Grading detection method and system for gradient color car cover film

Abstract

The invention provides a graded detection method and system for a gradient color vehicle cover film, and relates to the technical field of vehicle cover films, and the method comprises the steps: obtaining the spectral reflectivity data of a to-be-detected vehicle cover film in a visible light wave band, and extracting the chromaticity gradient distribution characteristics; configuring each dynamic change vector; dividing N detection grids, and synchronously measuring thickness uniformity data and surface roughness of each detection grid; performing multi-modal fusion on each dynamic change vector and the thickness uniformity data; and on the basis of the coupling coefficient matrix and the range value of the surface roughness, evaluating the comprehensive quality index of the to-be-detected vehicle cover film, performing dynamic grading by using a grading threshold interval, and drawing up defect mapping coordinates under the grade identification of the vehicle cover film. According to the method and the device, the technical problem of inaccurate quality evaluation or missing detection caused by lack of dynamic and accurate quality grading is solved, the film quality is dynamically graded through the grading threshold interval, the defect of the gradient color car cover film is quickly positioned, and the quality evaluation precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of car film, and particularly relates to a grading detection method and system for gradient color car film. BACKGROUND

[0002] Gradient color car film is a protective film for vehicles with a color gradient effect, usually used to enhance the aesthetic appearance of the vehicle body. Its characteristic is that the film surface presents a smooth transition from one color to another color. However, the traditional detection method of gradient color car film mostly relies on visual inspection or simple color difference measurement, which is difficult to comprehensively evaluate all quality indicators of the film surface. Especially in the detection of gradient color film, since the color is continuously changing, simply relying on color difference measurement is difficult to accurately detect subtle color differences or uneven gradient effects.

[0003] In addition, the detection of physical properties of gradient color car film (such as film thickness, surface roughness, etc.) usually focuses on a single feature, ignoring the mutual relationship between different physical properties. The traditional quality evaluation method often takes whether it is qualified as the only judgment standard, and cannot accurately grade the quality of the film, making it difficult to achieve detailed quality evaluation, resulting in missed detection or misjudgment of defects, and thus affecting the precision and accuracy of quality detection.

[0004] In summary, the existing technology has the technical problem of inaccurate quality evaluation or missed detection due to the reliance on a single physical property or visual inspection, lack of dynamic and accurate quality grading. SUMMARY

[0005] The purpose of the present application is to provide a grading detection method and system for gradient color car film, which solves the technical problem of inaccurate quality evaluation or missed detection due to the reliance on a single physical property or visual inspection, lack of dynamic and accurate quality grading in the prior art.

[0006] In view of the above problems, the present application provides a grading detection method and system for gradient color car film.

[0007] In a first aspect, the application provides a graded color car paint film grading detection method, which is realized by a graded color car paint film grading detection system. The graded color car paint film grading detection method comprises the following steps: obtaining spectral reflectance data of a to-be-tested car paint film in a visible light band, and extracting a color gradient distribution feature; based on the color gradient distribution feature, configuring a first dynamic change vector obtained by combining hue and saturation, a second dynamic change vector obtained by combining lightness and hue, and a third dynamic change vector obtained by combining lightness and saturation; according to the to-be-tested car paint film, dividing N detection grids, and synchronously measuring thickness uniformity data and surface roughness of each detection grid; performing multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector, and the thickness uniformity data to generate a coupling coefficient matrix; based on the coupling coefficient matrix and a range value of the surface roughness, evaluating a comprehensive quality index of the to-be-tested car paint film, dynamically grading the comprehensive quality index using a grading threshold interval, and formulating a defect mapping coordinate under a car paint film grade mark.

[0008] Optionally, the to-be-tested car paint film is scanned to collect reflection spectrum data in a preset wavelength range, the reflection spectrum data is subjected to baseline correction processing to extract a reflectance feature value; based on the reflectance feature value, a continuous spectrum reflectance curve is reconstructed to determine a gradient change rate of adjacent wavelength reflectance; according to a ratio of a difference value of the gradient change rate of adjacent wavelength reflectance and a wavelength difference value, the color gradient distribution feature is determined.

[0009] Optionally, the reconstructed continuous spectrum reflectance curve is converted to generate a hue angle distribution thermogram; an abnormal mutation region in the hue angle distribution thermogram is identified and marked as a potential color difference defect; based on the potential color difference defect, a color gradient smoothness is quantified according to a hue angle standard deviation and a saturation gradient direction consistency index, and the color gradient smoothness is used to configure a direction component of the first dynamic change vector.

[0010] Optionally, a dual-wavelength laser interferometer is configured, wherein a first wavelength is used to measure a substrate thickness, and a second wavelength is used to detect a surface coating thickness; the dual-wavelength laser interferometer is used to determine an interference fringe shift amount of each detection grid corresponding to the N detection grids, to obtain a thickness difference distribution of the substrate and the coating; based on the thickness difference distribution of the substrate and the coating, a multi-scale correlation analysis is performed on the color gradient smoothness to configure a coupling degree index associated with the first dynamic change vector.

[0011] Optionally, for each detection grid, a local linear regression function of the color gradient smoothness and the thickness difference distribution of the substrate and the coating is established; a multi-scale correlation coefficient r corresponding to each detection grid is obtained by using the local linear regression function of the color gradient smoothness and the thickness difference distribution of the substrate and the coating for multi-scale correlation analysis; and a coupling degree index associated with the first dynamic change vector is configured according to the multi-scale correlation coefficient r.

[0012] Optionally, the spatial overlap degree of the abnormal mutation area and the thickness abnormal area is analyzed by using a sliding window to configure an overlap degree index; an accumulated variance contribution rate is determined based on the multi-scale correlation coefficient r and the overlap degree index; and a coupling degree index associated with the first dynamic change vector is established based on the direction component of the first dynamic change vector, in combination with the multi-scale correlation coefficient r and the accumulated variance contribution rate.

[0013] Optionally, singular value decomposition is performed on the coupling coefficient matrix to determine a first quality index corresponding to the maximum singular value; and a weighted sum of the first quality index and a logarithmic transformation of the range value of the surface roughness is obtained, wherein a weight coefficient is dynamically adjusted according to the application scene of the car film.

[0014] Optionally, when a scratch and / or a bubble is detected on the surface, a defect area ratio is introduced as a penalty factor; and the comprehensive quality index is corrected by using the penalty factor.

[0015] Optionally, a color compensation repair area is located according to the defect mapping coordinates under the car film grade identification; a local color compensation robot is started, and a micro-fluidic nozzle of the local color compensation robot is controlled for color compensation operation in combination with the color compensation repair area; and at the same time, a color gradient consistency test is performed on the color compensation repair area during the color compensation operation.

