-sheathed cables, comprising:

[0008] CT cross-sectional images of the outer sheath of the rubber-sheathed cable are acquired. The outer sheath region is identified and a polar coordinate system is established. The annular region of the outer sheath is mapped to a two-dimensional unfolded image through polar coordinate transformation. Radial statistical analysis is performed on the unfolded image to extract the radial gray-level sequence. A process reference curve is obtained by polynomial fitting of the radial gray-level sequence. The difference between the radial gray-level sequence and the process reference curve is calculated to obtain the radial residual sequence. Autocorrelation analysis is performed on the radial residual sequence to calculate the average correlation length. Based on the variance of the radial residual sequence and the average correlation length, the radial defect comprehensive index is determined. Circumferential gray-level sequences at multiple pre-selected radial positions in the unfolded image are extracted. The mean of the standard deviations of all circumferential gray-level sequences is calculated as the circumferential uniformity base index. Spectral analysis is performed on the circumferential gray-level sequence of the intermediate layer to obtain the dominant frequency amplitude. The circumferential defect intensity index is determined based on the circumferential uniformity base index and the dominant frequency amplitude. The vulcanization uniformity of the outer sheath of the rubber-sheathed cable is comprehensively evaluated based on the magnitude of the radial defect comprehensive index and the circumferential defect intensity index.

[0009] This invention utilizes polar coordinate transformation to expand the annular outer sheath into a rectangular domain, effectively separating radial and circumferential features and providing a regular data structure for subsequent directional analysis. By fitting the process baseline curve, this invention successfully isolates the normal low-frequency density trend caused by heat conduction, ensuring that the residual sequence contains only true information on sulfurization unevenness and avoiding inflated evaluation metrics. This invention introduces autocorrelation analysis to construct a comprehensive radial defect index, considering not only the fluctuation amplitude but also the spatial duration of the fluctuation, thus exhibiting higher sensitivity to large-scale cluster defects. Simultaneously, this invention constructs a circumferential defect intensity index through spectral analysis, accurately identifying periodic unevenness caused by equipment mechanical pulsation. Finally, this invention achieves precise grading of sulfurization quality and preliminary location of fault sources through a dual-index comprehensive evaluation.

[0010] Preferably, the step of mapping the annular region of the outer sheath into a two-dimensional unfolded image through polar coordinate transformation includes: establishing a polar coordinate system with the midpoint between the center of the outer edge and the center of the inner edge of the outer sheath as the origin. The annular region is mapped to an expanded image, and the gray values ​​of the pixels in the expanded image satisfy the expression: In the formula, Indicates the unfolded image at the 1st Line number The grayscale value at the column, Indicates the original CT image in coordinates grayscale value, and These represent the x and y coordinates of the origin of the polar coordinate system, respectively. Indicates the first The radius length corresponding to each radial position Indicates the first The angle value corresponding to each angle sampling point.

[0011] This invention establishes a polar coordinate system with the geometric center as the origin and performs transformations to eliminate geometric distortions caused by slight eccentricity during cable manufacturing. It transforms irregular annular regions into regular rectangular images, so that each column corresponds to a radial profile and each row corresponds to a circumferential profile. This decouples radial heat conduction characteristics from circumferential mechanical distribution characteristics in geometric space, providing a standardized data foundation for the subsequent targeted extraction of vulcanization uniformity indicators in different dimensions.

[0012] Preferably, the extraction of the radial grayscale sequence includes: for each radial position in the unfolded image, calculating the arithmetic mean of the pixel values ​​in all angular directions at that radial position as the representative grayscale of that radial position; and arranging the representative grayscale values ​​of all radial positions in radial order to form a radial grayscale sequence.

[0013] This invention effectively suppresses random noise and local gray-level abrupt changes caused by tiny impurity clusters during CT imaging by averaging the unfolded image along the circumference, using statistical principles. This enhances the signal-to-noise ratio of the trend signal reflecting the overall density change of the material, enabling the extracted radial gray-level sequence to more stably and accurately characterize the common radial sulfidation law of the entire cross section, providing a reliable basis for constructing a high-precision process reference curve.

