Coating stone impact resistance test evaluation method and system based on AI assistance

Abstract

The invention relates to the technical field of coating stone impact resistance testing, and provides a coating stone impact resistance testing evaluation method and system based on AI assistance. According to the method, a tested sample image is obtained, an image processing algorithm is utilized to automatically identify a damage area and distinguish different damage depth levels (such as finish paint, electrophoresis layer and metal substrate damage), and then comprehensive damage degree parameters are calculated based on damage area proportion and depth information. And finally, generating an automatic evaluation result (such as a digital rating or a visual report) by comparing a preset rating standard. According to the method, full-process automatic evaluation is realized, subjective factor interference is eliminated, the objectivity, accuracy and efficiency of evaluation are remarkably improved, and reliable and repeatable technical support is provided for coating performance detection.

Description

Technical Field

[0001] This invention relates to the field of coating stone impact resistance testing technology, and more specifically, to an AI-assisted coating stone impact resistance testing and evaluation method and system. Background Technology

[0002] Currently, evaluation methods for coating stone impact resistance tests mainly rely on manual visual inspection or semi-automated techniques based on image processing. As a representative of existing technologies, a patent document discloses an image detection method based on coating stone impact resistance test results. It proposes a detection process involving image preprocessing, morphology analysis, and classification statistics, employing machine learning algorithms (such as logistic regression models) to process damage feature data, aiming to achieve intelligent rating. However, this method relies on complex image processing steps (such as translation registration and differential processing) and training the model with historical data, which may lead to insufficient algorithm stability: for example, when damage features are inconsistent or frame difference thresholds are improperly set, classification results are prone to bias, and it requires a large amount of labeled data, making it difficult to implement practically.

[0003] Another patent document discloses a method for evaluating the degree of stone chip damage to paint films. It provides a simpler approach by calculating the proportion of damaged pixels through grayscale conversion and binarization, and combining user evaluations to improve accuracy. However, this method is overly simplistic. Its binarization process relies on manually adjusting feature thresholds, lacks adaptability, and struggles to accurately distinguish damage depth (such as differences between topcoat, electrophoretic layer, or metal substrate). The final evaluation is based solely on pixel proportions, ignoring multidimensional damage features, and is easily affected by lighting or acquisition angle.

[0004] More fundamentally, neither of the above two solutions integrates a dedicated hardware system; image acquisition relies on general-purpose equipment, leading to errors introduced by factors such as light source and angle, thus failing to guarantee consistent image quality. These problems, combined with the inherent defects of manual methods (such as subjectivity and unstable reproducibility), make it difficult for existing technologies to meet the demands of high-precision, repeatable industrial inspection, especially when dealing with minor damage or large batches of samples, where they are inefficient.

[0005] Therefore, it is necessary to provide a new solution to address the core issues of objectivity, consistency, and accurate quantification in the stone impact resistance test of coatings that existing technologies have failed to resolve. Summary of the Invention

[0006] This invention addresses the technical problems existing in the prior art by providing an AI-assisted method and system for evaluating the stone impact resistance of coatings. It solves the technical problems of strong subjectivity, unstable reproducibility, and inability to efficiently quantify multi-dimensional parameters such as damage area and depth caused by relying on manual visual inspection or simple image processing.

[0007] According to a first aspect of the present invention, an AI-assisted method for evaluating the stone impact resistance of coatings is provided, comprising: S1, Obtain sample images after the stone impact resistance test of the coating; S2, The sample image is analyzed using an image processing algorithm to identify the damaged area and distinguish different levels of damage depth; S3, Calculate the damage degree parameters based on the area ratio and damage depth information of the damaged area; S4. The damage degree parameters are compared with preset rating standards to generate automated evaluation results.

[0008] Based on the above technical solution, the present invention can also be improved as follows.

[0009] Optionally, step S1 includes: S101, the coated sample that has completed the stone impact resistance test is positioned in a standardized fixing method, and the surface orientation of the sample is corrected by using grid markings to eliminate the sampling angle deviation and deformation error; S102, adjust the illumination parameters under diffuse lighting environment to make the light evenly distributed on the sample surface and generate the original image of gray scale distribution and damage depth correlation; S103: Collects digital image data of the sample surface, performs image quality verification simultaneously, and triggers a re-acquisition mechanism for images that do not meet the clarity or exposure standards. S104, Save the verified sample image.

