Image-based green strength prediction method and system for ceramic grinding wheel and storage medium

Abstract

This invention relates to an image-based method, system, and storage medium for predicting the strength of ceramic grinding wheel green bodies. The method includes the following steps: acquiring an original image of the surface of the ceramic grinding wheel green body to be tested; identifying and extracting the formula identification information of the ceramic grinding wheel green body from the coded region of the original digital image; calling a strength prediction model pre-associated with the ceramic grinding wheel green body; processing the original image to segment and extract the abrasive particle region and the pore region in the original image; calculating a uniformity index characterizing the uniformity of abrasive distribution based on the binarized image of the abrasive particle region; calculating a defect density index characterizing the density of pores based on the binarized image of the pore region; fusing the uniformity index and the defect density index to generate green body structural feature parameter information; and inputting the green body structural feature parameter information into the strength prediction model for calculation to output the predicted strength value of the ceramic grinding wheel green body.

Description

Technical Field

[0001] This invention relates to the field of ceramic grinding wheel green strength testing technology, specifically to an image-based method, system, and storage medium for predicting the strength of ceramic grinding wheel greens. Background Technology

[0002] Ceramic grinding wheels are bonded abrasives made primarily from abrasive materials (such as corundum and silicon carbide) and ceramic binders, through mixing, molding, drying, and high-temperature sintering. They are widely used in the machining field. In the stage after green body forming and before sintering, the green body has relatively low green strength. Sufficient green strength is crucial for ensuring the smooth operation of subsequent handling, kiln loading, and other processes, as well as the final product yield. If the green body strength is insufficient, it is highly susceptible to edge chipping, breakage, or even complete fracture during handling, resulting in direct economic losses.

[0003] Because ceramic grinding wheel blanks are fragile and have extremely low strength, they are difficult to effectively and non-destructively test and quantify using conventional mechanical methods. Therefore, there is currently no unified standard in the industry for testing and characterizing the strength of ceramic grinding wheel blanks. Traditional assessment methods mainly rely on the experience and intuition of workers, which are highly subjective and have poor repeatability. While some preliminary quantitative attempts can obtain strength reference values, they can only be used for sampling inspection and cannot achieve full, non-destructive testing of every blank on the production line. More importantly, these methods can only provide a macroscopic fracture strength result and cannot reveal the microscopic structural root causes that lead to that strength value, such as whether the abrasive particles are evenly distributed, whether the binder is fully filled, or whether there are harmful large-sized pores. Although machine vision technology has been widely used in industrial quality inspection, there are also existing technologies for the inspection of finished grinding wheels, such as using image recognition to identify macroscopic cracks on the surface of the finished product or analyzing the abrasive morphology of the working surface of the grinding wheel after use. However, these technologies focus on the surface of sintered finished products or in their used state. There is no mature technical solution that can perform non-destructive and rapid quantitative assessment of the internal structural uniformity of unsintered ceramic green bodies in this special stage, and establish a reliable predictive relationship between it and macroscopic mechanical strength.

[0004] This application relates to the development of an image-based method, system, and storage medium for predicting the strength of green ceramic grinding wheels, in order to solve the aforementioned problems. Summary of the Invention

[0005] To achieve the above-mentioned objectives and other advantages of the present invention, a first objective of the present invention is to provide an image-based method for predicting the strength of ceramic grinding wheel green bodies, comprising the following steps: Acquire raw images of the surface of the green ceramic grinding wheel to be tested; wherein, the raw digital images include coded areas for identifying the formula, abrasive particle areas reflecting the internal composition, and pore areas characterizing structural defects; The formula identification information of the ceramic grinding wheel green body is identified and extracted from the coded region of the original digital image; The strength prediction model pre-associated with the ceramic grinding wheel green body is invoked based on the formula identification information; The original image is processed to segment and extract the abrasive particle region and the pore region in the original image; Based on the binarized image of the abrasive particle region, a uniformity index characterizing the uniformity of the abrasive distribution is calculated. Based on the binarized image of the pore region, a defect density index characterizing the density of the pores is calculated. The uniformity index and the defect density index are fused to generate green body structural characteristic parameter information; The structural characteristic parameters of the green body are input into the strength prediction model for calculation, so as to output the strength prediction value of the ceramic grinding wheel green body.