[0016] In a second aspect, the application also provides a graded color car paint film grading detection system for performing the graded color car paint film grading detection method of the first aspect, wherein the graded color car paint film grading detection system comprises: a feature extraction module configured to obtain spectral reflectance data of a to-be-tested car paint film in a visible light band and extract a chroma gradient distribution feature; a change vector configuration module configured to configure, based on the chroma gradient distribution feature, a first dynamic change vector obtained by combining hue and saturation, a second dynamic change vector obtained by combining lightness and hue, and a third dynamic change vector obtained by combining lightness and saturation; a detection network division module configured to divide N detection grids according to the to-be-tested car paint film and synchronously measure thickness uniformity data and surface roughness of each detection grid; a multi-modal fusion module configured to perform multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector, and the thickness uniformity data to generate a coupling coefficient matrix; and a dynamic grading module configured to evaluate a comprehensive quality index of the to-be-tested car paint film based on the coupling coefficient matrix and a range value of the surface roughness, perform dynamic grading on the comprehensive quality index using a grading threshold interval, and formulate defect mapping coordinates under a car paint film grade mark.

[0017] The one or more technical solutions provided in the application have at least the following beneficial effects: By obtaining spectral reflectance data of a to-be-tested car paint film in a visible light band and extracting a chroma gradient distribution feature, a first dynamic change vector obtained by combining hue and saturation, a second dynamic change vector obtained by combining lightness and hue, and a third dynamic change vector obtained by combining lightness and saturation are configured based on the chroma gradient distribution feature, N detection grids are divided according to the to-be-tested car paint film, thickness uniformity data and surface roughness of each detection grid are synchronously measured, multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector, and the thickness uniformity data is performed to generate a coupling coefficient matrix, and a comprehensive quality index of the to-be-tested car paint film is evaluated based on the coupling coefficient matrix and a range value of the surface roughness, dynamic grading is performed on the comprehensive quality index using a grading threshold interval, and defect mapping coordinates under a car paint film grade mark are formulated. That is, by obtaining spectral reflectance data of a car paint film in a visible light band and extracting a chroma gradient distribution feature, a comprehensive quality index of the car paint film is comprehensively evaluated, dynamic grading of the car paint film quality is realized, defect mapping coordinates under a grade mark are accurately formulated, defects of a graded color car paint film are quickly located, and the accuracy of quality evaluation is improved.

[0018] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the specific embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, characteristics and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0020] Figure 1 The flowchart of the grading detection method of the gradual color car paint film of the present application.

[0021] Figure 2 The structural schematic diagram of the grading detection system of the gradual color car paint film of the present application.

[0022] Explanation of reference signs: feature extraction module 11, change vector configuration module 12, detection network division module 13, multi-modal fusion module 14, dynamic grading module 15. DETAILED DESCRIPTION

[0023] The present application provides a gradual color car paint film grading detection method and system, which solves the technical problem that the quality evaluation is inaccurate or missed in the prior art due to the lack of dynamic and accurate quality grading caused by the excessive reliance on single physical characteristics or visual inspection. By obtaining the spectral reflectance data of the car paint film in the visible light band and extracting the chroma gradient distribution characteristics, the comprehensive quality index of the car paint film is comprehensively evaluated, the dynamic grading of the car paint film quality is realized, the defect mapping coordinates under the grade mark are accurately drafted, and then the defects of the gradual color car paint film are quickly located, and the precision of quality evaluation is improved.

[0024] The technical solutions in the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described here. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application. In addition, it should be noted that only parts related to the present application are shown in the drawings for convenience of description, not all.

[0025] Embodiment one, please refer to the attached Figure 1 The application provides a kind of gradient color car paint film grading detection method, wherein the gradient color car paint film grading detection method is executed by a kind of gradient color car paint film grading detection system, and the gradient color car paint film grading detection method specifically includes the following steps: S100: obtaining the spectral reflectance data of the car paint film to be tested in the visible light band, and extracting the chroma gradient distribution characteristics.

[0026] Further, S100 of the application includes: The car paint film to be tested is scanned, the reflectance spectrum data in the preset wavelength range is collected, the reflectance spectrum data is subjected to baseline correction processing, and the reflectance characteristic value is extracted; based on the reflectance characteristic value, the continuous spectral reflectance curve is reconstructed, the gradient change rate of adjacent wavelength reflectance is determined; according to the ratio of the difference value of the gradient change rate of adjacent wavelength reflectance and the wavelength difference value, the chroma gradient distribution characteristics are determined.

[0027] Specifically, the car paint film to be tested is scanned, and the reflectance spectrum data in the preset wavelength range is collected by a spectral scanning device, which contains the reflection intensity of light of different wavelengths. The preset wavelength range is usually determined based on the visible light band, such as the electromagnetic wave band between about 400 nanometers and 700 nanometers. Since the reflectance spectrum data may be affected by instrument or environmental factors, baseline correction is usually required to eliminate these unnecessary effects. The correction process removes the effects of uneven instrument response, environmental interference and other factors, so that the reflectance spectrum data can more accurately reflect the optical properties of the car paint film itself. Baseline correction is a data preprocessing technique used to eliminate systematic errors introduced during instrument or measurement process, to ensure that the reflectance spectrum data accurately reflects the optical properties of the film itself.

[0028] Based on the data after baseline correction, the reflectance characteristic value is extracted, including the reflectance value at a specific wavelength, which represents the reflection characteristics of the car paint film at each wavelength band. The reflectance characteristic value refers to the numerical value of some key points (such as the reflectance value at some wavelengths) in the reflectance spectrum, which represents the reflection ability of the film surface to light of a specific wavelength, and is the basis for analyzing the optical properties of the film. Using the reflectance characteristic value, the continuous spectral reflectance curve is reconstructed to show the reflection behavior of the car paint film to light of each wavelength, indicating the reflectance change of the car paint film in the entire preset wavelength range, and directly determining the color distribution and gradient characteristics of the car paint film.

[0029] Based on the continuous spectral reflectance curve, the gradient change rate of reflectance between adjacent wavelengths is calculated, which is realized by calculating the difference of reflectance between each two adjacent wavelength points. The gradient change rate represents the speed of reflectance change between two adjacent wavelengths. High gradient change rate generally represents that the optical properties of the film change greatly, while low gradient change rate represents that the reflectance properties of the film are more uniform. For example, the reflectance changes from 0.55 to 0.60 at wavelengths of 450 nm to 500 nm, and the gradient change is small; the reflectance changes from 0.58 to 0.62 at wavelengths of 550 nm to 600 nm, and the gradient change is slightly larger.

[0030] By calculating the ratio of the difference of the gradient change rate of reflectance of adjacent wavelengths to the difference of wavelengths, the chroma gradient distribution feature is extracted, which describes the smoothness or gradient effect of the color change of the car paint film. For example, from 0.55 to 0.60, the gradient is (0.60-0.55) / (500-450)=0.05 / 50=0.001, wherein the change rate has no unit, the wavelength unit is nanometer, and the result is a unitless value. The chroma gradient distribution feature is a feature for describing the color change of the car paint film, which is determined based on the ratio of the gradient change rate of reflectance of adjacent wavelengths to the difference of wavelengths. By using the spectral reflectance data and the gradient change rate, the chroma gradient distribution feature of the car paint film is accurately extracted, which helps to deeply understand the color change of the car paint film.