[0010] Optionally, in step S2, the analysis of the sample image using an image processing algorithm to identify the damaged area includes: S201 uses a thresholding algorithm to divide the pixel grayscale values ​​in the sample image into multiple discrete intervals, with each interval associated with a specific damage depth level, thus achieving preliminary differentiation of the damage area. S202, the region segmentation algorithm is used to identify the contour of the damaged region and segment out the damaged region; S203, classify the segmented damaged areas according to depth level. The classification categories include at least topcoat damage, electrophoretic layer damage and metal substrate damage. Determine the depth information of the damaged areas based on pixel distribution or rule matching. S204: Extract the geometric feature parameters of the damaged area, including at least area, perimeter and distribution density, and generate a comprehensive feature vector by combining the depth level.

[0011] Optionally, before step S201, the following may also be included: The sample image is preprocessed, and the preprocessing includes at least one of image scanning, noise reduction, and resolution normalization.

[0012] Optionally, determining the depth information of the damaged region based on pixel distribution or rule matching includes: By analyzing pixel grayscale values, the pixel distribution in the sample image is matched with preset rules, where the preset rules define the mapping relationship between grayscale ranges and damage depth levels. Based on the pixel matching results, each pixel is classified into a corresponding damage depth level, which includes at least topcoat damage, electrophoretic layer damage and metal substrate damage. Based on the pixel classification results, the regional distribution of each damage depth level is statistically analyzed to determine the depth information of the sample image.

[0013] Optionally, step S3 includes: S301, based on the image analysis results, the ratio of the number of pixels in different depth levels to the total number of pixels in the region is statistically analyzed to obtain the area proportion of each damage depth level; S302 assigns corresponding weighting coefficients to different damage depth levels; S303, combine the area proportion of each damage level with the corresponding weighting coefficient to calculate the weighted damage parameters; S304 integrates the weighted parameters of all damage levels and generates a comprehensive damage severity parameter through weighted summation or product combination.

[0014] Optionally, step S4 includes: S401, Obtain a preset rating standard, which is based on historical test data, industry standards or experimental calibration definitions, including rating levels or threshold ranges corresponding to different damage degree parameters. S402, The calculated damage degree parameter is matched and compared with the preset rating standard, and the rating range or level to which the parameter belongs is determined by the comparison algorithm; S403, based on the matching results, automatically generate evaluation results, including numerical ratings, category labels, or visual reports.

[0015] Optionally, step S4 also includes: S404, to verify or round the evaluation results.

[0016] According to a second aspect of the present invention, an AI-assisted coating stone impact resistance testing and evaluation system is provided, comprising: The image acquisition module is used to acquire sample images of the coating after the stone impact resistance test under standardized conditions. The image processing module is used to analyze the sample image through image processing algorithms to identify the damaged area and distinguish different levels of damage depth. The parameter calculation module is used to calculate the damage degree parameters based on the area ratio and damage depth information of the damaged area; The rating module is used to compare the damage degree parameters with preset rating standards and generate automated evaluation results.

[0017] Optionally, the image processing module is configured to analyze the sample image using a thresholding algorithm, dividing the pixel grayscale values ​​into multiple discrete intervals, each interval being mapped to a specific damage depth level, so as to achieve automatic identification and depth differentiation of the damage area.

[0018] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the processor is configured to implement the steps of the above-described AI-assisted coating stone impact resistance test evaluation method when executing a computer management program stored in the memory.

[0019] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer management program is stored, wherein the computer management program, when executed by a processor, implements the steps of the above-described AI-assisted coating stone impact resistance test evaluation method.