[0014] Preferably, the method for obtaining the radial residual sequence is as follows: for each radial position, the value of the radial position in the radial grayscale sequence is subtracted from the fitted value of the radial position in the process reference curve to obtain the radial residual at the radial position; the radial residuals at all radial positions are used to form a radial residual sequence; and the radial residual sequence is subjected to zero-mean processing.

[0015] This invention effectively removes the normal low-frequency process trends dominated by heat conduction mechanisms by calculating the radial residual sequence. This ensures that the retained signal primarily contains high-frequency abnormal fluctuations caused by factors such as uneven raw material mixing and localized temperature anomalies. Zero-mean processing further eliminates the influence of the DC component, ensuring that subsequent fluctuation analysis purely reflects the inhomogeneity of sulfurization. This avoids artificially inflated uniformity indicators caused by normal process density gradients, thus improving the specificity of defect detection.

[0016] Preferably, the calculation of the average correlation length includes: calculating the autocorrelation function of the zero-mean radial residual sequence: In the formula, Indicates the number of lag steps. The autocorrelation value at time, and They represent the first and the The zero-mean radial residuals at each location Indicates the total number of radial pixels. Indicates the number of lag steps. Indicates radial index; retrieves the first time the autocorrelation function drops to its maximum value. The number of lag steps corresponding to the time multiplier is used as the average correlation length.

[0017] Preferably, the radial defect composite index satisfies the expression: In the formula, Indicates the radial defect comprehensive index; This represents the value of the autocorrelation function at zero lag; This represents the average correlation length of the radial residual sequence after zero-mean normalization. This indicates the total number of radial pixels.

[0018] The radial defect composite index of this invention integrates the amplitude and spatial scale of residual fluctuations, making it highly sensitive to large-scale clustered defects but insensitive to isolated small noises. This solves the problem that simply relying on variance indices cannot distinguish defect types, and can more accurately assess the potential harm of large-scale mixing unevenness to the mechanical properties of cables, achieving in-depth quantification of radial vulcanization quality.

[0019] Preferably, the step of extracting the circumferential grayscale sequence of multiple pre-selected radial positions in the unfolded image includes: selecting A representative radial position , Extract the unfolded image. All the grayscale data in the row constitute the first row. A circumferential grayscale sequence, The number of radial positions selected. For the first The row index of the selected radial position in the unfolded image.

[0020] Preferably, the circumferential defect strength index satisfies the expression: In the formula, Indicates the circumferential defect strength index; Indicates the basic index of circumferential uniformity; Indicates the clock speed; Indicates the spectral amplitude at the dominant frequency; This represents the average spectral amplitude in the non-dominant frequency region; This is the adjustment coefficient.

[0021] This invention analyzes circumferential grayscale sequences at different depths and constructs a circumferential defect intensity index based on spectral characteristics. This effectively captures periodic density fluctuations caused by extruder screw pulsation or uneven die rotation. The spectral term in the circumferential defect intensity index highlights periodic components using the signal-to-noise ratio. The index is significantly amplified only when the energy of the periodic fluctuation is significantly higher than that of random noise. This allows for sensitive identification of hidden mechanical process defects, avoiding misjudgments caused by accidental noise and providing direct quantitative evidence for assessing equipment operational stability.

[0022] Preferably, the comprehensive evaluation includes: determining uniform sulfidation when the radial defect comprehensive index is less than a preset radial defect comprehensive threshold and the circumferential defect intensity index is less than a preset circumferential defect intensity threshold; determining radial clustering unevenness when the radial defect comprehensive index is greater than or equal to a preset radial defect comprehensive threshold and the circumferential defect intensity index is less than a preset circumferential defect intensity threshold; determining circumferential periodic unevenness when the radial defect comprehensive index is less than a preset radial defect comprehensive threshold and the circumferential defect intensity index is greater than or equal to a preset circumferential defect intensity threshold; and determining severe unevenness when the radial defect comprehensive index is greater than or equal to a preset radial defect comprehensive threshold and the circumferential defect intensity index is greater than or equal to a preset circumferential defect intensity threshold.

[0023] Preferably, the polynomial fitting employs the least squares method.