[0031] S200: Based on the chroma gradient distribution feature, a first dynamic change vector obtained by combining hue and saturation, a second dynamic change vector obtained by combining lightness and hue, and a third dynamic change vector obtained by combining lightness and saturation are configured.

[0032] Further, S200 of the present application includes: The reconstructed continuous spectral reflectance curve is converted to generate a hue angle distribution heat map; abnormal mutation regions in the hue angle distribution heat map are identified and marked as potential color difference defects; based on the potential color difference defects, the chroma gradient smoothness is quantified according to the hue angle standard deviation and the saturation gradient direction consistency index, and the chroma gradient smoothness is used to configure the direction component of the first dynamic change vector.

[0033] Specifically, according to the distribution characteristics of the chroma gradient, information about the color change on the surface of the car paint film is obtained, which describes the color change from one area to another, such as the change of hue, saturation and lightness. Through hue, saturation and lightness, it is evaluated whether the transition of color is smooth, and potential defect areas are further identified. Hue is one of the basic attributes of color, representing the type or name of color, such as red, blue, etc., which is usually represented by the angle of the color wheel, and is expressed in degrees in the HSV or HSL color model (e.g. 0° represents red, 120° represents green, and 240° represents blue). Saturation describes the purity or intensity of color, the higher the saturation, the brighter the color; low saturation makes the color more gray. Lightness refers to the brightness or light intensity of color, which generally describes the transition of color from black (low lightness) to white (high lightness).

[0034] According to the changes of hue and saturation, a first dynamic change vector is configured, the direction of which represents the trend of color change, and the size represents the intensity of color change. The reconstructed continuous spectral reflectance is converted into a curve, the reflectance data is converted into hue angle, and a hue angle distribution heat map is drawn according to the reflectance characteristics at different wavelengths. The hue angle data reflects the color type (hue) at each wavelength, and using a color model such as HSV (hue, saturation, lightness model) or HSL (hue, saturation, brightness model), based on the reflectance data at each wavelength, the corresponding hue angle value can be calculated.

[0035] Using the heat map tool, the hue angle data is presented in the form of a heat map, and the colors in the map represent different hue angles, which can intuitively show the distribution of the color on the surface of the car paint film, such as which area has a larger color change. In the generated hue angle distribution heat map, areas with sharp color changes are identified, which are usually represented by sudden jumps in hue angle. For example, the hue angle of some areas suddenly jumps from one color value to another extreme value, and such mutations usually indicate material non-uniformity or defect areas. When the hue angle mutation area is identified, these areas are marked as potential color difference defects, which are caused by production non-uniformity, raw material problems or unstable surface quality. For example, assuming that the hue angle of a certain area in the heat map jumps sharply from yellow (60°) to purple (270°), it indicates that the color of this area has changed abnormally, and there may be a color difference defect.

[0036] For the potential color difference defect area, the hue angle standard deviation of the area is calculated, reflecting the dispersion degree of the hue angle value. The larger the standard deviation, the more uneven the color change of the area. By calculating the gradient direction consistency of the saturation, the smoothness of the color change is judged. If the saturation gradient is consistent in direction within a certain area, it means that the color change is relatively smooth. If the direction changes suddenly, it means that the color transition in the area is not smooth, indicating the presence of defects. According to the hue angle standard deviation and the saturation gradient direction consistency index, the color gradual change smoothness of the area is quantified. The higher the color gradual change smoothness value, the smoother the color transition, and vice versa, indicating that there may be color difference defects. The color gradual change smoothness is used to quantify the smoothness of the color change. The smoother the gradual change, the higher the value of the color gradual change smoothness. According to the quantified color gradual change smoothness, the direction component of the first dynamic change vector is configured. The higher the color gradual change smoothness, the more consistent the direction of the first dynamic change vector, indicating that the color change is smoother.

[0037] Similarly, according to the changes in brightness and hue, the changes in brightness and type of color with space or time are analyzed to generate a second dynamic change vector. The direction of the vector represents the trend of the brightness change of the color, and the size represents the intensity of the brightness and hue change. According to the changes in brightness and saturation, the changes in brightness and purity of the color are analyzed. The brightness and saturation values in the reflectivity data are extracted, the changes in brightness and purity of the color on the car paint film surface are analyzed, and a third dynamic change vector is generated. The direction of the vector represents the trend of the brightness and purity change, and the size represents the intensity of the brightness and purity change. Through the multi-dimensional analysis of the dynamic change vector on the color of the car paint film surface, the fine detection of the color change is provided, and the color unevenness and potential defect area are effectively detected.

[0038] S300: According to the car paint film to be measured, N detection grids are divided, and the thickness uniformity data and surface roughness of each detection grid are measured synchronously.

[0039] Specifically, the car paint film to be measured is divided into N detection grids, and the surface of the car paint film is divided into multiple small areas (grids) for detecting the performance indicators (such as thickness uniformity, surface roughness, etc.) of these areas one by one. The number and size of the grids can be set according to the size of the car paint film and the required detection accuracy. Generally speaking, the smaller the grid, the more detailed the detection data obtained, but the more the calculation and analysis workload.

[0040] For each grid area, the thickness is measured by precise instruments such as laser thickness gauge, ultrasonic thickness gauge, etc., and recorded. The thickness of each grid is measured at multiple points to obtain the thickness distribution of the grid. By comparing the thickness of each measurement point, the thickness uniformity of each grid can be calculated. If the thickness of a certain grid changes greatly, it indicates that there may be a problem of uneven thickness in that area. The thickness uniformity of the car film refers to whether the thickness of the car film in different areas remains consistent. If the thickness is not uniform, it may cause fluctuations in the performance of the film layer, affecting the service life and appearance of the car film.

[0041] The surface roughness of each grid area is measured using roughness measuring instruments such as profilometer, white light interferometer, etc. The surface is scanned and the small undulations are recorded. The measured roughness value can be represented by parameters such as Ra (arithmetic mean roughness) or Rz (ten-point average roughness) to judge the smoothness of the surface of each grid area and evaluate the surface quality of the car film.

[0042] In order to ensure the accuracy of the thickness uniformity and surface roughness data of each grid, a synchronous measurement method is usually used, that is, the measurement equipment needs to collect the thickness and roughness data of each grid at the same time or as short as possible. Synchronous measurement can avoid measurement errors caused by time difference and ensure the consistency of the data.

[0043] By dividing the surface of the car film into multiple small grids and synchronously measuring the thickness uniformity and surface roughness, the quality of the car film can be evaluated in detail, and the possible defects on the surface of the car film can be found and quantified, improving the accuracy of quality detection.