[0020] This invention provides an AI-assisted method, system, electronic device, and storage medium for evaluating the stone impact resistance of coatings. It achieves automated identification and quantitative evaluation of coating damage through image processing algorithms. First, it acquires post-test sample images. Then, it analyzes image features using intelligent algorithms to accurately identify damaged areas and distinguish different damage depth levels (such as damage to topcoat, electrophoretic layer, and metal substrate). Subsequently, it calculates comprehensive damage parameters based on the damage area ratio and depth information. Finally, it automatically generates evaluation results by comparing them with preset rating standards. This invention transforms the traditional subjective evaluation process, which relies on manual visual inspection, into an objective and standardized AI analysis process. Firstly, it completely eliminates human interference, ensuring high repeatability and reproducibility of evaluation results. Secondly, it improves the scientific rigor and accuracy of evaluation indicators through the fusion calculation of multi-dimensional damage parameters (area and depth). Thirdly, it automates the entire process from image acquisition to rating output, significantly improving detection efficiency and providing quantifiable and traceable data support for coating performance evaluation. Attached Figure Description

[0021] Figure 1 A flowchart of an AI-assisted coating stone impact resistance test evaluation method provided by the present invention; Figure 2 A schematic diagram of a diffuse bright field and CCD camera 2 in conjunction with a certain embodiment is provided; Figure 3 A schematic diagram of the original sample image collected for one embodiment; Figure 4A comparative schematic diagram of various damaged areas after image processing, provided for one embodiment; Figure 5 A schematic diagram of rating results provided for one embodiment; Figure 6 A block diagram of an AI-assisted coating stone impact resistance testing and evaluation system provided by the present invention; Figure 7 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention; Figure 8 This is a schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention.

[0022] The attached diagram lists the components represented by each number as follows: 1. Light source box, 101. LED light source, 2. CCD camera, 3. Coating sample. Detailed Implementation

[0023] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0024] Figure 1 A flowchart of an AI-assisted coating stone impact resistance testing and evaluation method provided by this invention is shown below. Figure 1 As shown, the method includes steps S1 to S4: S1, Obtain sample images after the stone impact resistance test of the coating; S2, The sample image is analyzed using an image processing algorithm to identify the damaged area and distinguish different levels of damage depth; S3, Calculate the damage degree parameters based on the area ratio and damage depth information of the damaged area; S4. The damage degree parameters are compared with preset rating standards to generate automated evaluation results.

[0025] Understandably, given the deficiencies in the background technology, this invention proposes an AI-assisted method for evaluating the stone impact resistance of coatings. This method achieves automated identification and quantitative evaluation of coating damage through image processing algorithms. First, post-test sample images are acquired. Then, intelligent algorithms analyze image features to accurately identify damaged areas and distinguish different damage depth levels (such as damage to topcoat, electrophoretic layer, and metal substrate). Subsequently, comprehensive damage parameters are calculated based on the damage area ratio and depth information. Finally, evaluation results are automatically generated by comparing with preset rating standards.

[0026] This invention transforms the traditional subjective evaluation process, which relies on human visual inspection, into an objective and standardized AI analysis process. First, it completely eliminates the interference of human factors, ensuring a high degree of repeatability and reproducibility of evaluation results. Second, it improves the scientific nature and accuracy of evaluation indicators through the fusion calculation of multi-dimensional damage parameters (area and depth). Third, it achieves full automation from image acquisition to rating output, significantly improving detection efficiency and providing quantifiable and traceable data support for coating performance evaluation.

[0027] Based on the above technical solutions, the embodiments of the present invention can be further improved as follows.

[0028] In one possible embodiment, step S1, acquiring a sample image after the coating stone impact resistance test, includes sub-steps S101~S104: S101, the coated sample 3 that has completed the stone impact resistance test is positioned using a standardized fixing method, and the surface orientation of the sample is corrected using grid markings to eliminate sampling angle deviation and deformation error.

[0029] For example, in a practical implementation scenario, the coated sample 3, having completed the stone impact resistance test, is placed on a dedicated fixing device (such as a fixture with a grid scale). The operator adjusts the sample's position to align the grid markings (such as intersecting lines or coordinate points) with the reference lines of the acquisition system. For instance, the center point of the grid on the sample surface is aligned with the camera's optical axis, thereby correcting the sample's orientation and eliminating acquisition angle deviations and deformation errors caused by tilting or bending. This operation provides distortion-free input for subsequent image analysis, improving evaluation accuracy.

[0030] S102 adjusts the illumination parameters under diffuse lighting conditions to make the light evenly distributed on the sample surface, generating an original image that correlates grayscale distribution with damage depth.

[0031] For example Figure 2 A possible diffuse lighting environment is demonstrated, in which a diffuse bright field is formed near the test surface of the coating sample 3 by the light source box 1, and the lighting parameters are adjusted to generate a high-quality image by the coordinated operation of the diffuse bright field lighting and the CCD camera 2.