[0024] The beneficial effects of this invention are as follows: This invention successfully removes the interference of normal density gradient caused by heat conduction through polar coordinate transformation and baseline curve fitting, making the evaluation results directly point to the actual sulfurization defects; This invention introduces autocorrelation length and spectral analysis to construct a radial defect comprehensive index and a circumferential defect intensity index, which not only quantifies the degree of non-uniformity but also effectively distinguishes between radial cluster defects and circumferential periodic defects, providing clear guidance for adjusting the production process; This invention performs non-destructive full inspection based on CT images, significantly improving the coverage and efficiency of cable quality control. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating an evaluation method for the uniformity of vulcanization of the outer sheath of a rubber-sheathed cable according to the present invention.

[0026] Figure 2 This is a schematic CT image of a cable;

[0027] Figure 3 This is a schematic representation of the result of the Hough circle transformation;

[0028] Figure 4 This is a schematic representation of the unfolded image;

[0029] Figure 5 This is a schematic diagram illustrating the radial grayscale sequence and the process reference curve. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0032] This invention discloses a method for evaluating the uniformity of vulcanization of the outer sheath of rubber-sheathed cables, referring to... Figure 1 This includes steps S1 to S5:

[0033] S1. Obtain the CT cross-sectional image of the outer sheath of the rubber-sheathed cable, identify the outer sheath area and establish a polar coordinate system, and map the annular area of ​​the outer sheath into a two-dimensional unfolded image through polar coordinate transformation.

[0034] It should be noted that the uniformity of vulcanization of the outer sheath of a rubber-sheathed cable is essentially reflected in the spatial distribution characteristics of its material density. Industrial CT technology can non-destructively obtain density information inside an object, and the gray value of its reconstructed image is positively correlated with the local electron density. Therefore, this invention uses CT cross-sectional images for quantitative analysis.

[0035] Specifically, an industrial CT scanner is used to perform a tomographic scan of the target cable sample to obtain a high-resolution CT image, which is a grayscale image. For example, Figure 2 This is a CT image of a cable.

[0036] It should be noted that the CT image contains multiple structures such as conductors, insulating layers, fillers, and outer sheaths, while the object of interest in this invention is only the annular region formed by the outer sheath. In order to avoid introducing irrelevant grayscale information into non-target areas and causing subsequent analysis deviations, this invention accurately identifies the boundary of the outer sheath.

[0037] Specifically, the CT image is first subjected to Gaussian filtering to suppress imaging noise. Then, the Canny edge detection algorithm is used to extract edges from the Gaussian-filtered CT image. Hough circle transform is applied to fit the extracted edges to determine the center and radius of the outer edge of the outer sheath, as well as the center and radius of the inner edge of the outer sheath. For example, Figure 3 The result is the result of the Hough circle transform test.

[0038] It should be noted that, due to slight eccentricity during manufacturing, the two centers may not perfectly coincide. Using the outer edge center as the origin of polar coordinates directly would introduce geometric distortion, affecting the accuracy of subsequent transformations. Therefore, this invention uses the midpoint between the center of the outer edge and the center of the inner edge of the outer sheath as the origin of polar coordinates, and uses the radius of the inner edge and the radius of the outer edge as the boundary between the inner and outer radii. Furthermore, since the outer sheath has a ring-like structure in Cartesian coordinates, it is difficult to directly perform directional sectioning analysis along the radial and circumferential directions. Due to the lack of a regularized spatial mapping relationship, it is impossible to effectively separate radial and circumferential features. Therefore, this invention uses polar coordinate transformation to expand it into a rectangular domain.

[0039] Specifically, the midpoint between the center of the outer edge and the center of the inner edge of the outer sheath. Establish a polar coordinate system with the origin. , to the ring area , Mapped to a new two-dimensional unfolded image row index Corresponding radial position Column index Corresponding angle ,in The radius of the inner edge of the outer sheath. The radius of the outer edge of the outer sheath. These are the radial position coordinates in the polar coordinate system. These are the angular and directional coordinates in the polar coordinate system. The sampling step size is in the angular direction. The total number of sampling points in the angular direction, in this invention ,but .