[0006] Furthermore, the construction of the intensity prediction model includes the following steps: Selecting a variety of different ceramic grinding wheel formulations; According to the formula, multiple standard green test strips with the same formula and process parameters were prepared. Acquire standard digital images of the surface of each of the aforementioned standard green test strips; Obtain the green structure characteristic parameter information of the standard green test strip based on the standard digital image; Each of the standard green test strips is subjected to a destructive mechanical strength test to obtain the measured strength value of the standard green test strip; A training dataset is constructed based on the green structure characteristic parameters and measured strength values ​​of all the standard green test strips for each formulation. Using the training dataset, a regression analysis algorithm is used to train the intensity prediction model for each formula. The strength prediction model corresponding to each formula is associated with its formula identification information and stored in the model database.

[0007] Furthermore, each of the aforementioned standard green test strips is subjected to a destructive mechanical strength test, including the following steps: Place the standard green billet horizontally on the support platform of the testing equipment, aligning one end of the billet with the edge of the platform; and ensure that one side of the standard green billet contacts the pushing mechanism of the testing equipment. A constraint force perpendicular to the support platform is applied to the standard green test strip by the pressing component; The propulsion mechanism is driven to push the standard green test strip at a constant speed until the test strip breaks. Measure and record the length of the remaining portion of the test strip that remains on the support platform after breakage; Determine whether the length meets the preset threshold length; When the length meets the threshold length, the ratio of the remaining length to the original total length of the test strip is calculated and used as the measured strength value of the standard green test strip; When the length does not meet the threshold length, a new standard green test strip that meets the preset specifications is prepared or selected, and the destructive mechanical strength test steps are repeated.

[0008] Furthermore, the preset threshold length is any value between one-third and two-thirds of the original total length of the test strip.

[0009] Furthermore, the regression analysis algorithm includes at least one of linear regression, multinomial regression, support vector machine regression, or neural network regression algorithms.

[0010] Further, the step of identifying and extracting the formula identification information of the ceramic grinding wheel green body from the coded region of the original digital image includes the following steps: Identify and locate the coded region in the original digital image; The encoded region is preprocessed to obtain a preprocessed encoded region image; The initial identification information in the preprocessed encoded region image is obtained using character recognition or graphic code recognition technology. The initial identification information is matched with the pre-stored model database to obtain the formula identification information.

[0011] Further, the original image is processed to segment and extract the abrasive particle region and pore region from the original image, including the following steps: The original image is converted to grayscale to obtain a grayscale image; The grayscale image is filtered to obtain a filtered image; A threshold-based image segmentation algorithm is used to divide the filtered image into an abrasive particle region, a pore region, and a background region.

[0012] Furthermore, the step of employing a threshold-based image segmentation algorithm to divide the filtered image into abrasive particle regions, pore regions, and background regions includes the following steps: The pixel regions in the filtered image whose grayscale values ​​are higher than a first preset threshold are identified as abrasive particle regions. The connected regions in the filtered image whose gray values ​​are lower than the second preset threshold and whose area is greater than the preset minimum area are identified as pore regions. The remaining areas in the filtered image that do not meet the criteria for determining the abrasive particle area and the pore area are determined as background areas.

[0013] Further, the calculation of the uniformity index characterizing the uniformity of the abrasive distribution includes the following steps: The image of the abrasive particle region is divided into multiple grid units; Calculate the area ratio of the abrasive particle region in each of the grid cells; Calculate the statistical coefficient of variation of the area ratio based on the area ratio of all the grid cells; The reciprocal of the statistical coefficient of variation is used as the uniformity index.

[0014] Furthermore, the calculation of the defect density index characterizing the porosity includes the following steps: Calculate the total number and total area of ​​the pore regions; The ratio of the total area of ​​the pore region to the total area of ​​the original digital image region is defined as the area density; The ratio of the total number of pore regions to the total area of ​​the original digital image region is defined as the number density; The defect density index is obtained by weighted summation of the area density and the number density.

[0015] Furthermore, the process of fusing the uniformity index and the defect density index to generate green body structural characteristic parameter information includes the following steps: The uniformity index and the defect density index are respectively normalized. Assign preset weighting coefficients to the normalized uniformity index and the defect density index; The weighted product of the uniformity index and the defect density index is calculated according to the preset weighting coefficients to generate the green body structural characteristic parameter information.