[0032] Combination Figure 2 As shown, in actual operation, the diffused bright field light source box 1 (such as an LED light source 101 with an integrated diffuser plate) is activated, emitting uniform and soft light to cover the surface of the coated sample 3, avoiding localized reflections or shadows. Simultaneously, the operator adjusts the light source intensity, illumination angle, and camera exposure parameters (such as setting ISO and shutter speed via a software interface) to ensure uniform light distribution. This allows the CCD camera 2 (such as one equipped with a high-resolution sensor) to capture an image from a preset angle, generating an original image where grayscale values ​​are precisely correlated with damage depth. For example... Figure 3The image shown is a sample image captured by CCD camera 2. The deeper the damage, the weaker the reflected light and the lower the gray value. For example, the electrophoretic layer damage appears dark gray, while the paint damage appears light gray, providing reliable input for subsequent analysis.

[0033] This step eliminates environmental interference through standardized lighting control, ensuring image consistency and repeatability, and allowing damage depth information to be presented objectively.

[0034] S103, via Figure 2 The CCD camera 2 or other image acquisition device shown acquires digital image data of the sample surface and simultaneously performs image quality verification, triggering a re-acquisition mechanism for images that do not meet the sharpness or exposure standards.

[0035] S104, Save the verified sample image.

[0036] In this embodiment, a standardized image acquisition process ensures input data quality, thereby eliminating environmental interference and guaranteeing evaluation consistency and repeatability. Specifically, based on the attached... Figure 2 The spatial relationship between the diffuse bright-field illumination and the CCD camera 2 is shown. Diffuse illumination eliminates local reflections, and the CCD camera 2 captures the true damage morphology at an oblique angle. Combined with sample gridding and positioning to correct the acquisition angle, a standardized image with a precise correlation between grayscale values ​​and damage depth is generated. This embodiment solves the image distortion problem caused by uneven lighting and angle deviations in traditional acquisition methods, ensuring the input quality for subsequent AI analysis from the source. It provides a reliable physical basis for damage depth differentiation and area calculation, guaranteeing the objectivity of the evaluation results from the "image generation" stage.

[0037] In one possible embodiment, prior to performing step S2, the acquired sample images are preprocessed. This preprocessing includes at least one of image scanning, noise reduction, and resolution normalization.

[0038] Understandably, in this embodiment, the quality of sample images is optimized through techniques such as image scanning, noise reduction, and resolution standardization. Image scanning ensures comprehensive capture of image data, noise reduction reduces interference introduced by the environment or equipment (such as dust or uneven lighting), and resolution standardization unifies image size and clarity, thereby providing clean and consistent input for subsequent thresholding and damage analysis. This embodiment improves the accuracy and reliability of subsequent damage identification and depth differentiation by eliminating image distortion and variation factors, thus enhancing the reliability and repeatability of the evaluation method.

[0039] In one possible embodiment, step S2, which involves analyzing the sample image using an image processing algorithm to identify damaged areas, includes sub-steps S201 to S204.

[0040] S201, Threshold processing: By using a thresholding algorithm, the pixel grayscale values ​​in the sample image are divided into multiple discrete intervals, each interval being associated with a specific damage depth level, thus achieving preliminary differentiation of the damage area.

[0041] In practice, thresholding algorithms can be applied to the sample images. For example, using the Otsu automatic thresholding method, the image pixel grayscale values ​​are divided into four discrete intervals: 0-63, 64-127, 128-191, and 192-255 (the interval range is for illustrative purposes only and can be adjusted in practice). Each interval maps to a specific damage depth level: grayscale values ​​0-63 correspond to metal substrate damage (deep damage), grayscale values ​​64-127 correspond to electrophoretic layer damage (medium damage), grayscale values ​​128-191 correspond to topcoat damage (shallow damage), and grayscale values ​​192-255 correspond to undamaged areas. In this step, the algorithm automatically labels different grayscale intervals with different colors (such as green, red, blue, and gray) to achieve preliminary differentiation of damaged areas. For example... Figure 3 This demonstrates the effect of the original sample image (grayscale). Figure 4 This demonstrates the effect of color-coding different grayscale value ranges after thresholding, i.e., the mapping effect between grayscale ranges and color coding. Figure 4 The degree of damage in each region of the sample image can be seen intuitively.