[0040] The grayscale value of any pixel in the new two-dimensional image is calculated from the CT image using bilinear interpolation:

[0041]

[0042] In the formula, Indicates the unfolded image at the 1st Line number The grayscale value at the column, Indicates the original CT image in coordinates grayscale value, and These represent the x and y coordinates of the origin of the polar coordinate system, respectively. Indicates the first The radius length corresponding to each radial position Indicates the first The angle values ​​corresponding to each angle sampling point. and This is the result of trigonometric function operations. When... As it increases, the corresponding position moves from the inside to the outside; when As the size increases, the corresponding position rotates counterclockwise. This transformation flattens the original ring structure into a rectangular image, where each column represents a radial profile and each row represents a circumferential profile, providing a regular data structure for subsequent separation of radial and circumferential features. The total number of radial pixels is [missing information]. ,in The radius of the inner edge of the outer sheath. The radius of the outer edge of the outer sheath.

[0043] For example, Figure 4 To expand the image.

[0044] S2. Perform radial statistical analysis on the unfolded image, extract the radial grayscale sequence, and perform polynomial fitting on the radial grayscale sequence to obtain the process reference curve.

[0045] It should be noted that in the unfolded image, a single radial profile may be affected by CT imaging noise, small impurity clusters or local defects. If its grayscale distribution is directly used as the analysis object, the resulting trend is easily disturbed by random disturbances and cannot stably represent the radial sulfurization law common to the entire cross section. Therefore, this invention performs average processing on all circumferential profiles to suppress random disturbances and enhance the stability of the trend signal.

[0046] Specifically, for each radial position in the unfolded image, pixel values ​​in all angular directions at that radial position are extracted, and the arithmetic mean of these pixel values ​​is calculated to obtain a statistical value reflecting the overall grayscale level of that radial position, which serves as the representative grayscale value for that radial position. By traversing all radial positions and repeating this operation, the representative grayscale values ​​of all radial positions are combined into a sequence, which is then used as the radial grayscale sequence. This radial grayscale sequence can effectively suppress random noise interference, highlight the overall radial variation trend, and more stably and accurately characterize the common features of the entire cross-section, providing a reliable foundation for the subsequent construction of process reference curves.

[0047] It should be noted that the radial grayscale sequence contains two types of components: one is a low-frequency, smooth trend term caused by heat conduction, and the other is a high-frequency fluctuation term caused by uneven mixing of raw materials or local anomalies. If its variance is directly used as a uniformity index, it will still lead to an artificially high index. Therefore, this invention uses the least squares method to perform polynomial fitting on the radial grayscale sequence to construct a process baseline curve that can accurately describe the normal process trend.

[0048] Specifically, the least squares method is used to perform polynomial fitting on the radial grayscale sequence, and the fitting result is used as the process baseline curve. The choice of polynomial order must balance fitting ability and generalization performance. Too low an order cannot capture the nonlinear changes in the trend, while too high an order is easily affected by local noise. This embodiment uses a cubic polynomial for fitting, which is suitable for most gradual temperature gradient scenarios dominated by heat conduction. In other embodiments, implementers can choose the polynomial order according to the actual implementation situation. For example, when the sheath is thick and the temperature gradient is significantly nonlinear, a higher order polynomial can be selected.

[0049] It should be noted that the process baseline curve can effectively separate the low-frequency trend caused by heat conduction, providing a reliable benchmark for subsequent calculations of actual vulcanization non-uniformity and ensuring that the uniformity assessment is not affected by process trends. For example, Figure 5 This is a schematic diagram of the radial grayscale sequence and the process reference curve.

[0050] S3. Calculate the difference between the radial grayscale sequence and the process reference curve to obtain the radial residual sequence. Perform autocorrelation analysis on the radial residual sequence to calculate the average correlation length. Based on the variance of the radial residual sequence and the average correlation length, determine the radial defect comprehensive index.

[0051] It should be noted that the radial grayscale sequence obtained in the previous step reflects the average density distribution trend of the entire outer sheath cross section in the radial direction, while the process reference curve obtained by polynomial fitting characterizes the normal vulcanization gradient law dominated by the heat conduction mechanism. Since the presence of this trend component will significantly interfere with the subsequent identification of real defects, if the statistical fluctuation index is directly calculated based on the radial grayscale sequence, it will lead to an overestimation of the evaluation results. Therefore, this invention compares the radial grayscale sequence and the process reference curve point by point and extracts the deviation signal between the two as residual information reflecting the actual vulcanization unevenness.

