A method and system for detecting the groove pattern of a rubber-coated roller
By employing multispectral detection and intelligent residue identification technologies, the problem of inaccurate accuracy and prediction caused by residue interference in the morphology detection of grooves on coating rollers has been solved. This has enabled high-precision morphology detection and reliable life prediction, thereby improving production efficiency and product quality.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHONGQING WANGBIAN ELECTRIC GRP CORP
- Filing Date
- 2026-03-24
- Publication Date
- 2026-07-10
Smart Images
Figure CN122360331A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of coating roller inspection technology, and more specifically, to a method and system for inspecting the groove morphology of coating rollers. Background Technology
[0002] In continuous industrial production of steel strip coating, the coating roller is a core component. The morphology of its surface grooves directly determines the accuracy of coating metering and transfer, thus affecting the uniformity and quality of the steel strip coating. To ensure product quality and optimize production efficiency, accurate online detection and life prediction of the coating roller's groove morphology are crucial. Traditional detection methods are inefficient and prone to large errors, making it difficult to meet the high precision and real-time requirements of modern production lines.
[0003] Although laser scanning technology has been introduced to achieve non-contact online inspection, aiming to accurately quantify rubber roller wear and predict its lifespan, in practical applications, paint mist and solvent vapors generated during the coating process can form an uneven residue film on the grooved surface of the rubber roller. This uneven residue film has optical reflection properties that differ significantly from the rubber material of the roller itself. When the laser beam used for inspection scans these areas covered by the residue film, the laser echo signal is interfered with, manifesting as localized and irregular distortions in the intensity and shape of the laser signal. The mechanism of this interference lies in the fact that when the laser penetrates or reflects the residue film, its energy attenuation, scattering angle, and spectral response differ from when it directly irradiates the rubber surface, causing the signal captured by the receiver to no longer accurately reflect the true geometric position of the rubber surface.
[0004] The optical signal distortions caused by the residual film are directly reflected in the raw point cloud data generated by laser scanning. At specific locations along the groove contour, the point cloud data will exhibit numerous outliers, or the data points will show local deviations from the true contour. When subsequent piecewise functions are used to fit the groove contour to these point cloud data, especially when accurately calculating the wear-sensitive groove bottom fillet and groove edge, these outliers or locally deviated data will significantly affect the accuracy of the fitting, resulting in significant errors in the calculated groove depth, angle, or fillet radius. The fitting algorithm, attempting to account for all data points, will be "pulled off course" by these outliers, causing the final fitted curve to deviate from the true physical contour of the rubber roller.
[0005] To address outliers and data deviations caused by residual film, technicians typically attempt to enhance filtering during data preprocessing to eliminate interference. However, due to the randomness and complexity of the residual film's distribution, its interference characteristics on laser signals exhibit high uncertainty. In this situation, while general brute-force filtering algorithms remove these interference points, they inevitably "smooth out" some subtle geometric changes that manifest as true wear characteristics such as rounded groove edges and shallower groove bottoms. The consequence of this over-filtering is that the calculated wear amount is systematically underestimated, causing the system-reported roller wear condition to always be better than its actual physical state. This underestimation of wear directly misleads production managers' judgment of roller lifespan, leading to unexpected defective products, increased scrap rates, and even the need for emergency shutdowns to replace rollers, causing production interruptions and significant economic losses.
[0006] A further layer of complexity lies in the fact that coating production lines typically use different batches of paint, which may vary in chemical composition, viscosity, volatility, and other properties. This results in differences in the deposition characteristics of the paint mist on the roller surface and the optical properties of the cured residue film. Simultaneously, variations in the temperature and humidity of the workshop environment significantly affect the curing speed of the paint mist and the physical morphology of the residue film. These factors combined mean that the characteristics interfering with the laser signal are constantly changing. This means that a fixed compensation formula cannot be established through a one-time parameter calibration to correct for detection errors caused by the residue film. Therefore, the lifetime prediction regression formula established based on historical wear data and coating quality data will experience a significant decrease in prediction accuracy, or even frequent failure, when faced with such dynamically changing interference, failing to provide reliable lifetime predictions.
[0007] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention
[0008] The purpose of this application is to provide a method and system for detecting the groove morphology of coated rollers, aiming to solve the problems in the prior art such as decreased accuracy due to interference from residues, inaccurate life prediction, and inability to adapt to dynamic interference in the detection of groove morphology of coated rollers.
[0009] In a first aspect, this application provides a method for detecting the groove morphology of a coating roller, the method comprising the following steps: A1. Obtain the reflected signal intensity and original position data at the same location on the surface of the coating roller under at least two preset wavelengths; A2. Extract the spectral feature parameters of the corresponding position based on the intensity of the reflected signal; A3. Compare the spectral characteristic parameters with the preset substrate material characteristic range to identify whether the location belongs to a non-substrate region affected by residues, and determine the degree of interference if it belongs to a non-substrate region; A4. For locations identified as non-base regions, perform shape correction processing on the original location data according to the degree of interference to obtain corrected location data; wherein, the shape correction processing includes adjusting values according to a preset physical compensation model, or performing geometric inference based on the location data of adjacent base regions; A5. Use the corrected position data to perform contour fitting and extract the geometric feature parameters of the groove; A6. Calculate the remaining service life of the coating roller based on the geometric feature parameters and the preset wear limit.
[0010] Secondly, this application provides a system for detecting the groove morphology of a coating roller, the system comprising: The data acquisition module is used to acquire the intensity of reflected signals at the same location on the surface of the coating roller under at least two preset wavelengths, as well as the original position data. The feature extraction module is used to extract spectral feature parameters at the corresponding position based on the intensity of the reflected signal; The identification module is used to compare the spectral feature parameters with a preset substrate material feature range, identify whether the location belongs to a non-substrate region affected by residue interference, and determine the degree of interference when it belongs to a non-substrate region; The topography correction module is used to perform topography correction processing on the original location data for locations identified as non-base regions, based on the degree of interference, to obtain corrected location data; wherein, the topography correction processing includes adjusting values according to a preset physical compensation model, or performing geometric inference based on the location data of adjacent base regions; The contour fitting module is used to perform contour fitting using the corrected position data and extract the geometric feature parameters of the groove. The life calculation module is used to calculate the remaining service life of the coating roller based on the geometric feature parameters and the preset wear limit.