[0042] S202, Region Segmentation: A region segmentation algorithm is used to identify the contour of the damaged region and segment out the damaged region.

[0043] In practice, based on the thresholding results of S201, region segmentation algorithms (such as the watershed algorithm or edge detection) are used to identify the contours of damaged regions. For example, the algorithm first performs morphological operations (such as dilation and erosion) on the binarized image to eliminate noise, then detects the boundaries of connected regions, separating the damaged regions from the undamaged background and generating a binary mask image. After segmentation, each damaged region is independently labeled for easy subsequent analysis.

[0044] S203, Deep Classification: The segmented damaged areas are classified according to depth level. The classification categories include at least topcoat damage, electrophoretic layer damage, and metal substrate damage. The depth information of the damaged areas is determined based on pixel distribution or rule matching.

[0045] More specifically, by analyzing pixel grayscale values, the pixel distribution in the sample image is matched with preset rules, where the preset rules define the mapping relationship between grayscale ranges and damage depth levels. Based on the pixel matching results, each pixel is classified into a corresponding damage depth level, which includes at least topcoat damage, electrophoretic layer damage and metal substrate damage. Based on the pixel classification results, the regional distribution of each damage depth level is statistically analyzed to determine the depth information of the sample image.

[0046] In practice, this step classifies the damaged areas segmented by S202 according to depth levels. The classification is based on mapping rules defined in the thresholding process: for example, green areas (grayscale 0-63) are classified as metal substrate damage, red areas (grayscale 64-127) as electrophoretic layer damage, and blue areas (grayscale 128-191) as topcoat damage. The algorithm confirms depth information through pixel distribution statistics (such as the average grayscale value of pixels within the region) or rule matching (such as color code comparison) to ensure accurate classification. For example... Figure 3 The image described is a contrasting image, which reflects the relationship between grayscale values ​​and damage depth, supporting the classification logic of this step.

[0047] S204, Feature Extraction: Extract the geometric feature parameters of the damaged area, including at least area, perimeter and distribution density, and combine them with the depth level to generate a comprehensive feature vector.

[0048] In practice, geometric feature parameters are extracted for each damaged region. For example, the area (number of pixels), perimeter (boundary pixel length), and distribution density (number of damage points per unit area) of the damaged region are calculated. Combined with the depth level, a comprehensive feature vector is generated: for example, for a damaged region in an electrophoretic layer, the output vector is [area percentage = 5%, perimeter = 120 pixels, density = 0.1 points / mm², depth level = medium damage] (parameters are for illustrative purposes only). This comprehensive feature vector provides structured input for the subsequent damage level parameter calculation step S3.

[0049] It is understood that this embodiment fully automates the damage area identification process, significantly improves the objectivity and accuracy of the evaluation, eliminates subjective errors by replacing manual visual inspection with algorithms, enhances the consistency and repeatability of the test results, and improves processing efficiency, providing a reliable and quantifiable analytical basis for coating stone impact resistance testing.

[0050] In one possible embodiment, step S3 includes sub-steps S301 to S304.

[0051] S301, Calculate the area percentage: Based on the image analysis results, the ratio of the number of pixels in different depth levels to the total number of pixels in the region is statistically analyzed to obtain the area proportion of each damage depth level.

[0052] For example, in actual operation, the system reads the threshold and processes it as follows: Figure 4The image shown uses different colors to represent different damage depths. The ratio of the number of pixels in each color region to the total number of pixels in the region is calculated. For example, the pixel statistics are as follows: Blue area (paint damage) pixel count: 15,000 pixels Red area (electrophoretic layer damage) pixel count: 8,000 pixels Green area (damage to the metal substrate) Pixel count: 2,000 pixels Total effective area pixels: 100,000 pixels The area percentages for each level of damage are as follows: topcoat damage 15%, electrophoretic layer damage 8%, and metal substrate damage 2% (these percentages are for illustrative purposes only).

[0053] S302, Assigning weighting coefficients: Different weighting coefficients are assigned to different damage depth levels to reflect the impact of depth differences on the overall damage severity.