[0044] S400: Perform multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector, and the thickness uniformity data to generate a coupling coefficient matrix.

[0045] Further, the S400 of the present application comprises: A dual-wavelength laser interferometer is configured, wherein the first wavelength is used to measure the substrate thickness, and the second wavelength is used to detect the surface coating thickness; the dual-wavelength laser interferometer is used to determine the interference fringe offset of each detection grid corresponding to the N detection grids, and obtain the thickness difference distribution of the substrate and the coating; based on the thickness difference distribution of the substrate and the coating, and the multi-scale correlation analysis of the chroma gradual change smoothness, a coupling degree index associated with the first dynamic change vector is configured.

[0046] For each detection grid, a local linear regression function of the color gradient smoothness and the thickness difference distribution of the substrate and coating is established; using the local linear regression function of the color gradient smoothness and the thickness difference distribution of the substrate and coating, a multi-scale correlation analysis is performed to obtain a multi-scale correlation coefficient r corresponding to each detection grid; and according to the multi-scale correlation coefficient r, a coupling degree index associated with the first dynamic change vector is configured.

[0047] The spatial overlap degree of the abnormal mutation region and the thickness abnormal region is analyzed by using a sliding window, and an overlap degree index is configured; based on the multi-scale correlation coefficient r and the overlap degree index, an accumulated variance contribution rate is determined; and based on the direction component of the first dynamic change vector, in combination with the multi-scale correlation coefficient r and the accumulated variance contribution rate, a coupling degree index associated with the first dynamic change vector is established.

[0048] Specifically, the color gradient distribution characteristics (i.e., the first, second, and third dynamic change vectors) and the thickness uniformity data are fused in multiple modalities to comprehensively evaluate the color and thickness of the car film and improve the accuracy of quality evaluation. First, the three dynamic change vectors and the thickness uniformity data are standardized to eliminate the influence caused by unit differences between different data, so that different data can be compared and fused on the same scale. The coupling coefficient matrix is generated by calculating the correlation between different modalities. The correlation between the first, second, and third dynamic change vectors and the thickness uniformity data is calculated by methods such as correlation coefficient or covariance. According to the correlation data, a coupling coefficient matrix is constructed to reflect the correlation strength between different data modalities. The coupling coefficient matrix is a mathematical matrix used to describe the correlation strength between multiple variables (such as thickness uniformity and color gradient) and evaluate the correlation between different detection indexes to provide a basis for subsequent defect evaluation.

[0049] The coupling coefficient represents the correlation or coupling degree between different features. Each element in the matrix represents the coupling degree between two features, reflecting the relationship strength between them. The construction of the matrix is usually based on the correlation calculation between different features. The coupling coefficients calculated from the first, second, and third dynamic change vectors and the thickness uniformity data are arranged in matrix form to obtain a coupling coefficient matrix, and each matrix element corresponds to the coupling degree between two features. According to the coupling coefficient matrix, the quality of each region is evaluated. A high coupling coefficient indicates a strong correlation between color gradient and thickness uniformity, usually indicating good quality of the car film; on the contrary, a low coupling coefficient may indicate uneven quality or defects.

[0050] A dual-wavelength laser interferometer is a device that uses the principle of laser interference to measure the thickness of an object's surface or interior. It emits two different wavelengths of laser beams to interfere with each other, and then measures the thickness of the object's surface or interior. The two wavelengths can be used to measure different levels of thickness, such as the thickness of the substrate and the thickness of the coating. The substrate is the main material layer of the car film, usually metal or plastic, which serves as the support for the coating. The substrate thickness refers to the actual thickness of this layer of material. The surface coating is the outer protective layer of the car film, usually composed of a layer of paint or film, and its thickness affects the durability and performance of the film.

[0051] When configuring a dual-wavelength laser interferometer, two different wavelengths of laser beams are selected based on the optical properties of the material being measured. The first wavelength is used to measure the thickness of the substrate, and the second wavelength is used to measure the thickness of the coating. The two wavelengths of laser beams will interfere with the surfaces of the substrate and coating, respectively. Using two different wavelengths of laser light can effectively distinguish the thickness of the coating and substrate, avoiding the problem of interference with different material layers that may occur when using only one wavelength.

[0052] The laser interferometer measures the displacement of the interference fringes by shining a laser beam on the surface being measured. Each grid point produces a set of interference fringes that reflect the thickness variation at that point. The different thicknesses of the substrate and coating cause differences in the phase of the reflected waves, resulting in a shift in the interference fringes. The thickness difference between the substrate and coating is calculated by calculating the shift in the interference fringes. There is a certain mathematical relationship between the shift and the thickness difference. For example, if the shift in the interference fringes is Δx, then the thickness difference Δd can be calculated by the ratio of the shift in the interference fringes to the wavelength of the light.

[0053] Measure N detection grids and calculate the thickness difference between the substrate and coating for each grid point. For each grid point, use the shift in the interference fringes and the wavelength information to calculate the thickness difference between the substrate and coating using the corresponding physical formula. Calculate the thickness difference for all grid points to obtain the thickness difference distribution between the substrate and coating, including the difference at each grid point, and visually display whether the film surface is uniform, whether there are weak areas or defect areas.

[0054] For each monitoring grid, a local linear regression model is established to describe the relationship between the color gradient smoothness and the substrate-to-coat thickness difference. The color gradient smoothness and the substrate-to-coat thickness difference of each detection grid are extracted. Using the local linear regression technique, the relationship between the color gradient smoothness and the thickness difference is calculated on each grid. The local linear regression function is fitted to the neighborhood data of each grid point, which can ensure that even if the conditions of different grids are significantly different, local characteristics can be modeled. Local linear regression is a statistical method for regression analysis in local regions of data, which establishes a local model by fitting the neighborhood data of each point to obtain the trend or relationship within the local region.

[0055] Through multi-scale correlation analysis, the correlation between the color gradient smoothness and the thickness difference distribution at different scales is analyzed. Multi-scale analysis is a technique for identifying and analyzing patterns and correlations at different scales (levels). Multi-scale analysis refers to analyzing the correlation between color gradient smoothness and thickness difference at different scales (i.e., different sizes of grids). Different scales may reveal different quality problems (such as large-scale unevenness or local minor defects). Multi-scale analysis involves detection grids of different sizes, which can be adjusted according to actual needs. Larger scales can capture global trends, while smaller scales can help reveal local details. At each scale, the correlation between the color gradient smoothness and the thickness difference distribution is calculated, and the multi-scale correlation coefficient r is obtained by calculating the Pearson correlation coefficient formula. The multi-scale correlation coefficient r is a numerical value used to measure the linear correlation between the color gradient smoothness and the thickness difference at different scales, usually the Pearson correlation coefficient. The r values calculated at multiple scales can represent the correlation strength at different scales.