[0052] Specifically, for each radial position, the representative gray level at that radial position is subtracted from the fitted value of that radial position in the process reference curve to obtain the radial residual at that radial position. The radial residuals at all radial positions are then used to construct a radial residual sequence.

[0053] It should be noted that the radial residual sequence obtained by this invention effectively strips away the gradual process trend, so that the radial residual sequence mainly contains high-frequency abnormal fluctuations caused by factors such as uneven mixing of raw materials, local temperature anomalies, or improper stirring, thus providing a clean data foundation for subsequent accurate evaluation.

[0054] It should be further explained that it is difficult to distinguish the type of defect by the standard deviation or variance of the residual signal alone. For example, a large number of small random fluctuations and a few large-scale cluster defects may have similar statistical amplitudes, but the latter is far more harmful to the mechanical performance of the cable than the former. In order to improve the depth of the assessment and the engineering guidance value, this invention introduces a signal correlation analysis method to quantify the spatial persistence characteristics of residual fluctuations.

[0055] Specifically, before performing correlation calculations, the radial residual sequence is first subjected to zero-mean processing. This involves subtracting the overall arithmetic mean of the radial residual sequence from each data point in the original sequence, resulting in a signal with a mean of zero. This operation aims to eliminate the DC offset component in the signal, ensuring that subsequent autocorrelation analysis reflects only the similarity between fluctuating components, unaffected by the overall bias, thereby improving the accuracy and stability of spatial correlation measurement.

[0056] Furthermore, the autocorrelation function of the radial residual sequence after zero-mean normalization is calculated:

[0057]

[0058] In the formula, Indicates the number of lag steps. The autocorrelation value at time, and They represent the first and the The zero-mean radial residuals at each location Indicates the total number of radial pixels. This represents the number of lag steps, with a value range of [value missing]. to , Indicates radial index. When hour, The variance of the radial residual sequence after zero-mean normalization represents the fluctuation power; as... Increase Generally, it exhibits a monotonically decreasing trend, and its decay rate directly reflects the radial continuity of grayscale anomalies.

[0059] It should be noted that the decay rate of the autocorrelation function is directly related to the spatial expansion characteristics of the defect. To transform it into a quantifiable physical indicator, this invention defines the average correlation length. For the autocorrelation function to first drop to its maximum value The number of lag steps corresponding to the time multiplication factor:

[0060]

[0061] In the formula, This represents the average correlation length of the radial residual sequence after zero-mean normalization. This represents the value of the autocorrelation function at zero lag. This represents the reciprocal of the natural constant. When the average correlation length... A larger value indicates stronger spatial persistence of the radial residual fluctuations after zero-mean normalization, corresponding to larger-scale sulfidation heterogeneous clusters; when the average correlation length... The smaller the value, the faster the radial fluctuations decay after zero-meaning, corresponding to small or isolated defects.

[0062] To further enhance the comprehensiveness of the assessment, this invention defines a radial defect comprehensive index to simultaneously reflect the amplitude and spatial scale of residual fluctuations:

[0063]

[0064] In the formula, Indicates the radial defect comprehensive index; Let the value of the autocorrelation function at zero lag be equal to the variance of the radial residual sequence after zero mean normalization. The standard deviation of the radial residual sequence after zero-mean normalization reflects the magnitude of non-uniformity; This represents the average correlation length of the radial residual sequence after zero-mean normalization. Indicates the total number of radial pixels; This is used to normalize the correlation length, eliminating scale effects caused by differences in image resolution or sheath thickness. When The larger or When the value is larger, the radial defect comprehensive index increases. A larger value indicates the presence of significant and large-scale radial inhomogeneities. The radial defect composite index is highly sensitive to large-scale clusters; it only becomes significant when fluctuations are weak and the spatial scale is small. This will keep the level low, thus enabling in-depth quantification and risk classification of the true degree of uneven sulfurization.

[0065] S4. Extract the circumferential grayscale sequences from multiple pre-selected radial positions in the unfolded image, calculate the mean of the standard deviations of all circumferential grayscale sequences as the basic index of circumferential uniformity, perform spectral analysis on the circumferential grayscale sequences of the intermediate layer to obtain the main frequency amplitude, and determine the circumferential defect strength index based on the basic index of circumferential uniformity and the main frequency amplitude.