[0054] For example, in practical operation, based on the principle of analyzing both the damaged area and the damaged depth of the paint film, the following coefficients are set: • Topcoat damage (shallow layer): Weighting factor 1.0 • Electrophoretic layer damage (middle layer): Weighting factor 2.0 • Damage to metal substrate (deep): Weighting factor 3.0 The coefficient values ​​are based on industry experience that the greater the depth, the greater the damage, but they are adjustable in practice and are not limited to this example.

[0055] S303, Calculate the weighted damage parameters: The area proportion of each damage level is combined with the corresponding weighting coefficient to obtain the weighted damage parameters.

[0056] For example, in actual operation, the calculation result is: Weighted parameter for topcoat damage: 15% × 1.0 = 0.15 Damage weighting parameter for electrophoretic layer: 8% × 2.0 = 0.16 Damage weighting parameter for metal substrate: 2% × 3.0 = 0.06 This step converts the area data into depth-weighted values, which quantifies the severity of the damage.

[0057] S304, Generate comprehensive damage severity parameters: By integrating the weighted parameters of all damage levels, a comprehensive damage severity parameter is generated through weighted summation or product combination.

[0058] For example, the calculated comprehensive damage parameter is 0.15 + 0.16 + 0.06 = 0.37.

[0059] This parameter integrates information on damage area and depth, providing a single, comparable quantitative indicator for subsequent rating.

[0060] It is understood that this embodiment integrates the damage area ratio and depth information by weighting coefficients to generate comprehensive damage degree parameters, thereby improving the scientificity and reliability of the coating's stone impact resistance test evaluation.

[0061] In one possible embodiment, step S4 includes sub-steps S401 to S403.

[0062] S401, Obtain the preset rating criteria: The system retrieves preset rating standards from the database. These rating standards are based on historical test data, industry standards (such as automotive coating stone chip resistance test standards), or experimental calibration definitions, and include rating levels or threshold ranges corresponding to different damage degree parameters.

[0063] For example, the rating criteria divide the damage severity parameter into three threshold ranges: 0-0.2: Corresponds to Grade A rating (minor damage). 0.2-0.5: Corresponds to Grade B rating (moderate damage). 0.5 or above: corresponding rating C (severe damage).

[0064] S402, Level Matching: The system uses a comparison algorithm (such as interval matching or fuzzy logic) to match and compare the calculated damage level parameter (such as 0.37) with the preset rating standard.

[0065] For example, if the algorithm detects that the damage level parameter (0.37) falls within the range [0.2, 0.5], it automatically assigns a rating of B. The matching process is completed in real time, requiring no manual intervention and avoiding subjective errors.

[0066] S403, Generate evaluation results: Based on the matching results, an evaluation result is automatically generated, which includes a numerical rating, category label, or visual report.

[0067] For example, the automatically generated evaluation results are represented as follows: Digital rating: Output "B" Category tag: Marked "Moderate damage, maintenance recommended" Visualized report: Displays tables on the PC or mobile app interface, including the original image (e.g.) Figure 3 Thresholding images (e.g.) Figure 4) and rating results (e.g. Figure 5 (See the example table). The results can be output to a user interface (such as a screen display) or a storage system (such as a database) to support subsequent tracing and analysis.

[0068] Understandably, this embodiment achieves intelligent rating of coating stone chip resistance test results through an automated process. First, it acquires preset rating standards (such as threshold ranges or grade mappings) defined based on historical data, industry standards, or experimental calibration. Then, it uses a comparison algorithm to match the calculated damage degree parameters with the standards, determining the rating range or grade to which the parameters belong. Finally, based on the matching results, it automatically generates diverse evaluation results, including digital ratings, classification labels, or visual reports. Its technical effect is to completely eliminate the subjectivity and inconsistency of manual evaluation, improve the objectivity, accuracy, and reproducibility of the rating, significantly improve detection efficiency through a fully automated comparison and output process, ensure standardized and traceable results, and provide efficient and reliable decision support for coating performance evaluation.

[0069] In one possible embodiment, step S4 further includes: S404, to verify or round the evaluation results.