[0056] Sliding window analysis is a common signal or data processing method that defines a window on the data and slides the window on the data to calculate and analyze the spatial overlap between the abnormal mutation area and the thickness abnormal area. The abnormal mutation area refers to the area with sharp or discontinuous color change in the color gradient map of the car film surface, which may indicate coating defects or coating unevenness. The thickness abnormal area refers to the area with significant thickness variation in the car film thickness distribution map, which usually indicates uneven or abnormal thickness of the substrate or coating. Spatial overlap refers to the degree of overlap between the abnormal mutation area and the thickness abnormal area in space, which is usually determined by calculating the ratio of the intersection part to the union part of the two areas.

[0057] A fixed-size window (e.g., 3x3, 5x5, or 7x7 grid) is defined, which slides over the image, and the abnormal region and the thickness abnormal region within each window are analyzed. On the color gradient map and the thickness distribution map of the car paint film, the window is slid, and the spatial positions of the abnormal region and the thickness abnormal region are checked in each window. For each sliding window, the spatial overlap between the abnormal region and the thickness abnormal region is calculated, i.e., the ratio of the intersection to the union of the two regions, and the overlap index reflects the degree of overlap between the two abnormal regions within the window.

[0058] The cumulative variance contribution rate is a weighted value that measures the contribution of multiple features (e.g., color gradient smoothness, multi-scale correlation coefficient, overlap index) to the overall quality assessment. According to the multi-scale correlation coefficient and the overlap index, the variance of each feature (e.g., color gradient smoothness, correlation coefficient, overlap) is calculated. Then, the contribution of each feature to the total variance is calculated, i.e., the ratio of the variance of each feature to the total variance, and the total variance is obtained by summing all feature variances. The cumulative variance contribution rate represents the cumulative contribution of each feature to the total variance, and the contribution rates of various features are gradually summed until all features are considered, understanding the importance and influence of each feature in the overall quality assessment.

[0059] The directional component of the first dynamic change vector is composed of the changes in hue and saturation, describing the color change trend of the car paint film. By calculating the changes in hue and saturation on each detection grid, a directional component can be obtained to represent the directionality of color change. The coupling index combines the directional component of the first dynamic change vector (representing the direction of color change) with the multi-scale correlation coefficient r and the cumulative variance contribution rate, such as coupling index = a*r + b*first dynamic change vector directional component + c*cumulative variance contribution rate, where a, b, and c are the influence weights of the multi-scale correlation coefficient, the first dynamic change vector directional component, and the cumulative variance contribution rate, respectively.

[0060] When the coupling index is high, it means that multiple features such as color gradient, coating thickness change, etc. are highly correlated, and the quality of the car paint film is relatively good. When the coupling index is low, it indicates that the relationship between these features is weak, and there may be irregular defects or uneven quality. The coupling index is used to quantify the relationship between multiple quality features (e.g., color gradient smoothness, thickness distribution, color difference, etc.) and their overall impact on quality assessment, reflecting the coordination degree and correlation of these features in the quality assessment of the car paint film.

[0061] By using dual-wavelength laser interferometer, local regression analysis and multi-scale correlation analysis, high-precision analysis of car paint film quality can be achieved, especially the establishment of a precise quantitative relationship between thickness difference and color change. Sliding window analysis and overlap index help to accurately identify potential defect areas on the surface of car paint film, especially for defects such as uneven coating thickness or gradual color change. By establishing the coupling index, the quality characteristics of each area of the car paint film are comprehensively evaluated, which helps to improve the overall quality level of the car paint film.

[0062] S500: Based on the coupling coefficient matrix and the range of surface roughness, the comprehensive quality index of the tested car paint film is evaluated, the comprehensive quality index is dynamically graded using the grading threshold interval, and the defect mapping coordinates under the grade identification of the car paint film are formulated.

[0063] Specifically, according to the coupling coefficient matrix and the range of surface roughness, weighted summation is performed to obtain the comprehensive quality index of the car paint film. The setting of the weighting coefficient can be adjusted according to the actual demand and the importance of the detection data, and the specific process is described in the corresponding steps. For the sake of brevity of the specification, only a simple summary is given. According to the range of the comprehensive quality index, the grading threshold interval is set. The grading threshold interval divides the comprehensive quality index into different grades, usually including high, medium and low quality grades. According to different quality requirements, the car paint film is classified to provide products with different quality levels for customers. For example, the comprehensive quality index greater than or equal to 0.8 is excellent, greater than or equal to 0.6 and less than 0.8 is good, greater than or equal to 0.4 and less than 0.6 is qualified, and less than 0.4 is unqualified. According to the comprehensive quality index of the car paint film, the grading threshold is dynamically adjusted according to the actual application scenario and customer requirements.

[0064] Using the grading threshold interval, the comprehensive quality index is dynamically graded to determine the quality grade of the car paint film and assign it the corresponding grade identification. Each grade identification corresponds to different quality standards and defect tolerance. For example, the excellent quality index is within the excellent range, indicating that the surface quality of the car paint film is very high and is suitable for high-demand scenarios; the good quality index is within the good range, indicating that the surface quality of the car paint film is good and is suitable for daily use; the qualified quality index is within the qualified range, and the car paint film meets the basic quality requirements and is suitable for general scenarios; the unqualified quality index is within the unqualified range, and the car paint film has obvious defects and needs to be repaired or remanufactured.

[0065] On the basis of comprehensive quality evaluation, the surface defects of the car film, such as scratches, bubbles, color difference, etc., are detected, and the defect area is marked. Combined with the position and size of the defect, the defect mapping coordinates are determined. The defect mapping coordinates refer to mapping the position and type of the defect to a coordinate system according to the car film detection results, and marking the specific position of the defect on the surface of the car film, which helps to locate the repair area or perform quality control. According to the type and position of the surface defects of the car film, image processing algorithms or sensor data are used to generate the mapping coordinates of the defects. The evaluation results and the defect mapping coordinates are fed back to the control center, and the local color compensation robot performs color compensation repair.

[0066] By coupling the coefficient matrix and the range of surface roughness, various quality factors are considered comprehensively and accurately to evaluate the quality of the car film. Through comprehensive quality evaluation and grading, the quality grade of the car film can be clearly marked, which helps to flexibly evaluate the quality of the car film. Through the generation of defect mapping coordinates and grade marking, the defects of the car film are quickly and accurately located and repaired, which can effectively guide the quality control and repair work, thereby ensuring the quality stability and efficient production of the car film and improving the accuracy of quality evaluation.

[0067] Further, the S500 of the present application includes: The coupling coefficient matrix is singular value decomposed to determine the first quality index corresponding to the maximum singular value. The range of the surface roughness is logarithmically transformed and weightedly summed with the first quality index, wherein the weight coefficient is dynamically adjusted according to the application scenario of the car film.