[0016] A second objective of this invention is to provide an image-based system for predicting the strength of green ceramic grinding wheels, comprising the following modules: An image acquisition module is used to acquire raw digital images of the surface of the green ceramic grinding wheel to be inspected; wherein, the raw digital image includes a coding area for identifying the formula, an abrasive particle area reflecting the internal composition, and a pore area characterizing structural defects; The image processing module is used to identify and extract the formula identification information of the ceramic grinding wheel green body from the coded area of ​​the original digital image; call the strength prediction model pre-associated with the ceramic grinding wheel green body according to the formula identification information; process the original image, segment and extract the abrasive particle area and the pore area in the original image; The data processing module is used to calculate a uniformity index representing the uniformity of the abrasive distribution based on the binarized image of the abrasive particle region; calculate a defect density index representing the density of the pores based on the binarized image of the pore region; and fuse the uniformity index and the defect density index to generate green body structural feature parameter information. The strength prediction module is used to input the structural characteristic parameter information of the green body into the strength prediction model for calculation, so as to output the strength prediction value of the ceramic grinding wheel green body.

[0017] Furthermore, it also includes the following modules: The model building module is used to select various ceramic grinding wheel formulations; prepare multiple standard green body test strips with the same formulation and process parameters according to the formulations; acquire standard digital images of the surface of each standard green body test strip; obtain the green body structural feature parameter information of the standard green body test strip based on the standard digital images; perform destructive mechanical strength tests on each standard green body test strip to obtain the measured strength value of the standard green body test strip; and construct a training dataset based on the green body structural feature parameter information and measured strength values ​​of all standard green body test strips under each formulation. Using the training dataset, a regression analysis algorithm is used to train and obtain the intensity prediction model corresponding to each formulation; the intensity prediction model corresponding to each formulation is associated with its formulation identification information and stored in the model database.

[0018] A third objective of this invention is to provide a readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements an image-based method for predicting the strength of ceramic grinding wheel green bodies.

[0019] Compared with the prior art, the beneficial effects of the present invention are: This invention relates to an image-based method, system, and storage medium for predicting the strength of ceramic grinding wheel green bodies. By combining multi-source image information such as formula identifiers, abrasive distribution, and porosity defects on the green body surface through image acquisition, a strength prediction model is constructed, incorporating image feature extraction and formula correlation prediction. By calling this strength detection model, rapid, non-destructive, and accurate assessment of green body strength under different formulas and process conditions is achieved. In particular, the dual-feature fusion architecture, including abrasive distribution uniformity and porosity defect density indices, generates comprehensive green body structural feature parameters, effectively quantifying key internal structural factors affecting green body strength and significantly improving the accuracy and process adaptability of strength prediction. This effectively overcomes the problems of missed and false judgments in traditional detection methods caused by formula differences, uneven abrasive distribution, and porosity aggregation. It also solves the problem of inaccurate strength assessment caused by the inability to non-destructively quantify internal structure in existing technologies, achieving real-time, online, and reliable detection of green body strength, providing strong support for quality control and process optimization in ceramic grinding wheel green body production.

[0020] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it according to the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Specific embodiments of the present invention are given in detail in the following embodiments and their accompanying drawings. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart of the image-based method for predicting the green strength of ceramic grinding wheels according to this application; Figure 2 This is a flowchart illustrating the process of identifying and extracting the formula identification information of the ceramic grinding wheel green body as described in Example 1. Figure 3 The flowchart for constructing the intensity prediction model described in Example 1 is shown below; Figure 4 This is a flowchart of the destructive mechanical strength test performed on each of the standard green test strips as described in Example 1; Figure 5 This is a schematic diagram of the dedicated testing equipment described in Example 1; Figure 6 This is a schematic diagram of the ceramic grinding wheel green body described in Example 1; Figure 7 This is a flowchart illustrating the process of segmenting and extracting the abrasive particle region and pore region from the original image as described in Example 1. Figure 8This is a flowchart illustrating the calculation of the uniformity index characterizing the uniformity of the abrasive distribution as described in Example 1; Figure 9 This is a flowchart illustrating the calculation of the defect density index characterizing the porosity as described in Example 1; Figure 10 This is a flowchart illustrating the generation of green body structural feature parameter information as described in Example 1; Figure 11 This is a schematic diagram of the image-based ceramic grinding wheel green strength prediction system in Example 2. Figure 1 ; Figure 12 This is a schematic diagram of the image-based ceramic grinding wheel green strength prediction system in Example 2. Figure 2 ; Figure 13 This is a schematic diagram of a computer-readable storage medium in Example 3; In the diagram: 100, controller; 110, control circuit; 120, control switch; 130, electric propeller; 140, propulsion mechanism; 150, holding component; 160, support platform. Detailed Implementation

[0022] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0023] In the following description, the use of suffixes such as "module," "part," or "unit" to denote elements is solely for the purpose of illustrative purposes and has no specific meaning in itself. Therefore, "module," "part," or "unit" may be used interchangeably.