[0066] It should be noted that uneven vulcanization may also manifest as circumferential periodic fluctuations, such as those caused by screw pulsation in the extruder or uneven rotation of the die. Such defects are difficult to reflect their periodic characteristics in a single standard deviation index, making it impossible to accurately locate the source of the fault. Therefore, this invention introduces spectrum analysis to reveal the dominant frequency characteristics of circumferential fluctuations.

[0067] Specifically, in the unfolded image, each row corresponds to a fixed radial position. All the grayscale values ​​arranged column-by-column in that row collectively record complete information about the continuous variation of material density with circumferential angle at a specific radius of the outer sheath. To evaluate the circumferential uniformity at different depths, this invention selects... A representative radial position , Extract the unfolded image. All the grayscale data in the row constitute the first row. A circumferential grayscale sequence, the circumferential grayscale sequence containing The grayscale values ​​arranged in angular order fully characterize the circumferential density fluctuation at that radial position. This represents the total number of sampling points in the angular direction. The number of radial positions selected. For the first The row index of the selected radial position in the unfolded image.

[0068] To cover material layers of different depths, this embodiment sets... They are located near the inner wall. Intermediate layer and near the outer skin Among them, This is the row index of the first selected radial position in the unfolded image, i.e., the radial position index closest to the inner wall. This is the row index of the second selected radial position in the unfolded image, i.e., the radial position index of the intermediate layer. The row index of the third selected radial position in the unfolded image, i.e., the radial position index closest to the outer skin. The radius of the inner edge of the outer sheath. The radius of the outer edge of the outer sheath. The floor symbol is used for rounding down. To buffer the distance and avoid the influence of boundary effects, Pick to In this embodiment, pixels are set In other embodiments, implementers may choose according to the actual implementation situation. The value and buffer distance The value of can be adjusted, for example, the number of sampling layers can be increased for thick-walled sheaths.

[0069] Furthermore, the standard deviation of each circumferential grayscale sequence is calculated. The standard deviation of all circumferential grayscale sequences The arithmetic mean of is used as the basic index for circumferential uniformity, where For the first The standard deviation of a circumferential grayscale sequence.

[0070] It should be noted that the standard deviation of each circumferential grayscale sequence reflects the severity of the circumferential density fluctuation at that layer. The larger the standard deviation, the more significant the grayscale fluctuation and the worse the uniformity at that layer. The arithmetic mean of the standard deviations of all circumferential grayscale sequences comprehensively reflects a unified index of the circumferential fluctuation level at multiple depth levels, which can effectively suppress the impact of local anomalies at a single layer on the overall assessment and improve the representativeness and robustness of the results.

[0071] Furthermore, in order to identify whether grayscale fluctuations have clear periodic characteristics, this invention performs a fast Fourier transform on the circumferential grayscale sequence located in the intermediate layer, transforming it from the angle domain to the frequency domain to obtain the corresponding discrete spectrum, and obtains the frequency index with the largest amplitude among the non-zero frequencies in the discrete spectrum as the main frequency.

[0072] It should be noted that the energy of the dominant frequency component directly reflects the intensity of the periodic density fluctuation. The higher the energy, the more obvious the non-uniformity of the material at that spatial frequency, and the greater the potential harm to the mechanical properties of the cable. Therefore, this invention determines the circumferential defect intensity index based on the circumferential uniformity basic index and the spectral amplitude of the dominant frequency.

[0073] Specifically, the circumferential defect strength index satisfies the following expression:

[0074]

[0075] In the formula, Indicates the circumferential defect strength index; It represents the circumferential uniformity basic index, reflecting the amplitude of fluctuations; This represents the dominant frequency, which is the index of the frequency with the largest amplitude among the non-zero frequencies in the discrete spectrum. Indicates the spectral amplitude at the dominant frequency; This represents the average spectral amplitude in the non-dominant frequency region, used to characterize the background noise level. Less than the smallest positive number To avoid computational overflow, it is stipulated that... In this embodiment, the smallest positive number In other embodiments, implementers may set the appropriate parameters according to the actual implementation situation. ,but Must meet ; This is an adjustment coefficient used to control the enhancement weight of the periodic component. It is only applied when the energy of the periodic fluctuation is significantly higher than the random noise. Only then will it be effectively amplified, thus avoiding misjudging random noise as a serious defect.