[0068] Specifically, the coupling coefficient matrix is singular value decomposed. Singular value decomposition is a matrix decomposition method in linear algebra, which decomposes a matrix into the product of three matrices: a left singular matrix (representing the direction of data features), a diagonal matrix (containing singular values), and a right singular matrix (representing the projection of data in another feature space). Through singular value decomposition, the main information of the matrix can be extracted, and the maximum singular value usually represents the most important feature of the matrix. The maximum singular value is the maximum value in the matrix, which usually represents the most important feature in the matrix. In quality evaluation, the maximum singular value is usually related to the dominant factor of the quality of the car film. Therefore, the maximum singular value can be used as the first quality index. The first quality index is the maximum singular value obtained by singular value decomposition, which represents the most significant quality information in the coupling coefficient matrix. The first quality index is usually the matrix feature corresponding to the maximum singular value, which reflects the dominant factor of the quality of the car film.

[0069] The range of surface roughness is the difference between the maximum and minimum values, reflecting the degree of surface fluctuation. The range of roughness is logarithmically transformed to compress the roughness data with a large range of variation, making it easier to handle and compare in analysis. Logarithmic transformation is a common data preprocessing method, especially suitable for data that grows exponentially or has extreme values. Logarithmic transformation of the range of surface roughness helps to compress the data range and reduce the impact of outliers.

[0070] The weight coefficients are dynamically adjusted according to the application scenarios of the car film. Different application scenarios may have different requirements for chroma gradient (first quality indicator) and surface roughness, such as when the quality of high-end car paint is required, chroma gradient may be more important than surface roughness; while in other application scenarios, surface roughness may have higher priority. Based on the actual use environment, customer requirements or specific application characteristics of the car film, the weight coefficients are adjusted. According to the determined weight coefficients, the first quality indicator (maximum singular value) and the logarithmically transformed range of surface roughness are combined by weighted summation to obtain the comprehensive quality index. Comprehensive quality index = w1 * first quality indicator + w2 * logarithmically transformed roughness range, where w1 and w2 are dynamically adjusted weight coefficients. The comprehensive quality index obtained by calculation comprehensively evaluates the quality of the car film, taking into account both color characteristics (through maximum singular value) and surface characteristics (through roughness range), allowing for multi-faceted quality monitoring of the car film.

[0071] Further, the present application further comprises the following steps: When scratches and / or bubbles are detected on the surface, the area ratio of defects is introduced as a penalty factor; using the penalty factor, the comprehensive quality index is modified.

[0072] Specifically, surface scratches refer to linear or irregular cracks or scratches on the surface of the car film, usually caused by external force or friction. Bubbles refer to gas clumps present on the surface or inside the car film, usually formed during the film coating process due to incomplete gas removal, resulting in uneven surface, affecting visual effect and functionality. Through high-resolution cameras or laser scanning, etc., detect whether there are scratches on the surface of the car film. Bubble detection can be achieved through surface inspection (visual inspection or detection methods based on optical principles). Bubbles usually appear optically as uneven surfaces or abnormal reflected light. Using a laser interferometer or high-resolution optical equipment, the presence and size of bubbles can be accurately measured.

[0073] When detecting scratches or bubbles on the surface of the gradient color car paint film, the area of the defect region is extracted, which is achieved through image processing techniques, usually by calculating the number of pixels or the actual surface area of the defect region. The area of the defect region is divided by the total surface area of the car paint film to obtain the defect area ratio. The penalty factor is a coefficient used to modify the comprehensive quality index, which is calculated based on the defect area ratio, and the penalty factor = 1 + a * defect area ratio, where a is a constant coefficient used to adjust the intensity of the penalty. When the car paint film surface has defects such as scratches, bubbles, etc., the penalty factor is used to adjust the quality index to reflect the negative impact of these defects on the overall quality.

[0074] The comprehensive quality index is modified using the penalty factor, and the modified comprehensive quality index can be represented by the formula: modified comprehensive quality index = original comprehensive quality index * (1 - penalty factor). Through the modified comprehensive quality index, quality control personnel can better evaluate the actual quality of the car paint film. If the modified comprehensive quality index is lower than the preset quality standard, it indicates that the car paint film surface has more serious defects (such as scratches, bubbles, etc.). By introducing the defect area ratio as the penalty factor, the impact of surface defects on the overall quality can be effectively considered, enhancing the accuracy of quality evaluation. Even if the car paint film is good in other aspects, but if there are obvious scratches or bubbles, it can still make accurate evaluation. By combining defect area ratio and quality modification, defects can be more sensitively identified, especially in the detection of small defects or low visibility defects, ensuring that no potential problems are missed.

[0075] Further, the present application further comprises the following steps: According to the defect mapping coordinates under the grade identification of the car paint film, the color repair area is located; the local color repair robot is started, and the micro-fluidic nozzle of the local color repair robot is controlled for color repair operation in combination with the color repair area; and the colorimetric gradient consistency of the color repair area is tested during the color repair operation.

[0076] Specifically, according to the detected surface defects of the car paint film, the types, positions and sizes of the defects are included. The defect position data is converted into a coordinate system to identify the area of the car paint film surface that needs to be repaired. The defect mapping coordinates refer to a spatial coordinate system established by detecting the position and type of the defect on the surface of the car paint film. The position of each defect has a corresponding coordinate on the surface of the car paint film, which can be used to locate the defect area for repair or other processing. According to the grade identification of the car paint film, the defects are classified and mapped. Different grades of car paint film may have different types and sizes of defects. Using the mapping coordinate information, the repair area is accurately located, thereby improving the repair efficiency and accuracy.

[0077] After determining the color correction repair area, start the color correction robot, and the microfluidic nozzle on the robot is automatically adjusted by the control center to perform accurate color correction operation in the specified area. The color correction robot can be accurately positioned through image recognition and a coordinate system, ensuring that the color correction operation is only performed in the required area, avoiding waste and incorrect spraying. The microfluidic nozzle accurately adjusts the spraying amount and spraying range according to the size and shape of the defect area. The control center of the robot can adjust the working state of the nozzle in real time to ensure that the coating uniformly covers the defect area to achieve the ideal color repair effect. The microfluidic nozzle is a precise spraying device that can accurately control the spraying amount, spraying position and spraying speed of the coating, and uses the principle of fluid mechanics to adjust the flow of liquid through a small control unit to ensure the uniformity and precision of the coating.

[0078] During the color correction operation, the color hue gradient consistency of the color correction repair area is monitored in real time, including detection of hue, saturation and brightness, to ensure smooth color change in the color correction area and natural transition with the surrounding area, avoiding obvious color difference or color blocks. The color hue gradient consistency verification algorithm is used to analyze the hue, saturation and brightness changes of the color correction area to evaluate the transition effect with the surrounding area. When inconsistent color hue gradient is detected, the color correction operation is automatically fed back and adjusted to ensure that the repair effect meets the requirements. If the color hue gradient consistency verification fails, multiple adjustments of the spraying are performed for optimization until the color gradient of the color correction area and the surrounding area is consistent, achieving the best repair effect. The color hue gradient consistency verification refers to detecting the color hue change of the car paint film repair area to ensure that the color of the repair area and the color of the surrounding area are consistent in the color hue gradient. By detecting the smooth transition of parameters such as hue, saturation and brightness, the repair effect is ensured to be natural and not abrupt.