[0011] Beneficial Effects: This application provides a method and system for detecting the groove morphology of a coated roller. By introducing multispectral reflectance signal intensity and original position data, and combining the comparison of spectral characteristic parameters with the characteristic range of the substrate material, it can effectively identify non-substrate areas on the coated roller surface affected by residue interference, and correct the morphology of the original position data according to the degree of interference. This correction process includes adjusting values according to a preset physical compensation model or performing geometric inference based on the position data of adjacent substrate areas, thereby overcoming the problems of optical signal distortion and inaccurate point cloud data caused by residue film in traditional laser scanning technology under paint mist and solvent vapor environments. By contour fitting the corrected position data, more accurate groove geometric characteristic parameters can be extracted, avoiding the smoothing effect of excessive filtering on the true wear characteristics and solving the problem of underestimated wear. Finally, based on accurate geometric characteristic parameters and preset wear limits, the remaining service life of the coated roller is calculated, improving the accuracy and reliability of life prediction, effectively avoiding unexpected defective products, increased scrap rate, and production interruptions and economic losses such as emergency shutdowns, thereby significantly improving the accuracy of online inspection of coated rollers and the operating efficiency of the production line. Attached Figure Description
[0012] Figure 1 A flowchart illustrating a method for detecting the groove morphology of a coating roller provided in this application.
[0013] Figure 2 This is a schematic diagram of a detection system for the groove morphology of a coating roller provided in this application.
[0014] Labeling Explanation: 1. Data Acquisition Module; 2. Feature Extraction Module; 3. Recognition Module; 4. Shape Correction Module; 5. Contour Fitting Module; 6. Lifetime Calculation Module. Detailed Implementation
[0015] 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. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0016] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0017] Please refer to Figure 1 This application discloses a method for detecting the groove morphology of a coating roller in some embodiments, the method comprising the following steps: A1. Obtain the reflected signal intensity and original position data at the same location on the surface of the coating roller under at least two preset wavelengths; A2. Extract the spectral feature parameters of the corresponding position based on the intensity of the reflected signal; A3. Compare the spectral characteristic parameters with the preset substrate material characteristic range to identify whether the location belongs to a non-substrate region affected by residues, and determine the degree of interference if it belongs to a non-substrate region; A4. For locations identified as non-base regions, perform shape correction processing on the original location data according to the degree of interference to obtain corrected location data; wherein, the shape correction processing includes adjusting values according to a preset physical compensation model, or performing geometric inference based on the location data of adjacent base regions; A5. Use the corrected position data to perform contour fitting and extract the geometric feature parameters of the groove; A6. Calculate the remaining service life of the coating roller based on the geometric feature parameters and the preset wear limit.
[0018] The "coating roller" is a core component used to coat the surface of a steel strip with paint. Its surface is usually engraved with grooves of a specific shape to precisely control the metering and transfer of paint.
[0019] "Groove morphology" refers to the geometry of these grooves, including but not limited to parameters such as groove depth, groove width, groove corners, and fillet radius. These parameters directly affect the coating quality.
[0020] "Reflected signal intensity" refers to the energy of the reflected light captured by the receiver after the laser beam irradiates the surface of the rubber roller. Its intensity is affected by various factors such as surface material, morphology, and residues.
[0021] "Original location data" refers to the three-dimensional coordinate information of surface points obtained directly through methods such as laser scanning, and it usually exists in the form of point clouds.
[0022] "Spectral characteristic parameters" refer to numerical values extracted from the intensity of reflected signals at different wavelengths that can characterize the optical properties of a material, such as the ratio or normalized difference of reflection intensity at different wavelengths.
[0023] "Base material characteristic range" refers to the known distribution range of the intensity of reflected signals or spectral characteristic parameters of pure rubber roller material at different wavelengths, used to distinguish the rubber roller body from residues.
[0024] "Non-substrate area" refers to the area on the roller surface covered by residue, whose optical properties differ from the substrate material. Conversely, "substrate area" refers to the area on the roller surface not covered by residue.
[0025] "Interference level" quantifies the degree of influence of residues on the detection signal, and can be determined based on the degree of deviation between the spectral characteristic parameters and the characteristic range of the substrate material.
[0026] "morphological correction processing" refers to the process of correcting the original position data that has been disturbed by residues in order to restore its true geometric morphology, including two methods: physical compensation and geometric inference.
[0027] The "physical compensation model" is a model based on the optical properties of the residue (such as refractive index and thickness) and is used to correct optical path errors.
[0028] "Geometric inference" uses the geometric information of adjacent undisturbed regions to predict the true shape of the disturbed region.
[0029] "Corrected position data" refers to point cloud data that has undergone morphology correction processing, which more accurately reflects the true surface morphology of the rubber roller. It includes morphology-corrected position data of non-base regions and original position data of base regions.
[0030] "Contour fitting" refers to using mathematical methods (such as piecewise function fitting) to fit discrete point cloud data into a continuous curve in order to extract the geometric features of the groove.
[0031] "Geometric feature parameters" are quantitative indicators of the groove profile obtained by fitting, such as groove depth, groove width, cross-sectional area, etc.
[0032] "Preset wear limit" is the allowable wear range of the grooved geometric feature parameters set according to the design requirements of the rubber roller and the coating quality standards.
[0033] "Remaining service life" is the estimated length of time a rubber roller can continue to be used, based on its current wear condition and wear trend.