[0076] It should be noted that the adjustment coefficient The value of needs to balance the evaluation sensitivity and stability. If the value is too small, the response to periodic defects will be insufficient, reducing diagnostic capability; if... If the value is too high, it may excessively amplify slight fluctuations, leading to an increase in the false alarm rate. In this embodiment, based on statistical analysis and verification experiments on multiple batches of rubber-sheathed cable samples, the following settings are configured: This value enables effective enhanced response to periodic defects under typical operating conditions while maintaining stable operation of the overall assessment system. In other embodiments, implementers can adjust the value according to specific cable specifications, CT image resolution, or quality control level. The value can be adjusted, for example, to 1.2 in high-precision detection scenarios and to 0.5 in coarse screening scenarios, to adapt to different detection needs.

[0077] S5. The uniformity of vulcanization of the outer sheath of the rubber-sheathed cable is comprehensively evaluated based on the radial defect comprehensive index and the circumferential defect strength index.

[0078] Specifically, based on production process requirements and analysis of historical qualified sample data, a radial defect comprehensive threshold and a circumferential defect intensity threshold are preset; in this embodiment, the radial defect comprehensive threshold is set to... The circumferential defect strength threshold is This value is determined based on the 95th percentile of the statistical distribution of the radial defect composite index and the circumferential defect strength index of a batch of known qualified samples. In other embodiments, implementers can set specific values ​​for the radial defect composite threshold and the circumferential defect strength threshold according to the actual implementation situation, for example, adjusting the threshold according to different cable models or quality grades.

[0079] Specifically, if the radial defect comprehensive index is less than the preset radial defect comprehensive threshold and the circumferential defect strength index is less than the preset circumferential defect strength threshold, the product is judged to be uniformly vulcanized and qualified.

[0080] If the radial defect comprehensive index is greater than or equal to the preset radial defect comprehensive threshold and the circumferential defect strength index is less than the preset circumferential defect strength threshold, it is determined that the radial clusters are uneven and unqualified. At this time, the staff is reminded to check the raw material mixing uniformity or vulcanization temperature gradient control.

[0081] If the radial defect comprehensive index is less than the preset radial defect comprehensive threshold and the circumferential defect strength index is greater than or equal to the preset circumferential defect strength threshold, it is judged as circumferential periodic unevenness and unqualified. At this time, the staff is reminded to check the running stability of the extruder screw or the installation accuracy of the mold.

[0082] If the radial defect comprehensive index is greater than or equal to the preset radial defect comprehensive threshold, and the circumferential defect strength index is greater than or equal to the preset circumferential defect strength threshold, it is judged as severely uneven and unqualified. At this time, the staff is reminded to conduct a comprehensive inspection of the raw material supply, mixing process, extrusion molding and heat conduction control system of the vulcanization production line, and focus on checking the consistency of equipment cooperative operation status and process parameters.

Claims

1. A method for evaluating the uniformity of vulcanization of the outer sheath of rubber-sheathed cables, characterized in that, include: Obtain CT cross-sectional images of the outer sheath of the rubber-sheathed cable, identify the outer sheath area and establish a polar coordinate system, and map the annular area of ​​the outer sheath into a two-dimensional unfolded image through polar coordinate transformation; Radial statistical analysis is performed on the unfolded image to extract the radial grayscale sequence. Polynomial fitting is performed on the radial grayscale sequence to obtain the process reference curve. The difference between the radial grayscale sequence and the process reference curve is calculated to obtain the radial residual sequence. Autocorrelation analysis is performed on the radial residual sequence to calculate the average correlation length. Based on the variance of the radial residual sequence and the average correlation length, the radial defect comprehensive index is determined. Extract the circumferential grayscale sequences from multiple pre-selected radial positions in the unfolded image, calculate the mean of the standard deviations of all circumferential grayscale sequences as the basic index of circumferential uniformity, perform spectral analysis on the circumferential grayscale sequences of the intermediate layer to obtain the main frequency amplitude, and determine the circumferential defect strength index based on the basic index of circumferential uniformity and the main frequency amplitude. The uniformity of vulcanization of the outer sheath of the rubber-sheathed cable is comprehensively evaluated based on the magnitude of the radial defect comprehensive index and the circumferential defect strength index. Calculating the average correlation length includes: calculating the autocorrelation function of the zero-mean radial residual sequence. In the formula, Indicates the number of lag steps. The autocorrelation value at time, and They represent the first and the The zero-mean radial residuals at each location Indicates the total number of radial pixels. Indicates the number of lag steps. Indicates radial index; retrieves the first time the autocorrelation function drops to its maximum value. The number of lag steps corresponding to the time multiplication factor is used as the average correlation length; The radial defect composite index satisfies the following expression: In the formula, Indicates the radial defect comprehensive index; This represents the value of the autocorrelation function at zero lag; This represents the average correlation length of the radial residual sequence after zero-mean normalization. The circumferential defect strength index satisfies the following expression: In the formula, Indicates the circumferential defect strength index; Indicates the basic index of circumferential uniformity; Indicates the clock speed; Indicates the spectral amplitude at the dominant frequency; This represents the average spectral amplitude in the non-dominant frequency region; This is the adjustment coefficient.