[0079] Through the precise control of the local color correction robot and the microfluidic nozzle, local repair can be performed in the specified defect area, avoiding unnecessary coating and resource waste. The color hue gradient consistency verification effectively ensures that the transition of the repair area color and the surrounding area is natural and smooth, avoiding obvious color difference and inconsistency, and improving the overall aesthetic appearance and quality standard of the car paint film.

[0080] In summary, the gradual color car paint film grading detection method provided by the present application has the following beneficial effects: The spectral reflectance data of the to-be-tested car film in the visible light band is acquired, and the colorimetric gradient distribution characteristics are extracted; based on the colorimetric gradient distribution characteristics, a first dynamic change vector combined by hue and saturation, a second dynamic change vector combined by lightness and hue, and a third dynamic change vector combined by lightness and saturation are configured; N detection grids are divided according to the to-be-tested car film, and the thickness uniformity data and the surface roughness of each detection grid are synchronously measured; the first dynamic change vector, the second dynamic change vector, and the third dynamic change vector are subjected to multi-modal fusion with the thickness uniformity data to generate a coupling coefficient matrix; based on the coupling coefficient matrix and the range value of the surface roughness, the comprehensive quality index of the to-be-tested car film is evaluated, the comprehensive quality index is dynamically classified using a classification threshold interval, and the defect mapping coordinates under the grade identification of the car film are formulated. That is, the spectral reflectance data of the car film in the visible light band is acquired, the colorimetric gradient distribution characteristics are extracted, the comprehensive quality index of the car film is comprehensively evaluated, the dynamic classification of the quality of the car film is realized, the defect mapping coordinates under the grade identification are accurately formulated, and then the defects of the gradient color car film are quickly located, and the precision of the quality evaluation is improved.

[0081] In the embodiment two, based on the same inventive concept as the gradient color car film grading detection method in the aforementioned embodiment one, the present application further provides a gradient color car film grading detection system, please refer to the attached Figure 2 The gradient color car film grading detection system comprises: A feature extraction module 11 is configured to acquire the spectral reflectance data of the to-be-tested car film in the visible light band and extract the colorimetric gradient distribution characteristics; a change vector configuration module 12 is configured to configure a first dynamic change vector combined by hue and saturation, a second dynamic change vector combined by lightness and hue, and a third dynamic change vector combined by lightness and saturation based on the colorimetric gradient distribution characteristics; a detection network division module 13 is configured to divide N detection grids according to the to-be-tested car film and synchronously measure the thickness uniformity data and the surface roughness of each detection grid; a multi-modal fusion module 14 is configured to perform multi-modal fusion of the first dynamic change vector, the second dynamic change vector, and the third dynamic change vector with the thickness uniformity data to generate a coupling coefficient matrix; and a dynamic classification module 15 is configured to evaluate the comprehensive quality index of the to-be-tested car film based on the coupling coefficient matrix and the range value of the surface roughness, dynamically classify the comprehensive quality index using a classification threshold interval, and formulate the defect mapping coordinates under the grade identification of the car film.

[0082] Further, the feature extraction module 11 in the gradient color car film grading detection system is further configured to: The reflectance spectrum data in a preset wavelength range is collected by scanning the car paint film to be tested, the reflectance spectrum data is subjected to baseline correction processing, and a reflectance characteristic value is extracted; based on the reflectance characteristic value, a continuous spectrum reflectance curve is reconstructed, and a gradient change rate of adjacent wavelength reflectance is determined; and according to a ratio of a difference value of the gradient change rate of adjacent wavelength reflectance and a wavelength difference value, the chroma gradient distribution characteristic is determined.

[0083] Further, the change vector configuration module 12 in the graded detection system for the gradient color car paint film is further used for: The reconstructed continuous spectrum reflectance curve is converted to generate a hue angle distribution heat map; an abnormal mutation region in the hue angle distribution heat map is identified and marked as a potential color difference defect; based on the potential color difference defect, a chroma gradient smoothness is quantified according to a hue angle standard deviation and a saturation gradient direction consistency index, and the chroma gradient smoothness is used to configure a direction component of the first dynamic change vector.

[0084] Further, the multi-modal fusion module 14 in the graded detection system for the gradient color car paint film is further used for: A dual-wavelength laser interferometer is configured, wherein a first wavelength is used to measure a substrate thickness, and a second wavelength is used to detect a surface coating thickness; the dual-wavelength laser interferometer is used to determine an interference fringe offset amount of each detection grid corresponding to the N detection grids, and to obtain a thickness difference distribution of the substrate and the coating; based on the thickness difference distribution of the substrate and the coating, a multi-scale correlation analysis is performed on the chroma gradient smoothness to configure a coupling degree index associated with the first dynamic change vector.

[0085] Further, the multi-modal fusion module 14 in the graded detection system for the gradient color car paint film is further used for: For each detection grid, a local linear regression function of the chroma gradient smoothness and the thickness difference distribution of the substrate and the coating is established; the local linear regression function of the chroma gradient smoothness and the thickness difference distribution of the substrate and the coating is used to perform a multi-scale correlation analysis to obtain a multi-scale correlation coefficient r corresponding to each detection grid; and based on the multi-scale correlation coefficient r, a coupling degree index associated with the first dynamic change vector is configured.

[0086] Further, the multi-modal fusion module 14 in the graded detection system for the gradient color car paint film is further used for: The spatial overlap degree of the abnormal mutation region and the thickness abnormal region is analyzed by using a sliding window to configure an overlap degree index; based on the multi-scale correlation coefficient r and the overlap degree index, an accumulated variance contribution rate is determined; and based on the direction component of the first dynamic change vector, the multi-scale correlation coefficient r and the accumulated variance contribution rate are combined to establish the coupling degree index associated with the first dynamic change vector.

[0087] Furthermore, the dynamic grading module 15 in the gradient color car cover film grading detection system is further used to: Perform singular value decomposition on the coupling coefficient matrix to determine the first quality index corresponding to the maximum singular value; perform logarithmic transformation on the range value of the surface roughness and perform weighted summation with the first quality index, wherein the weight coefficient is dynamically adjusted according to the application scenario of the car cover film.

[0088] Furthermore, the dynamic grading module 15 in the gradient color car cover film grading detection system is further used to: When scratches and / or bubbles are detected on the surface, the defect area ratio is introduced as a penalty factor; and the comprehensive quality index is corrected using the penalty factor.