[0034] The detection method of this application is achieved through the following steps: In step A1, the reflected signal intensity and original position data at the same location on the coated roller surface are acquired at at least two preset wavelengths. This step is fundamental to the entire detection process, aiming to obtain sufficient information to distinguish the roller substrate material from residues and to preliminarily obtain the surface morphology. For example, a multispectral laser scanning system can be used, containing at least two power and gain-calibrated laser emitters that simultaneously emit laser beams of different wavelengths onto the surface to be tested. These laser beams can be visible, near-infrared, or mid-infrared light, with the specific wavelength selection depending on the optical absorption and reflection characteristics of the roller material and common residues. Simultaneously, a receiver equipped with a narrowband filter of the corresponding wavelength is used to simultaneously capture the signal reflected from the surface to be tested by the laser beam. The narrowband filter ensures that the receiver only captures reflected light of a specific wavelength, thereby obtaining multispectral reflection information at the same location. In addition, the system also simultaneously acquires the distance information of this location to determine the reflected signal intensity and the original position data. Another implementation method is to use a tunable laser that rapidly switches the emission wavelength in a short time, and combine it with a single broadband receiver and a spectrum analyzer to obtain the reflected signal intensity at different wavelengths. Raw location data can be obtained through non-contact ranging techniques such as laser triangulation, time-of-flight method, or phase modulation method.
[0035] In step A2, spectral feature parameters at the corresponding locations are extracted based on the intensity of the reflected signal. The purpose of this step is to transform the original multispectral reflected signal intensity into more discriminative feature values for subsequent residue identification. For example, the reflected signal intensity at different wavelengths can be directly used as spectral feature parameters. Alternatively, the intensity ratio between reflected signals at different wavelengths can be calculated, for example, dividing the reflected intensity at wavelength λ1 by the reflected intensity at wavelength λ2. This ratio can effectively eliminate the influence of fluctuations in light source intensity and distance variations, highlighting the spectral characteristics of the material itself. Alternatively, the normalized difference in intensity between reflected signals at different wavelengths can be calculated. This normalized difference also has good robustness in distinguishing differences in the spectral responses of different materials.
[0036] In step A3, the spectral characteristic parameters are compared with a preset substrate material characteristic range to identify whether the location belongs to a non-substrate region affected by residue interference, and if it belongs to a non-substrate region, the degree of interference is determined. This step is crucial for identifying residues. The preset substrate material characteristic range can be obtained through extensive measurement and statistical analysis of clean rubber roller substrate materials, forming a multi-dimensional characteristic space or numerical range. For example, if the spectral characteristic parameter is the ratio of reflection intensity at two wavelengths, then the substrate material characteristic range is a specific ratio range. When the spectral characteristic parameter at a certain location falls within this range, it is determined to be a substrate region; otherwise, it is determined to be a non-substrate region. After identifying the non-substrate region, the degree of interference also needs to be determined. For example, the degree of interference can be quantified based on the degree to which the spectral characteristic parameter deviates from the substrate material characteristic range. The greater the deviation, the higher the degree of interference.
[0037] In step A4, for locations identified as non-substrate regions, morphological correction processing is performed on the original location data according to the degree of interference, resulting in corrected location data. This step is one of the core innovations of this application, aiming to eliminate the influence of residues on morphological measurements. Morphological correction processing includes two main methods: adjusting values based on a preset physical compensation model, or performing geometric inference based on the location data of adjacent substrate regions. For example, when the interference level is low, a physical compensation model can be used. This model can be pre-established, for example, by experimentally measuring the effect of residues of different thicknesses on the laser optical path at different wavelengths, thereby establishing an optical path compensation lookup table or function. When a non-substrate region is detected, the thickness and optical properties of the residue are estimated based on its spectral characteristic parameters, and then the physical compensation model is used to perform optical path compensation on the original location data, adjusting its values to reflect the true substrate surface position. When the interference level is high, the residue may be too thick or uneven, making it difficult to accurately apply the physical compensation model. In this case, geometric inference can be used. For example, trend fitting is performed using location data of adjacent basal regions at the edge of non-basal regions. Through interpolation or extrapolation, the true basal topography within the non-basal regions is inferred and replaced with the original location data. The corrected location data includes the topographically corrected location data of the non-basal regions and the original location data of the basal regions.
[0038] In step A5, contour fitting is performed using the corrected position data to extract the geometric feature parameters of the groove. This step aims to accurately quantify the geometry of the groove from the corrected point cloud data. For example, the corrected position data can be denoised to eliminate measurement noise. Then, piecewise function fitting is performed according to a preset groove cross-sectional model (e.g., V-groove, U-groove, or composite groove). Piecewise function fitting can better adapt to the complex geometry of the groove, such as the bottom fillet and the groove edge. Based on the fitted piecewise function curve, the groove depth, groove width, and cross-sectional area, among other geometric feature parameters, can be accurately calculated.
[0039] In step A6, the remaining service life of the coating roller is calculated based on the geometric feature parameters and preset wear limits. This step is the final life prediction stage. The preset wear limits are determined based on the roller's design life, material properties, and coating quality requirements; for example, the groove depth cannot be lower than a certain threshold, and the groove width cannot exceed a certain range. By comparing the currently extracted geometric feature parameters with these wear limits, the wear state of the roller can be assessed. For example, the wear amount of each geometric feature parameter can be calculated, and combined with historical wear data and a wear rate model, the time required for the roller to reach its wear limit can be predicted, thereby calculating the remaining service life.
[0040] The method for detecting the groove morphology of coated rollers proposed in this application can effectively solve the limitations of traditional detection methods when dealing with residual films generated in the coating process by introducing multispectral information and intelligent morphology correction mechanism.