[0070] For example, in practice, the evaluation results are processed as follows: 1. Verification Processing: The system compares the current evaluation results with similar sample data in the historical database. If a significant deviation is found (such as damage parameters being abnormally higher than the historical average), the system automatically triggers a review mechanism to re-examine the accuracy of each step of the image processing. 2. Rounding: Numerical results are formatted according to preset rules. For example, the percentage of damaged area of ​​0.1567 is rounded to 0.16 (retaining two decimal places), and the rating parameter of 0.37 is rounded to 0.4 (retaining one decimal place), to ensure that the data conforms to the industry standard presentation specifications. 3. Logical verification: Check the logical consistency between the rating results and image features. If a contradiction is found, such as "the proportion of damaged area of ​​metal substrate is high but the rating is minor damage", an abnormality mark will be automatically marked for manual review.

[0071] Figure 6 A structural diagram of an AI-assisted coating stone impact resistance testing and evaluation system provided in this embodiment of the invention is shown below. Figure 6 As shown, an AI-assisted coating stone impact resistance testing and evaluation system includes an image acquisition module, an image processing module, a parameter calculation module, and a rating module, wherein: The image acquisition module is used to acquire sample images of the coating after the stone impact resistance test under standardized conditions. The image processing module is used to analyze the sample image through image processing algorithms to identify the damaged area and distinguish different levels of damage depth. The image processing module is configured to analyze the sample image using a thresholding algorithm, dividing the pixel grayscale values ​​into multiple discrete intervals, each interval being mapped to a specific damage depth level, so as to achieve automatic identification and depth differentiation of the damage area. The parameter calculation module is used to calculate the damage degree parameters based on the area ratio and damage depth information of the damaged area; The rating module is used to compare the damage degree parameters with preset rating standards and generate automated evaluation results.

[0072] It is understood that the AI-assisted coating stone impact resistance testing and evaluation system provided by the present invention corresponds to the AI-assisted coating stone impact resistance testing and evaluation method provided in the foregoing embodiments. The relevant technical features of the AI-assisted coating stone impact resistance testing and evaluation system can be referred to the relevant technical features of the AI-assisted coating stone impact resistance testing and evaluation method, and will not be repeated here.

[0073] Please see Figure 7 , Figure 7 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 7 As shown, this embodiment of the invention provides an electronic device 700, including a memory 710, a processor 720, and a computer program 711 stored in the memory 710 and executable on the processor 720. When the processor 720 executes the computer program 711, it performs the following steps: S1, Obtain sample images after the stone impact resistance test of the coating; S2, The sample image is analyzed using an image processing algorithm to identify the damaged area and distinguish different levels of damage depth; S3, Calculate the damage degree parameters based on the area ratio and damage depth information of the damaged area; S4. The damage degree parameters are compared with preset rating standards to generate automated evaluation results.

[0074] Please see Figure 8 , Figure 8 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 8 As shown, this embodiment provides a computer-readable storage medium 800, on which a computer program 711 is stored. When the computer program 711 is executed by a processor, it performs the following steps: S1, Obtain sample images after the stone impact resistance test of the coating; S2, The sample image is analyzed using an image processing algorithm to identify the damaged area and distinguish different levels of damage depth; S3, Calculate the damage degree parameters based on the area ratio and damage depth information of the damaged area; S4. The damage degree parameters are compared with preset rating standards to generate automated evaluation results.

[0075] This invention provides an AI-assisted method, system, and storage medium for evaluating the stone impact resistance of coatings. It automates the evaluation of stone impact resistance testing through AI-assisted image processing and analysis. First, standardized image acquisition is used to obtain post-test sample images. Then, image processing algorithms (such as thresholding and region segmentation) are used to intelligently identify damaged areas and accurately distinguish different damage depth levels (such as damage to topcoat, electrophoretic layer, and metal substrate). Subsequently, a comprehensive damage degree parameter is calculated based on the damage area ratio and depth information. Finally, an automated evaluation result (such as a digital rating or visual report) is generated by comparing the result with a preset rating standard. This invention eliminates the subjectivity and reproducibility issues of traditional manual evaluation. Full-process automation improves the objectivity, accuracy, and efficiency of the evaluation. Simultaneously, multi-dimensional parameter quantification (area and depth fusion) ensures the scientific reliability of the evaluation results, providing consistent and traceable decision support for coating performance evaluation and significantly enhancing the standardization and practicality of the testing process.

[0076] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0077] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0078] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0079] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0080] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0081] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0082] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An AI-assisted based coating stone chip test evaluation method, characterized in that, The method comprises the following steps: S1, acquiring a sample image after a stone chip test of a coating; S2, analyzing the sample image by an image processing algorithm to identify a damage area and distinguish different damage depth levels; S3, calculating a damage degree parameter based on the area proportion of the damage area and damage depth information; S4, comparing the damage degree parameter with a preset rating standard to generate an automatic evaluation result.