[0089] Furthermore, the dynamic grading module 15 in the gradient color car cover film grading detection system is further used to: According to the defect mapping coordinates under the car cover film grade mark, locate the color repair area; start the local color repair robot, and control the microfluidic nozzle of the local color repair robot to perform color repair operations in combination with the color repair area; at the same time, during the color repair operation, perform a chromaticity gradient consistency test on the color repair area. The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The gradient color car cover film grading detection method and specific examples in Example 1 are also applicable to a gradient color car cover film grading detection system in this embodiment. Through the above detailed description of a gradient color car cover film grading detection method, those skilled in the art can clearly understand a gradient color car cover film grading detection system in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0090] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0091] Obviously, for those skilled in the art, several improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A method for grading detection of a gradient color car paint film, characterized in that, The method comprises the following steps: Obtain the spectral reflectance data of the test car film in the visible light band, and extract the color gradient distribution characteristics; Based on the color gradient distribution characteristics, configure the first dynamic change vector combined by hue and saturation, the second dynamic change vector combined by lightness and hue, and the third dynamic change vector combined by lightness and saturation; According to the test car film, divide N detection grids, and synchronously measure the thickness uniformity data and surface roughness of each detection grid; Multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector and the thickness uniformity data generates a coupling coefficient matrix; Based on the coupling coefficient matrix and the range value of the surface roughness, evaluate the comprehensive quality index of the test car film, dynamically classify the comprehensive quality index using a grading threshold interval, and formulate the defect mapping coordinates under the car film grade label.

2. The method for grading detection of a gradient color car paint film according to claim 1, wherein, Obtain the spectral reflectance data of the test car film in the visible light band, and extract the color gradient distribution characteristics, comprising: Scan the test car film, collect the reflectance spectrum data in the preset wavelength range, perform baseline correction processing on the reflectance spectrum data, and extract the reflectance characteristic value; Based on the reflectance characteristic value, reconstruct the continuous spectral reflectance curve, and determine the gradient change rate of adjacent wavelength reflectance; According to the ratio of the gradient change rate difference value and the wavelength difference value of adjacent wavelength reflectance, determine the color gradient distribution characteristics.

3. The method for grading the color-changing car paint film according to claim 1, wherein Based on the color gradient distribution characteristics, configure the first dynamic change vector combined by hue and saturation, comprising: Convert the reconstructed continuous spectral reflectance curve to generate a hue angle distribution heat map; Identify the abnormal mutation area in the hue angle distribution heat map and mark it as a potential color difference defect; Based on the potential color difference defect, quantize the color gradual change smoothness according to the hue angle standard deviation and the saturation gradient direction consistency index, and the color gradual change smoothness is used to configure the direction component of the first dynamic change vector.

4. The method for grading detection of a gradient color car paint film according to claim 3, wherein Multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector and the thickness uniformity data generates a coupling coefficient matrix, which further comprises: Configure a dual-wavelength laser interferometer, wherein the first wavelength is used to measure the substrate thickness, and the second wavelength is used to detect the surface coating thickness; Using the dual-wavelength laser interferometer, determine the interference fringe shift amount of each detection grid corresponding to the N detection grids, and obtain the thickness difference distribution of the substrate and the coating; Based on the thickness difference distribution of the substrate and the coating, and the color gradual change smoothness, perform multi-scale correlation analysis to configure the coupling degree index associated with the first dynamic change vector.

5. The method for grading detection of a gradient color car film according to claim 4, wherein, Based on the thickness difference distribution of the substrate and the coating, and the color gradual change smoothness, perform multi-scale correlation analysis to configure the coupling degree index associated with the first dynamic change vector, comprising: For each detection grid, establish a local linear regression function of the color gradual change smoothness and the thickness difference distribution of the substrate and the coating; Performing multi-scale correlation analysis using the local linear regression function of the color gradient smoothness and the thickness difference distribution of the substrate and the coating layer to obtain a multi-scale correlation coefficient r corresponding to each detection grid; According to the multi-scale correlation coefficient r, a coupling degree index associated with the first dynamic change vector is configured.

6. The method for grading detection of a gradient color car film according to claim 5, wherein, According to the multi-scale correlation coefficient r, a coupling degree index associated with the first dynamic change vector is configured, including: Using a sliding window to analyze the spatial overlap degree of the abnormal mutation region and the thickness abnormal area, and configuring an overlap degree index; Based on the multi-scale correlation coefficient r and the overlap degree index, the cumulative variance contribution rate is determined; Based on the direction component of the first dynamic change vector, combined with the multi-scale correlation coefficient r and the cumulative variance contribution rate, the coupling degree index associated with the first dynamic change vector is established.

7. The method of claim 1, wherein the step of detecting the color of the film is performed by a colorimeter. Based on the coupling coefficient matrix and the range value of the surface roughness, the comprehensive quality index of the tested car paint film is evaluated, which further includes: Performing singular value decomposition on the coupling coefficient matrix to determine a first quality index corresponding to the maximum singular value; After logarithmic transformation of the range value of the surface roughness, the first quality index is weighted and summed, wherein the weight coefficient is dynamically adjusted according to the application scenario of the car paint film.

8. The method for grading detection of a gradient color car film according to claim 7, wherein, When detecting that the surface has scratches and / or bubbles, introduce the defect area ratio as a penalty factor; Using the penalty factor, the comprehensive quality index is modified.

9. The method of claim 8, wherein the step of detecting the color of the film is performed by a colorimeter. The comprehensive quality index is dynamically classified, and the defect mapping coordinates under the car paint film grade identification are also included. According to the defect mapping coordinates under the car paint film grade identification, the color repair area is located; Start the local color repair robot, and control the micro-fluidic nozzle of the local color repair robot for color repair operation combined with the color repair area; At the same time, the color gradient consistency of the color repair area is tested during the color repair operation.

10. A graded color car paint film grading system characterized by, The steps of the gradual color car paint film grading detection method according to any one of claims 1-9 are implemented by the gradual color car paint film grading detection system, which includes: A feature extraction module is configured to obtain the spectral reflectance data of the tested car paint film in the visible light band and extract the color gradient distribution features; A change vector configuration module is configured to configure a first dynamic change vector combined by hue and saturation, a second dynamic change vector combined by lightness and hue, and a third dynamic change vector combined by lightness and saturation based on the color gradient distribution features; A detection network division module is configured to divide N detection grids according to the tested car paint film, and simultaneously measure the thickness uniformity data and surface roughness of each detection grid; A multi-modal fusion module is configured to perform multi-modal fusion of the first dynamic change vector, the second dynamic change vector, the third dynamic change vector, and the thickness uniformity data to generate a coupling coefficient matrix; A dynamic classification module is configured to evaluate the comprehensive quality index of the tested car paint film based on the coupling coefficient matrix and the range value of the surface roughness, use a classification threshold interval to dynamically classify the comprehensive quality index, and determine the defect mapping coordinates under the car paint film grade identification.

Images