[0041] Specifically, in step A1, by acquiring the reflected signal intensity at the same location on the coating roller surface at at least two preset wavelengths, along with the original position data, this application can obtain richer optical information than single-wavelength laser scanning. Traditional methods rely solely on the distance information of a single-wavelength laser. When the laser beam encounters a residue film, its echo signal is distorted due to the optical properties of the residue, causing the original position data to deviate from the true morphology. This application, however, by acquiring multispectral reflected signal intensity, provides a spectroscopic basis for subsequently distinguishing between the substrate material and the residue. In steps A2 and A3, spectral feature parameters are extracted based on the reflected signal intensity and compared with a preset substrate material feature range, accurately identifying non-substrate areas affected by residue interference and determining the degree of interference. This is one of the core innovations of this application. Traditional methods struggle to distinguish between signal distortion caused by residue and true morphological changes, often requiring strong filtering, which simultaneously smooths out the true wear characteristics. This application analyzes the intensity of reflected signals at different wavelengths and utilizes the difference in spectral response between residues and substrate materials to accurately identify the presence and extent of residue influence, avoiding misjudgment of the true morphology data. In step A4, for locations identified as non-substrate areas, morphology correction processing is performed on the original location data according to the degree of interference, resulting in corrected location data. This step is crucial for solving residue interference in this application. Traditional methods either ignore residue interference, leading to errors, or over-filter, resulting in information loss. This application adopts different correction strategies based on the degree of interference: for mild interference, a physical compensation model is used to adjust the values, and the optical path error is accurately corrected by estimating the optical properties and thickness of the residue; for severe interference, geometric inference is used, utilizing reliable data from adjacent substrate areas to reconstruct the morphology of the damaged area. This hierarchical correction strategy ensures the accuracy of the correction while avoiding damage to the true morphological features, allowing the corrected location data to more accurately reflect the actual morphology of the rubber roller. In steps A5 and A6, contour fitting is performed using the corrected position data to extract the geometric feature parameters of the grooves. Based on these parameters and a preset wear limit, the remaining service life of the coating roller is calculated. Since the aforementioned steps effectively eliminate the influence of residues on the morphology data, this application can obtain more accurate groove geometric feature parameters. Compared to traditional methods where wear is underestimated due to residue interference, this application provides a more accurate wear assessment, making life prediction more reliable. This avoids unexpected defective products, production interruptions, and economic losses caused by misjudging the roller's lifespan.
[0042] In summary, this application overcomes the shortcomings of existing technologies in terms of accuracy and reliability for detecting groove morphology and predicting lifespan of coated rollers in complex industrial environments by introducing multispectral detection, intelligent residue identification, and a graded morphology correction mechanism. Its core innovation lies in its ability to effectively distinguish between residue interference and actual morphology changes, and to perform targeted data correction, thereby obtaining high-precision groove morphology data. This provides solid technical support for the precise maintenance and efficient production of coated rollers.
[0043] In some implementations, step A2 includes: The intensity ratio between reflected signals of different wavelengths is calculated, or the intensity normalized difference between reflected signals of different wavelengths is calculated, and these are used as the spectral characteristic parameters.
[0044] By calculating the ratio between the reflected signal intensities at different wavelengths, the influence of environmental factors such as fluctuations in light source intensity and differences in detector sensitivity on the measurement results can be effectively eliminated, thereby more accurately characterizing the inherent spectral properties of the material. For example, if the reflected signal intensities I(λ1) and I(λ2) at wavelengths λ1 and λ2 are obtained, the intensity ratio can be expressed as I(λ1) / I(λ2) or I(λ2) / I(λ1).
[0045] Alternatively, by calculating the intensity normalized difference between reflected signals at different wavelengths, the absorption or reflection characteristics of a material within a specific wavelength range can be highlighted. The intensity normalized difference is typically expressed as (I(λ1)-I(λ2)) / (I(λ1)+I(λ2)). This calculation method normalizes spectral characteristic parameters to a certain range, facilitating subsequent comparison and analysis. Both calculation methods aim to extract characteristic quantities sensitive to material type and state from multispectral reflectance information, thereby enabling subsequent identification of non-substrate regions affected by residue interference.
[0046] The proposed solution effectively distinguishes the substrate material of the coated roller from potential residues (such as ink, coating fragments, etc.) by calculating the ratio or normalized difference between the intensity of reflected signals at different wavelengths. This is because the optical response characteristics of the substrate material and residues often differ significantly at different wavelengths. For example, some residues may exhibit strong absorption or reflection at specific wavelengths, while the substrate material displays different spectral curves. By processing the ratio or normalized difference, these differences are amplified and transformed into easily identifiable spectral feature parameters, thus providing a reliable basis for identifying non-substrate regions in subsequent step A3. This processing method reduces the interference of external environmental noise on spectral feature extraction and improves the stability and accuracy of the feature parameters.
[0047] In some implementations, step A3 includes: A301. Calculate the deviation vector between the spectral characteristic parameters and the preset characteristic range of the substrate material; A302. Compare the magnitude of the deviation vector with a preset deviation threshold to identify whether the location belongs to a non-substrate region affected by residues; A303. If the location belongs to a non-substrate region affected by residues, the degree of interference at the location is divided into a first interference level or a second interference level based on the magnitude of the deviation vector; the deviation level of the first interference level is lower than that of the second interference level.
[0048] The preset substrate material characteristic range refers to the expected distribution range or reference value of the spectral characteristic parameters that the coating roller substrate material should possess under ideal conditions without residual interference. The deviation vector is a quantitative representation of the difference between the spectral characteristic parameters and the preset substrate material characteristic range; its direction indicates the direction of deviation, and its magnitude quantifies the degree of deviation. For example, when the spectral characteristic parameters are multi-dimensional vectors (such as when there are more than three preset wavelengths, the spectral characteristic parameters have three elements according to the permutation and combination results), the deviation vector can be a vector pointing from the current measurement point to the center or boundary of the substrate material characteristic range.
[0049] The magnitude of the deviation vector is used to measure the "distance" or degree of difference between the spectral characteristics of the current location and the standard spectral characteristics of the substrate material. A preset deviation threshold is a critical value used to determine whether the degree of deviation is sufficient to classify the region as non-substrate; this threshold can be set based on experimental data, statistical analysis, or experience.
[0050] When the location is identified as a non-base region disturbed by residue, the degree of disturbance can be further subdivided into a first disturbance level and a second disturbance level based on the magnitude of the deviation vector. The first disturbance level represents a milder disturbance, with a deviation degree lower than that represented by the more severe disturbance in the second disturbance level.
[0051] This application's solution introduces the concept of a deviation vector, transforming the simple comparison of spectral characteristic parameters with the characteristic range of the substrate material into a quantitative calculation of the degree of deviation. By calculating the magnitude of the deviation vector, the difference between the spectral characteristics at the current location and the standard spectral characteristics of the substrate material can be accurately measured, thus avoiding the ambiguity that may exist in traditional simple comparisons. Furthermore, by comparing the magnitude of the deviation vector with a preset deviation threshold, non-substrate regions affected by residue interference can be identified more objectively and accurately. In addition, the degree of interference is divided into a first interference level or a second interference level based on the magnitude of the deviation vector, making the assessment of residue interference more refined and providing a more targeted basis for subsequent morphology correction processing. For example, different interference levels may correspond to different types of residues or residues of different thicknesses, thus requiring different correction strategies.