[0024] Example 1 This invention provides an image-based method for predicting the strength of ceramic grinding wheel green bodies, such as... Figure 1 As shown, the specific steps include the following: S101, Acquire the original image of the surface of the green ceramic grinding wheel to be tested; wherein, the original digital image includes a coding area for identifying the formula, an abrasive particle area reflecting the internal composition, and a pore area characterizing structural defects; S102, Identify and extract the formula identification information of the ceramic grinding wheel green body from the coded area of ​​the original digital image; S103, The strength prediction model pre-associated with the ceramic grinding wheel green body is invoked according to the formula identification information; S104, Process the original image to segment and extract the abrasive particle region and pore region in the original image; S105, Based on the binarized image of the abrasive particle region, calculate the uniformity index characterizing the uniformity of the abrasive distribution; S106, Based on the binarized image of the pore region, calculate the defect density index that characterizes the density of the pores; S107, The uniformity index and the defect density index are fused to generate green body structural characteristic parameter information; S108, the structural characteristic parameter information of the green body is input into the strength prediction model for calculation, so as to output the strength prediction value of the ceramic grinding wheel green body.

[0025] In some embodiments, the predicted strength value output in step S108 is represented by a hardness grade code, which corresponds to a strength ratio measured by a destructive mechanical test. The strength ratio is the ratio of the average remaining length of the ceramic grinding wheel green specimen after fracture to the original total length; the lower the ratio, the higher the actual strength of the ceramic grinding wheel green specimen.

[0026] For example, taking the WA-F60 formulation as an example, the correspondence between each hardness grade code and the strength ratio is shown in Table 1. Code H corresponds to a strength ratio of 0.60–0.62, representing the lowest strength of the ceramic grinding wheel green body; code S corresponds to a strength ratio of 0.38–0.40, representing the highest strength of the ceramic grinding wheel green body. From H to S, the strength ratio decreases progressively, indicating that the strength of the ceramic grinding wheel green body increases sequentially. For example, if the strength prediction model outputs code K, it indicates that the predicted strength of the ceramic grinding wheel green body is in the lower-middle range, corresponding to a strength ratio of 0.55–0.57; if it outputs code R, it indicates that its predicted strength is relatively high, corresponding to a strength ratio of 0.40–0.43.

[0027] Through the above correspondence, this application realizes the mapping from image features to strength grades with clear process meanings, providing a reliable basis for rapid, non-destructive strength evaluation and grading of ceramic grinding wheel green bodies.

[0028] Table 1. Predicted Strength Values ​​of WA-F60 Ceramic Grinding Wheel Formulation hardness Remaining length of test strip (average / total length) H 0.60-0.62 J 0.57-0.60 K 0.55-0.57 L 0.52-0.55 M 0.50-0.52 N 0.48-0.50 P 0.45-0.48 Q 0.43-0.45 R 0.40-0.43 S 0.38-0.40 In some embodiments, the original digital image in step S101 is acquired by an image acquisition device, which may include one or more of a resolution scanner, a line scan camera, an industrial camera, a near-infrared imager, or a polarized light imaging device.

[0029] In a preferred embodiment, the image acquisition device involved in this application is an industrial camera configured with a uniform surface light source and a fixed-focus lens for acquisition; wherein, the industrial camera is configured to place the observation surface of the green blank to be inspected within its depth of field, and the uniform surface light source illuminates the observation surface at a preset angle and light intensity to ensure that the coded area, the abrasive particle area and the pore area have clear outlines and sufficient contrast in the original digital image.