2. The method for evaluating the uniformity of vulcanization of the outer sheath of rubber-sheathed cables according to claim 1, characterized in that, The process of mapping the annular region of the outer sheath into a two-dimensional unfolded image through polar coordinate transformation includes: Establish a polar coordinate system with the midpoint between the center of the outer edge and the center of the inner edge of the outer sheath as the origin. The annular region is mapped to an expanded image, and the gray values ​​of the pixels in the expanded image satisfy the expression: In the formula, Indicates the unfolded image at the 1st Line number The grayscale value at the column, Indicates the original CT image in coordinates grayscale value, and These represent the x and y coordinates of the origin of the polar coordinate system, respectively. Indicates the first The radius length corresponding to each radial position Indicates the first The angle value corresponding to each angle sampling point.

3. The method for evaluating the uniformity of vulcanization of the outer sheath of rubber-sheathed cables according to claim 1, characterized in that, The extraction of the radial grayscale sequence includes: For each radial position in the unfolded image, calculate the arithmetic mean of the pixel values ​​in all angular directions at that radial position, and use it as the representative gray level for that radial position; arrange the representative gray levels of all radial positions in radial order to form a radial gray level sequence.

4. The method for evaluating the uniformity of vulcanization of the outer sheath of a rubber-sheathed cable according to claim 1, characterized in that, The method for obtaining the radial residual sequence is as follows: For each radial position, the value of that radial position in the radial grayscale sequence is subtracted from the fitted value of that radial position in the process reference curve to obtain the radial residual at that radial position; the radial residuals at all radial positions are used to form a radial residual sequence; the radial residual sequence is then zero-mean processed.

5. The method for evaluating the uniformity of vulcanization of the outer sheath of a rubber-sheathed cable according to claim 1, characterized in that, The extraction of circumferential grayscale sequences from multiple pre-selected radial positions in the unfolded image includes: Select A representative radial position , Extract the unfolded image. All the grayscale data in the row constitute the first row. A circumferential grayscale sequence, The number of radial positions selected. For the first The row index of the selected radial position in the unfolded image.

6. The method for evaluating the uniformity of vulcanization of the outer sheath of a rubber-sheathed cable according to claim 1, characterized in that, The comprehensive evaluation includes: If the radial defect comprehensive index is less than a preset radial defect comprehensive threshold and the circumferential defect intensity index is less than a preset circumferential defect intensity threshold, the system is judged to be uniformly sulfurized. If the radial defect comprehensive index is greater than or equal to a preset radial defect comprehensive threshold and the circumferential defect intensity index is less than a preset circumferential defect intensity threshold, the system is judged to be radially clustered uneven. If the radial defect comprehensive index is less than a preset radial defect comprehensive threshold and the circumferential defect intensity index is greater than or equal to a preset circumferential defect intensity threshold, the system is judged to be circumferentially periodic uneven. If the radial defect comprehensive index is greater than or equal to a preset radial defect comprehensive threshold and the circumferential defect intensity index is greater than or equal to a preset circumferential defect intensity threshold, the system is judged to be severely uneven.

7. The method for evaluating the uniformity of vulcanization of the outer sheath of a rubber-sheathed cable according to claim 1, characterized in that, The polynomial fitting uses the least squares method.

Images