[0052] In practical applications, relying solely on the degree of deviation of spectral feature parameters for judgment may result in insufficient accuracy in identifying areas of residual interference. This is especially true when the degree of deviation of spectral features is at a critical level, making it difficult to accurately distinguish between minor surface defects and actual residual interference, thus affecting the accuracy of subsequent morphology correction.
[0053] Therefore, in some preferred embodiments, in step A1, the degree of polarization of the reflected signal under a preset polarization state is also obtained; The preset deviation thresholds include a first deviation threshold and a second deviation threshold, wherein the first deviation threshold is less than the second deviation threshold. Step A302 includes: If the magnitude of the deviation vector is less than the first deviation threshold, then the location is determined to belong to the base region that is not affected by the residue. If the magnitude of the deviation vector is greater than the second deviation threshold, then the location is determined to belong to a non-base region affected by residue interference. If the magnitude of the deviation vector is greater than or equal to the first deviation threshold and less than or equal to the second deviation threshold, the degree of polarization is compared with the standard polarization range of the substrate material. If the degree of polarization is lower than the lower limit of the standard polarization range, the location is determined to belong to a non-substrate region affected by residues; otherwise, the location is determined to belong to a substrate region unaffected by residues.
[0054] The degree of polarization refers to the characteristic of the direction of vibration of the electric field vector of a light wave, which can reflect the microscopic structural information of the interaction between light and matter. When acquiring reflected signals, a polarizer can be placed at the laser transmitter end and / or an analyzer at the receiver end to obtain the intensity of the reflected signal under a specific polarization state, and then the degree of polarization of the reflected light can be calculated. Preset polarization states can include linear polarization, circular polarization, or elliptical polarization, and the specific selection can be optimized according to the optical properties of the material and residue under test.
[0055] The first and second deviation thresholds are two key boundaries used to define the degree of deviation of spectral characteristic parameters. The first deviation threshold defines the region where the spectral characteristic parameters are highly consistent with the characteristic range of the substrate material, i.e., a clearly defined substrate region. The second deviation threshold defines the region where the spectral characteristic parameters deviate significantly from the characteristic range of the substrate material, i.e., a clearly defined non-substrate region. A "middle region" is formed between these two thresholds. For the location within this region, it is difficult to make a clear judgment based solely on the spectral deviation magnitude; therefore, polarization degree is introduced as an auxiliary criterion. The standard polarization degree range of the substrate material refers to the range of polarization degrees that the reflected light from the surface of the coating roller substrate material should have under ideal conditions without residue interference. This range can be obtained through measurement and statistical analysis of standard substrate materials. When the magnitude of the deviation vector is between the first and second deviation thresholds, by comparing the measured polarization degree with the lower limit of this standard polarization degree range, it is possible to further distinguish whether the deviation is caused by residues or other factors. Typically, the presence of residues alters the polarization characteristics of the reflected light, leading to a decrease in polarization degree.
[0056] This application's solution introduces the degree of polarization of the reflected signal under a preset polarization state as an auxiliary judgment criterion, and sets a first deviation threshold and a second deviation threshold, effectively solving the problem of insufficient accuracy that may exist when relying solely on the degree of deviation of spectral feature parameters for identification. Specifically, when the degree of deviation of the spectral feature parameters is very small (less than the first deviation threshold) or very large (greater than the second deviation threshold), it can be quickly and reliably determined as a substrate region or a non-substrate region. However, when the degree of deviation is in an intermediate ambiguous region (greater than or equal to the first deviation threshold and less than or equal to the second deviation threshold), the spectral information is insufficient to make a clear judgment. At this time, the degree of polarization, as an independent and complementary optical property, can provide additional identification information. Since residues usually have different optical refractive indices, absorption characteristics, and surface roughness than the substrate material, these differences will affect the polarization state of the reflected light. For example, some residues may cause a depolarization effect in the reflected light, resulting in a reduction in the degree of polarization. Therefore, by comparing the degree of polarization with the standard polarization range of the substrate material, regions that have not significantly deviated from the spectral features but are actually affected by residues can be identified more accurately, thereby avoiding misjudgment or omission.
[0057] In some implementations, step A4 includes: A401. When the interference level is the first interference level, the step of adjusting the value according to the preset physical compensation model is executed, and the original position data is optically compensated using the estimated value of the refractive index and thickness of the residue at the corresponding wavelength. A402. When the interference level is the second interference level, the step of performing geometric inference based on the location data of adjacent base regions is executed, and trend fitting is performed using the base position coordinates of the edge of the non-base region to replace the original location data in the non-base region.
[0058] When the interference level is identified as Level 1, it typically means that the residue has a relatively small impact on the reflected signal, for example, the residue layer is thin or its optical properties are not significantly different from the substrate material. In this case, a physical compensation model can be used to adjust the original position data. This physical compensation model is based on optical principles and is used to estimate the impact of the residue layer on the optical path. The estimated refractive index and thickness of the residue at the corresponding wavelength are key parameters for optical path compensation. These parameters can be obtained from a pre-established multispectral feature database. This database records the standard reflectance distribution range of the substrate material at different wavelengths and the influence curves of residues of different thicknesses on the reflectance characteristics of the substrate material. By comparing the actually measured spectral feature parameters with the data in the database, the thickness of the residue can be estimated, and then combined with its refractive index (which can be pre-measured) for optical path compensation to eliminate the error caused by the residue on the position data.
[0059] Furthermore, when the interference level is identified as Level II, this typically indicates a significant impact of the residue on the reflected signal. For example, the residue layer may be thick, or its optical properties may differ significantly from the substrate material, making accurate compensation through physical models difficult or unreliable. In this case, a method based on geometric inference from the location data of adjacent substrate regions is more suitable. This method utilizes the substrate location coordinates at the edge of the non-substrate region for trend fitting. Specifically, it identifies the location data of undisturbed substrate regions surrounding the non-substrate region and, based on this reliable substrate location data, predicts or infers the true substrate morphology within the non-substrate region through interpolation, extrapolation, or other trend fitting algorithms. Thus, the original location data within the non-substrate region is replaced with data obtained through geometric inference, effectively eliminating morphology measurement errors caused by severe interference.