[0030] Specifically, it should be understood that this application acquires visual information that comprehensively reflects the internal structural quality of the ceramic grinding wheel green body under inspection through a non-contact image acquisition method in a single shot. An industrial camera equipped with a uniform surface light source and a fixed-focus lens is used to capture images of the green body surface placed at the inspection station, obtaining high-resolution raw digital images. During image acquisition, the angle and intensity of the light source need to be adjusted to ensure uniform overall illumination without strong reflections or shadows, thus clearly presenting three key areas simultaneously: the coding area, typically text, barcodes, or QR codes printed or affixed to the green body surface containing formula and model information; the abrasive particle area, which appears in the image as a collection of dots or clumps of varying shapes and colors, the distribution and morphology of which directly reflect the mixing and forming state of the abrasive and binder; and the pore area, which appears in the image as pore-like features, the size, number, and distribution of which are crucial for assessing internal defects in the green body. By simultaneously capturing these three types of information in a single image, a complete and consistent raw data foundation is provided for subsequent formula identification, structural quantification analysis, and strength prediction.

[0031] In some embodiments, the process of identifying and extracting the formula identification information of the ceramic grinding wheel green body from the coded region of the original digital image in step S102, such as... Figure 2 As shown, it includes the following steps: S1021, Identify and locate the encoded region in the original digital image; S1022, Perform image preprocessing on the encoded region to obtain a preprocessed encoded region image; S1023, Initial identification information in the preprocessed encoded region image is obtained using character recognition or graphic code recognition technology; S1024, Match the initial identification information with the pre-stored model database to obtain the formula identification information.

[0032] In a preferred embodiment, step S1021 employs a feature-matching-based algorithm to determine the location of the coded region in the image. Specifically, a template image of the standard coded region is pre-stored, visual features are extracted from the original digital image, and a matching calculation is performed between the visual features and those of the template image. By evaluating the matching degree, the region in the original image most similar to the template is located, thereby determining the image coordinate range of the coded region. This method can still achieve accurate and automatic positioning of the coded region even when there is a certain deviation in the placement position of the raw material.

[0033] In a preferred embodiment, the image preprocessing in step S1022 includes grayscale conversion, noise suppression using Gaussian filtering or median filtering, and contrast enhancement and binarization processing using histogram equalization or adaptive threshold segmentation, in order to improve the separation of the identification features and the background in the image, eliminate interference introduced during the imaging process, and provide a clear input image for subsequent recognition steps.

[0034] In a preferred embodiment, step S1023 calls the corresponding recognition module based on the representation of the encoded region. Specifically, if the encoding is in character form, an optical character recognition module is used to convert the character sequence in the image into a text string; if the encoding is in graphic code form, a graphic code decoding module is called to parse the encoded data string to obtain the initial identification information in the preprocessed encoded region image.

[0035] In a preferred embodiment, step S1024, which involves matching the initial identification information with a pre-stored model database, specifically includes data cleaning and verification processes in the database. When a unique and valid formula record is successfully matched, the formula identification information is acquired and output for subsequent processes. If the matching fails, exception handling is performed according to preset rules.

[0036] In some embodiments, the intensity prediction model construction described in step S103 is as follows: Figure 3 As shown, it includes the following steps: S1031, selects a variety of different ceramic grinding wheel formulas; S1032, according to the formula, prepare multiple standard green test strips with the same formula and process parameters; S1033, Acquire a standard digital image of the surface of each of the standard green test strips; S1034, Obtain the green structure characteristic parameter information of the standard green test strip based on the standard digital image; S1035, each of the standard green test strips is subjected to a destructive mechanical strength test to obtain the measured strength value of the standard green test strip; S1036, Based on the green structure characteristic parameter information and measured strength value of all the standard green test strips under each formula, a training dataset is constructed; S1037, Using the training dataset, a regression analysis algorithm is used to train the model to obtain the intensity prediction model corresponding to each formula. S1038, The strength prediction model corresponding to each formula is associated with its formula identification information and stored in the model database.

[0037] In a preferred embodiment, the size of the standard green test strip in step S1032 can be selected and is not limited to 200mm long × 30mm wide × 20mm high, etc. According to the test results, the height of the test strip is a variable range that can be adjusted in 1mm increments, thereby classifying the strength of the ceramic grinding wheel green body into different grades.

[0038] In a preferred embodiment, the regression analysis algorithm in step S1037 includes at least one of linear regression, multinomial regression, support vector machine regression, or neural network regression algorithms.