2. The AI-assisted coating stone impact test evaluation method of claim 1, wherein, Step S1 comprises: S101, positioning a coating sample plate after a stone chip test by a standardized fixing method, correcting the surface orientation of the sample plate by using a grid identification to eliminate the deviation of the collection angle and deformation error; S102, adjusting the illumination parameters in a diffuse lighting environment to make the light uniformly distributed on the surface of the sample plate to generate an original image with a gray scale distribution associated with damage depth; S103, collecting digital image data of the sample surface, and synchronously implementing image quality checking to trigger a re-collection mechanism for images that do not meet the definition or exposure standard; S104, saving the sample image that passes the checking.

3. The AI-assisted coating stone impact test evaluation method of claim 1, wherein, In step S2, the analysis of the sample image by the image processing algorithm to identify the damage area comprises: S201, dividing the pixel gray scale values in the sample image into multiple discrete intervals by a threshold processing algorithm, each interval being associated with a specific damage depth level to achieve preliminary differentiation of the damage area; S202, identifying the contour of the damage area by a region segmentation algorithm to segment the damage area; S203, classifying the segmented damage area according to the depth level to determine the depth information of the damage area according to the pixel distribution or rule matching; S204, extracting the geometric feature parameters of the damage area and generating a comprehensive feature vector in combination with the depth level.

4. The AI-assisted coating stone impact test evaluation method of claim 3, wherein, Before step S201, it further comprises: preprocessing the sample image, the preprocessing comprising at least one of image scanning, noise elimination and resolution standardization.

5. The AI-assisted coating stone impact test evaluation method of claim 3, wherein, The determination of the depth information of the damage area according to the pixel distribution or rule matching comprises: analyzing the pixel gray scale values to match the pixel distribution in the sample image with a preset rule, wherein the preset rule defines the mapping relationship between the gray scale interval and the damage depth level; based on the pixel matching result, classifying each pixel into the corresponding damage depth level, the damage depth level at least including face paint damage, electrophoretic layer damage and metal substrate damage; according to the pixel classification result, counting the area distribution of each damage depth level to determine the depth information of the sample image.

6. The AI-assisted coating stone impact test evaluation method of claim 1, wherein, Step S3 comprises: S301, based on the image analysis result, counting the proportion of the pixel number of each depth level area to the total area pixel number to obtain the area proportion of each damage depth level; S302, assigning a corresponding weighting coefficient to each damage depth level; S303, combining and calculating the area proportion of each damage level with the corresponding weighting coefficient to obtain a weighted damage parameter; S304, integrating the weighted parameters of all damage levels to generate a comprehensive damage degree parameter by weighted summation or product combination.

7. The AI-assisted coating stone impact test evaluation method of claim 1, wherein, Step S4 comprises: S401, obtain a preset rating standard, the rating standard is defined based on historical test data, industry specifications or experimental calibration, including rating levels or threshold intervals corresponding to different damage degree parameters; S402, match the calculated damage degree parameter with the preset rating standard, and determine the rating interval or level to which the parameter belongs through an algorithm; S403, automatically generate an evaluation result based on the matching result, the evaluation result including a numerical rating, a classification label or a visual report.

8. The AI-assisted coating stone impact test evaluation method of claim 7, wherein, Step S4 further includes: S404, verifying or rounding the evaluation result.

9. An AI-assisted based coating stone chip test evaluation system, characterized in that, Including: An image acquisition module for acquiring sample images after the coating stone chip resistance test under standardized conditions; An image processing module for analyzing the sample images through an image processing algorithm to identify damage areas and distinguish different damage depth levels; A parameter calculation module for calculating damage degree parameters based on the area proportion of the damage area and damage depth information; A rating module for comparing the damage degree parameters with a preset rating standard to generate an automated evaluation result.

10. The AI-assisted coating stone impact test evaluation system of claim 9, wherein, The image processing module is configured to analyze the sample images through a threshold processing algorithm, divide pixel gray values into multiple discrete intervals, and map each interval to a specific damage depth level to achieve automatic identification and depth distinction of damage areas.

Images