[0060] This application's solution addresses the limitations of single correction methods when handling varying degrees of residue interference by dynamically selecting a morphology correction strategy based on the level of interference. Specifically, when the interference is mild (Level 1), the impact of residue on the optical path can be quantified and compensated using a physical model. By using the estimated refractive index and thickness of the residue to compensate for the optical path deviation caused by the thin layer of residue, the optical path deviation can be accurately corrected, thus restoring a near-realistic surface morphology. When the interference is severe (Level 2), the residue can cause significant distortion of the reflection signal, making accurate compensation based on the physical model difficult. In this case, by using reliable substrate location data from the edges of non-substrate regions for geometric inference, the severely interfered area can be effectively bypassed, and the morphology of that area can be reconstructed using available information from the surrounding area. This hierarchical correction strategy ensures that the most suitable correction method is used under different interference conditions, thereby improving the accuracy and robustness of morphology detection.
[0061] In some implementations, step A5 includes: A501. The corrected position data is denoised, and piecewise function fitting is performed according to the preset groove cross-section model; A502. Based on the piecewise function curve obtained by fitting, calculate the groove depth, groove width, and cross-sectional area, which are used as the geometric feature parameters.
[0062] Specifically, denoising the corrected location data is necessary before contour fitting. Denoising aims to eliminate random noise or outliers that may have been introduced during data acquisition, ensuring the accuracy of subsequent fitting. For example, various signal processing methods such as moving average, median filtering, and Gaussian filtering can be used to smooth the location data and reduce data fluctuations.
[0063] Subsequently, piecewise function fitting is performed according to the preset groove cross-section model. The preset groove cross-section model refers to a mathematical model established based on the typical geometry of the groove on the coating roller (e.g., V-shape, U-shape, rectangle, or composite shape). Piecewise function fitting involves dividing the entire contour of the groove into several sub-regions, each of which is fitted with a different function form to more accurately describe the complex geometry of the groove. For example, the bottom region of the groove may be fitted with a flat function or a circular arc function, while the sidewall region may be fitted with a straight line function or a curve function with varying slope. Through this piecewise fitting, the local features of the groove can be better captured, avoiding the bias that may be introduced by fitting a single function.
[0064] Based on the piecewise function curve obtained from the fitting, the geometric characteristic parameters of the groove can be accurately calculated. Among these, the groove depth refers to the vertical distance from the bottom of the groove to the highest point on the roller surface; the groove width refers to the horizontal width at the groove opening or at a specific depth; and the cross-sectional area refers to the cross-sectional area of the groove in the direction perpendicular to the roller axis. These geometric characteristic parameters are key indicators for evaluating the wear condition and performance of the coating roller.
[0065] The proposed solution effectively filters out measurement noise and ensures data purity by denoising the corrected position data. Based on this, a pre-defined groove cross-section model is used for piecewise function fitting. This fully utilizes prior knowledge of the groove geometry, decomposing the complex groove contour into more manageable local segments, and selectively choosing appropriate functions for fitting. This piecewise fitting strategy avoids errors that may result from single global fitting, especially when the groove shape is irregular or contains sharp corners, significantly improving the accuracy and robustness of the fitting. Therefore, the groove depth, groove width, and cross-sectional area, etc., calculated from the high-precision fitting curve, more realistically reflect the actual morphology of the coating roller groove, providing reliable basic data for subsequent wear assessment and life calculation.
[0066] The morphology correction process in step A4 includes adjusting values according to a preset physical compensation model or performing geometric inference based on the positional data of adjacent base regions. When using a physical compensation model for correction, the result may depend on the estimation of residual parameters, thus introducing a certain degree of uncertainty or error. If, in subsequent fitting processes, these data points corrected by the physical compensation model are treated the same as the original base region data or geometrically inferred data, potential estimation errors may propagate, thereby affecting the accuracy of contour fitting and the reliability of the extracted geometric feature parameters.
[0067] Therefore, in some preferred embodiments, in step A501, for positions where the morphology is corrected by adjusting the values according to a preset physical compensation model, the weight of the corresponding position data in the fitting process is reduced.
[0068] Specifically, reducing the weight of corresponding location data in the fitting process means assigning a relatively small weight coefficient to these location data corrected by the physical compensation model when performing contour fitting. For example, when using the least squares method for curve fitting, the sum of squared residuals is usually minimized. By introducing weight coefficients for different data points (used to weight the squared residuals of corresponding data points when calculating the sum of squared residuals), the fitting process can focus more on data points considered more reliable, while giving less influence to data points with lower reliability. These weight coefficients can be dynamically adjusted based on the estimation error of the physical compensation model, the degree of residual interference, the magnitude of the deviation vector, or other relevant parameters to ensure the accuracy and stability of the fitting results.
[0069] The solution proposed in this application effectively addresses the aforementioned problem by reducing the weight of positional data adjusted by the physical compensation model during the contour fitting process. When positional data is corrected using the physical compensation model, the correction result may contain certain uncertainties or errors because the model relies on estimations of residual parameters (such as refractive index and thickness). If fitting is performed indiscriminately, these potential errors may unnecessarily perturb the final fitted curve, thereby affecting the accurate extraction of groove geometric feature parameters. By reducing the weight of these data, the fitting algorithm, when seeking the optimal fitted curve, reduces its reliance on these potentially erroneous data points and instead refers more to more reliable data points (e.g., original base region data or data obtained through geometric inference). This allows the fitting process to better resist potential errors introduced by the physical compensation model, resulting in a smoother and more accurate groove contour.
[0070] In some implementations, step A6 includes: A601. Obtain the remaining wear margin based on the geometric feature parameters and the preset wear limit; A602. Obtain the wear rate based on the changing trend of the geometric feature parameters; A603. Calculate the remaining service life of the coating roller based on the remaining wear margin and the wear rate; A604. When the remaining service life is lower than the preset warning threshold, output maintenance or replacement suggestions.