[0039] In a preferred embodiment, step S1035 involves performing a destructive mechanical strength test on each of the standard green test strips, such as... Figure 4 As shown, it includes the following steps: S10351, The standard green billet test strip is placed horizontally on the support platform of the testing equipment, with one end of the test strip aligned with the edge of the platform; one side of the standard green billet test strip is brought into contact with the pushing mechanism of the testing equipment. S10352, A constraint force perpendicular to the support platform is applied to the standard green test strip by the pressing component; S10353, drive the propulsion mechanism to push the standard green test strip at a constant speed until the test strip breaks; S10354, Measure and record the length of the remaining portion of the test strip that remains on the support platform after breakage; S10355, Determine whether the length meets the preset threshold length; S10356a, when the length meets the threshold length, calculate the ratio of the remaining length to the original total length of the test strip, and use it as the measured strength value of the standard green test strip; S10356b, when the length does not meet the threshold length, the standard green test strip that meets the preset specifications is re-prepared or selected, and the destructive mechanical strength test steps are repeated.

[0040] In a preferred embodiment, the dedicated detection device mentioned in step S10351, such as... Figure 5 As shown, it includes: The controller 100 is electrically connected to the control line 110; Control switch 120 connected to the control circuit 110; An electric thruster 130 driven by the controller 100; The propulsion mechanism 140 is connected to the electric thruster 130 in a driving manner; Holding component 150; Support platform 160 for supporting the standard green test strip; The pressing component 150 is disposed above the support platform 160 and is used to press the test strip; the controller 100 receives the signal from the control switch 120 through the control line 110 and controls the electric pusher 130 to drive the push mechanism 140 to perform linear motion, so as to push the standard green test strip placed on the support platform 160.

[0041] In a preferred embodiment, the preset threshold length is any value between one-third and two-thirds of the original total length of the test strip.

[0042] For example, this application selects three formulation examples for destructive testing to obtain the measured strength values ​​of the standard green test strips: Formula 1: Take 121kg F60 GC abrasive, 19.36kg ceramic binder, 1.57kg dextrin powder, and 3.6kg dextrin liquid, and mix them according to the conventional ceramic grinding wheel forming material mixing process to form a ceramic grinding wheel green body as shown. Figure 6 As shown, the ceramic grinding wheel green body includes an abrasive particle region 200 and a pore region 300. A sample is taken from the molding material and pressed into five rectangular test strips with dimensions of 155mm×22mm×20mm according to a density of 1.97kg / cm3. The test strips are placed flat on the support platform, the pressing component is adjusted to be close to a section of the test strip, the switch is turned on, and the pushing mechanism pushes the test strip forward at a constant speed of 10mm / s until the test strip breaks. The length of the test strip remaining on the support platform is measured. The above operation is repeated, and five sets of data are measured and recorded.

[0043] Formula 2: Take 255 kg of F80 A abrasive, 30.6 kg of ceramic binder, 3.06 kg of dextrin powder, and 8.64 kg of water glass, and mix them according to the conventional ceramic grinding wheel forming material mixing process; take a portion of the sample from this forming material and press it into 5 rectangular test strips with a density of 2.16 kg / cm3 and a specification of 155 mm × 22 mm × 20 mm; place the test strips flat on the support platform, adjust the pressing component to fit the test strip tightly, start the switch, and push the test strip forward at a constant speed of 10 mm / s until the test strip breaks, and measure the length of the test strip remaining on the support platform; repeat the above operation, measure 5 sets of data, and record them.

[0044] Formula 3: Take 153.5 kg of F36 WA abrasive, 18.42 kg of ceramic binder, 2.76 kg of dextrin powder, and 2.76 kg of dextrin liquid, and mix them according to the conventional ceramic grinding wheel molding material mixing process; take a portion of the sample from this molding material and press it into 5 rectangular test strips with a density of 2.32 kg / cm3 and a specification of 155 mm × 22 mm × 20 mm; place the test strips flat on the support platform, adjust the pressing component to fit the test strip tightly, start the switch, and push the test strip forward at a constant speed of 10 mm / s until the test strip breaks, and measure the length of the test strip remaining on the support platform; repeat the above operation, measure 5 sets of data, and record them.

[0045] The strength measurement data of the ceramic grinding wheel green blanks prepared by the above three formulations are shown in Table 2.

[0046] Table 2. Measurement data of green strength of ceramic grinding wheels By analyzing Formulas 1, 2, and 3, the order of green strength of ceramic grinding wheels from highest to lowest is Formula 3 > Formula 2 > Formula 1. In other words, the smaller the value, the greater the green strength within the same range. This method solves the difficulty in measuring the green strength of ceramic grinding wheels, effectively quantifying it, and is of paramount importance for process optimization and formula improvement of ceramic grinding wheels.

[0047] This application systematically selects multiple formulations and prepares several standard test strips with strictly consistent process parameters for each formulation, establishing a dedicated database for each formulation. This ensures that the final trained strength prediction model not only has high prediction accuracy for specific formulations but also broad applicability to green body testing with different formulations, providing a unified and reliable technical foundation for mixed formulation testing on the production line. This application precisely correlates the green body structural feature parameters extracted from images with measured strength values. Specifically, in destructive mechanical strength testing, through standardized testing procedures and strict data validity judgment—requiring the remaining fracture length to be within a preset reasonable range (e.g., 1 / 3 to 2 / 3 of the total length)—invalid data caused by extreme defects in the test strips themselves or abnormal testing operations are effectively eliminated. This ensures that the data input into each model training set is authentic, comparable, and meaningful, fundamentally improving the reliability and generalization ability of the constructed model. Furthermore, since the model establishes an intrinsic mapping relationship between microstructural features (abrasive uniformity, porosity) and final strength, when the predicted strength is unqualified, its specific causes can be traced and quantitatively analyzed (e.g., whether it is due to uneven abrasive distribution or porosity defects). This provides direct and objective data guidance for the precise adjustment of upstream processes such as mixing, molding, and venting, and realizes closed-loop feedback and continuous optimization of the production process by quality inspection.

[0048] In some embodiments, the original image described in step S104 is processed to segment and extract the abrasive particle region and the pore region from the original image, such as... Figure 7 As shown, it includes the following steps: S1041, The original image is converted to grayscale to obtain a grayscale image; S1042, The grayscale image is filtered to obtain a filtered image; S1043, using a threshold-based image segmentation algorithm, the filtered image is divided into an abrasive particle region, a pore region, and a background region.

[0049] In a preferred embodiment, step S1043, which employs a threshold-based image segmentation algorithm to divide the filtered image into an abrasive particle region, a pore region, and a background region, includes the following steps: The pixel regions in the filtered image whose grayscale values ​​are higher than a first preset threshold are identified as abrasive particle regions. The connected regions in the filtered image whose gray values ​​are lower than the second preset threshold and whose area is greater than the preset minimum area are identified as pore regions. The remaining areas in the filtered image that do not meet the criteria for determining the abrasive particle area and the pore area are determined as background areas.

[0050] In some embodiments, the calculation of the uniformity index characterizing the uniformity of the abrasive distribution in step S105, such as... Figure 8 As shown, it includes the following steps: S1051, the image of the abrasive particle region is divided into multiple grid units; S1052, Calculate the area ratio of the abrasive particle region in each of the grid cells; S1053, Calculate the statistical coefficient of variation of the area ratio based on the area ratio of all the grid cells; S1054, the reciprocal of the statistical coefficient of variation is used as the uniformity index.

[0051] In some embodiments, the calculation of the defect density index characterizing the porosity in step S106, such as... Figure 9 As shown, it includes the following steps: S1061, Calculate the total number and total area of ​​the pore regions; S1062, the ratio of the total area of ​​the pore region to the total area of ​​the original digital image region is defined as the area density; S1063, the ratio of the total number of pore regions to the total area of ​​the original digital image region is defined as the number density; S1064, the area density and the number density are weighted and summed to obtain the defect density index.

[0052] In some embodiments, the fusion of the uniformity index and the defect density index in step S107 to generate green body structural characteristic parameter information, such as... Figure 10 As shown, it includes the following steps: S1071, The uniformity index and the defect density index are normalized respectively; S1072, assign preset weighting coefficients to the normalized uniformity index and the defect density index; S1073, calculate the weighted product of the uniformity index and the defect density index according to the preset weight coefficient to generate the green body structural characteristic parameter information.