[0071] Specifically, in step A601, the geometric feature parameters refer to the geometric features of the groove extracted by the contour fitting module, such as groove depth, groove width, and cross-sectional area. The preset wear limit refers to the maximum degree of wear or minimum size limit allowed by each geometric feature parameter of the coating roller under normal working conditions. The remaining wear margin can be understood as the difference between the current geometric feature parameters and the preset wear limit, representing the amount of wear that each geometric feature parameter can still withstand before the coating roller reaches the scrap standard.
[0072] In step A602, the trend of change of the geometric feature parameters refers to predicting their future trend by analyzing historical data of the geometric feature parameters obtained at different detection time points of the coating roller, such as through linear regression or exponential fitting. The wear rate refers to the amount of wear of each geometric feature parameter per unit time, reflecting the speed of wear of the coating roller.
[0073] In practical applications, step A603 calculates the remaining service life of the coating roller by dividing the remaining wear margin by the wear rate, which provides a dynamic prediction based on the current wear condition and wear trend.
[0074] Furthermore, in step A604, the warning threshold is a pre-set time or wear limit. When the calculated remaining service life is lower than the threshold, the system will automatically output maintenance or replacement suggestions to remind operators to take timely measures to avoid unexpected equipment downtime or product quality damage.
[0075] This application's solution addresses the shortcomings of traditional methods in lifespan assessment, namely insufficient accuracy and lack of timely warnings, by refining the calculation of remaining service life into obtaining remaining wear margin, obtaining wear rate, and performing calculations based on both, and by introducing an early warning mechanism. Specifically, step A601 provides foundational data for subsequent lifespan calculations by quantifying the gap between the current state and the scrapping standard. Step A602 dynamically captures the wear trend of the coating roller by analyzing historical data, making lifespan prediction no longer static but reflecting the dynamic nature of the actual wear process. Step A603 combines these two key parameters, making the calculated remaining service life more predictable and reliable. Finally, the early warning mechanism in step A604 ensures that timely alerts are issued before the coating roller reaches the critical wear point, thus providing a valuable time window for maintenance decisions.
[0076] In some implementations, step A601 includes: Based on the initial and actual values of each geometric characteristic parameter, calculate the wear amount of each geometric characteristic parameter respectively; Based on the wear amount of each geometric feature parameter and the corresponding preset wear limit, calculate the remaining wear margin of each geometric feature parameter; In step A602, the wear rate of each geometric feature parameter is obtained based on the changing trend of each geometric feature parameter. Step A603 includes: The initial remaining life of the coating roller is calculated based on the remaining wear margin and wear rate corresponding to each geometric characteristic parameter. The minimum initial remaining lifespan is selected as the final remaining lifespan of the coating roller.
[0077] The "various geometric characteristic parameters" can be understood as key dimensions or shape indicators describing the groove morphology of the coating roller, such as groove depth, groove width, and cross-sectional area. The "initial values" of these parameters refer to the standard values when the coating roller is brand new or just put into use, while the "actual values" are the current measured values extracted by the aforementioned contour fitting module. By comparing the initial values and actual values, the "wear amount of each geometric characteristic parameter" can be calculated separately, i.e., the actual wear degree of each parameter.
[0078] "Preset wear limit" refers to the maximum allowable wear amount or minimum allowable size set for each geometric feature parameter. Once this limit is reached, the coating roller is considered to have reached its service life. Based on the wear amount and the preset wear limit, the "remaining wear margin for each geometric feature parameter" can be calculated. This margin represents how much usable wear space each parameter has before reaching its wear limit.
[0079] The "changing trends of various geometric characteristic parameters" can be obtained by analyzing historical detection data, such as through linear regression, exponential fitting, or other time series analysis methods to predict the future trajectory of each parameter. Therefore, "obtaining the wear rate of each geometric characteristic parameter" refers to the amount of wear per parameter per unit time.
[0080] The initial remaining life is calculated independently for each geometric feature parameter, by dividing the remaining wear margin of that parameter by its wear rate. Since different geometric features parameters may have different wear rates and wear margins, multiple initial remaining lives will be obtained. To ensure the safe and reliable operation of the coating roller, this application selects the minimum initial remaining life as the final remaining service life of the coating roller. This means that once any critical geometric feature parameter is expected to reach its wear limit first, that point in time is considered the end of the entire coating roller's lifespan.
[0081] This application's solution addresses the limitations of a single overall lifespan assessment by refining the calculation of the remaining service life of the coating roller to individual geometric characteristic parameters and ultimately selecting the shortest initial remaining service life as the final remaining service life. Specifically, in step A601, by calculating the wear amount and remaining wear margin of each geometric characteristic parameter separately, the assessment of the coating roller's wear state becomes more refined, accurately reflecting the actual wear condition of each critical dimension. In step A602, by obtaining the wear rate of each geometric characteristic parameter separately, the prediction of the future wear trend of each parameter is ensured to be more accurate, avoiding prediction deviations caused by differences in the wear characteristics of different parameters. Finally, in step A603, by calculating the initial remaining service life of each parameter and selecting the minimum value as the final remaining service life, this solution can capture the most vulnerable or fastest-wearing link of the coating roller, thus providing a more conservative, safe, and realistic lifespan prediction. This method ensures that the coating roller will not fail unexpectedly due to premature wear of a single critical parameter, improving the reliability of the prediction.
[0082] refer to Figure 2 This application provides a system for detecting the groove morphology of a coating roller, the system comprising: Data acquisition module 1 is used to acquire the intensity of reflected signals at the same position on the surface of the coating roller under at least two preset wavelengths and the original position data (for details, please refer to step A1 above). Feature extraction module 2 is used to extract spectral feature parameters of the corresponding position based on the intensity of the reflected signal (for details, please refer to step A2 above). The identification module 3 is used to compare the spectral feature parameters with the preset substrate material feature range, identify whether the location belongs to the non-substrate region affected by residue interference, and determine the degree of interference when it belongs to the non-substrate region (the specific process can be referred to step A3 above). The topography correction module 4 is used to perform topography correction processing on the original location data for locations identified as non-base regions according to the degree of interference, so as to obtain corrected location data; wherein, the topography correction processing includes adjusting values according to a preset physical compensation model, or performing geometric inference based on the location data of adjacent base regions (the specific process can be referred to step A4 above). Contour fitting module 5 is used to perform contour fitting using the corrected position data and extract the geometric feature parameters of the groove (for details, please refer to step A5 above). The life calculation module 6 is used to calculate the remaining service life of the coating roller based on the geometric feature parameters and the preset wear limit (for details, please refer to step A6 above).
[0083] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.
Claims
1. A method for detecting the groove morphology of a coating roller, characterized in that, The method includes the following steps: A1. Obtain the reflected signal intensity and original position data at the same location on the surface of the coating roller under at least two preset wavelengths; A2. Extract the spectral feature parameters of the corresponding position based on the intensity of the reflected signal; A3. Compare the spectral characteristic parameters with the preset substrate material characteristic range to identify whether the location belongs to a non-substrate region affected by residues, and determine the degree of interference if it belongs to a non-substrate region; A4. For locations identified as non-base regions, perform shape correction processing on the original location data according to the degree of interference to obtain corrected location data; wherein, the shape correction processing includes adjusting values according to a preset physical compensation model, or performing geometric inference based on the location data of adjacent base regions; A5. Use the corrected position data to perform contour fitting and extract the geometric feature parameters of the groove; A6. Calculate the remaining service life of the coating roller based on the geometric feature parameters and the preset wear limit.
2. The method for detecting the groove morphology of a coating roller according to claim 1, characterized in that, Step A2 includes: The intensity ratio between reflected signals of different wavelengths is calculated, or the intensity normalized difference between reflected signals of different wavelengths is calculated, and these are used as the spectral characteristic parameters.
3. The method for detecting the groove morphology of a coating roller according to claim 1, characterized in that, Step A3 includes: A301. Calculate the deviation vector between the spectral characteristic parameters and the preset characteristic range of the substrate material; A302. Compare the magnitude of the deviation vector with a preset deviation threshold to identify whether the location belongs to a non-substrate region affected by residues; A303. If the location belongs to a non-substrate region affected by residues, the degree of interference at the location is divided into a first interference level or a second interference level based on the magnitude of the deviation vector; the deviation level of the first interference level is lower than that of the second interference level.
4. The method for detecting the groove morphology of a coating roller according to claim 3, characterized in that, In step A1, the degree of polarization of the reflected signal under a preset polarization state is also obtained; The preset deviation thresholds include a first deviation threshold and a second deviation threshold, wherein the first deviation threshold is less than the second deviation threshold. Step A302 includes: If the magnitude of the deviation vector is less than the first deviation threshold, then the location is determined to belong to the base region that is not affected by the residue. If the magnitude of the deviation vector is greater than the second deviation threshold, then the location is determined to belong to a non-base region affected by residue interference. If the magnitude of the deviation vector is greater than or equal to the first deviation threshold and less than or equal to the second deviation threshold, the degree of polarization is compared with the standard polarization range of the substrate material. If the degree of polarization is lower than the lower limit of the standard polarization range, the location is determined to belong to a non-substrate region affected by residues; otherwise, the location is determined to belong to a substrate region unaffected by residues.
5. The method for detecting the groove morphology of a coating roller according to claim 3, characterized in that, Step A4 includes: A401. When the interference level is the first interference level, the step of adjusting the value according to the preset physical compensation model is executed, and the original position data is optically compensated using the estimated value of the refractive index and thickness of the residue at the corresponding wavelength. A402. When the interference level is the second interference level, the step of performing geometric inference based on the location data of adjacent base regions is executed, and trend fitting is performed using the base position coordinates of the edge of the non-base region to replace the original location data in the non-base region.
6. The method for detecting the groove morphology of a coating roller according to claim 1, characterized in that, Step A5 includes: A501. The corrected position data is denoised, and piecewise function fitting is performed according to the preset groove cross-section model; A502. Based on the piecewise function curve obtained by fitting, calculate the groove depth, groove width, and cross-sectional area, which are used as the geometric feature parameters.
7. The method for detecting the groove morphology of a coating roller according to claim 6, characterized in that, In step A501, for positions where the shape is corrected by adjusting the values according to a preset physical compensation model, the weight of the corresponding position data in the fitting process is reduced.
8. The method for detecting the groove morphology of a coating roller according to claim 1, characterized in that, Step A6 includes: A601. Obtain the remaining wear margin based on the geometric feature parameters and the preset wear limit; A602. Obtain the wear rate based on the changing trend of the geometric feature parameters; A603. Calculate the remaining service life of the coating roller based on the remaining wear margin and the wear rate; A604. When the remaining service life is lower than the preset warning threshold, output maintenance or replacement recommendations.
9. The method for detecting the groove morphology of a coating roller according to claim 8, characterized in that, Step A601 includes: Based on the initial and actual values of each geometric characteristic parameter, calculate the wear amount of each geometric characteristic parameter respectively; Based on the wear amount of each geometric feature parameter and the corresponding preset wear limit, calculate the remaining wear margin of each geometric feature parameter; In step A602, the wear rate of each geometric feature parameter is obtained based on the changing trend of each geometric feature parameter. Step A603 includes: The initial remaining life of the coating roller is calculated based on the remaining wear margin and wear rate corresponding to each geometric characteristic parameter. The minimum initial remaining lifespan is selected as the final remaining lifespan of the coating roller.
10. A detection system for the groove morphology of a coating roller, characterized in that, The system includes: The data acquisition module is used to acquire the intensity of reflected signals at the same location on the surface of the coating roller under at least two preset wavelengths, as well as the original position data. The feature extraction module is used to extract spectral feature parameters at the corresponding position based on the intensity of the reflected signal; The identification module is used to compare the spectral feature parameters with a preset substrate material feature range, identify whether the location belongs to a non-substrate region affected by residue interference, and determine the degree of interference when it belongs to a non-substrate region; The topography correction module is used to perform topography correction processing on the original location data for locations identified as non-base regions, based on the degree of interference, to obtain corrected location data; wherein, the topography correction processing includes adjusting values according to a preset physical compensation model, or performing geometric inference based on the location data of adjacent base regions; The contour fitting module is used to perform contour fitting using the corrected position data and extract the geometric feature parameters of the groove. The life calculation module is used to calculate the remaining service life of the coating roller based on the geometric feature parameters and the preset wear limit.