[0053] This application first performs grayscale conversion and filtering preprocessing on the acquired raw images of the green body surface. Then, using multi-threshold segmentation technology, it accurately separates the abrasive particle region, the significant pore region, and the background region in the image. Based on this, for the abrasive region, the variation in its area distribution is statistically analyzed using a grid to calculate an index characterizing the uniformity of abrasive distribution; for the pore region, its area proportion and number density are combined to calculate an index characterizing the severity of pore defects. Finally, the above two indices reflecting different physical meanings are normalized and weighted to generate a comprehensive green body structural feature parameter, which serves as the direct input for the subsequent strength prediction model. This realizes a paradigm shift in the assessment of key structural features (abrasive uniformity and pore defects) inside the green body from traditional subjective and qualitative evaluation to objective and quantitative analysis. Through standardized image processing and statistical algorithms, the ineffable "texture" is transformed into data indicators that can be accurately measured and compared. This not only provides an unprecedentedly refined and digital evaluation standard for green body quality, but more importantly, the generated comprehensive structural characteristic parameters are deeply coupled with multiple physical factors affecting green body strength, laying a solid data foundation for establishing a high-precision and high-reliability intelligent strength prediction model.

[0054] Example 2 This invention provides an image-based system for predicting the strength of green ceramic grinding wheels, such as... Figure 11 As shown, it includes the following modules: An image acquisition module is used to acquire raw digital images of the surface of the green ceramic grinding wheel to be inspected; wherein, the raw digital image includes a coding area for identifying the formula, an abrasive particle area reflecting the internal composition, and a pore area characterizing structural defects; The image processing module is used to identify and extract the formula identification information of the ceramic grinding wheel green body from the coded area of ​​the original digital image; call the strength prediction model pre-associated with the ceramic grinding wheel green body according to the formula identification information; process the original image, segment and extract the abrasive particle area and the pore area in the original image; The data processing module is used to calculate a uniformity index representing the uniformity of the abrasive distribution based on the binarized image of the abrasive particle region; calculate a defect density index representing the density of the pores based on the binarized image of the pore region; and fuse the uniformity index and the defect density index to generate green body structural feature parameter information. The strength prediction module is used to input the structural characteristic parameter information of the green body into the strength prediction model for calculation, so as to output the strength prediction value of the ceramic grinding wheel green body.

[0055] In some embodiments, such as Figure 12 As shown, it also includes the following modules: The model building module is used to select various ceramic grinding wheel formulations; prepare multiple standard green body test strips with the same formulation and process parameters according to the formulations; acquire standard digital images of the surface of each standard green body test strip; obtain the green body structural feature parameter information of the standard green body test strip based on the standard digital images; perform destructive mechanical strength tests on each standard green body test strip to obtain the measured strength value of the standard green body test strip; and construct a training dataset based on the green body structural feature parameter information and measured strength values ​​of all standard green body test strips under each formulation. Using the training dataset, a regression analysis algorithm is used to train and obtain the intensity prediction model corresponding to each formulation; the intensity prediction model corresponding to each formulation is associated with its formulation identification information and stored in the model database.

[0056] This application acquires multi-feature images of the green body surface through an image acquisition module, establishes a strength prediction model library with adaptive formula capabilities using a model building module, and finally achieves online intelligent determination of green body strength through an integrated processing and prediction module. By synergistically fusing the abrasive distribution uniformity index and the porosity defect density index, the core structural factors affecting green body strength are effectively quantified. A non-contact image acquisition and analysis method is adopted, avoiding sample loss and efficiency bottlenecks caused by traditional destructive testing. Through modular and streamlined system design, the entire process from image acquisition and feature extraction to strength prediction is automated, significantly improving the objectivity, accuracy, and online quality inspection efficiency of green body strength testing. This provides reliable technical support for the stable control and process optimization of ceramic grinding wheel production, and achieves accurate prediction and rapid evaluation of ceramic grinding wheel green body strength.

[0057] Example 3 This invention also provides a computer-readable storage medium, such as... Figure 13 As shown, it stores program instructions, which, when executed, implement the image-based method for predicting the strength of ceramic grinding wheel green bodies as described in Embodiment 1 above.

[0058] The program instructions are stored in a computer-readable storage medium (such as a CD-ROM, USB flash drive, or portable hard drive) or on a network, and include several computer program instructions to cause a computing device (such as a personal computer, server, or network device) to execute the above-described method according to the embodiments of this application.

[0059] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention, and other modifications can be easily made by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and examples shown and described herein.

[0060] The apparatus, electronic device, and non-volatile computer storage medium and method provided in the embodiments of this specification are corresponding. Therefore, the apparatus, electronic device, and non-volatile computer storage medium also have similar beneficial technical effects as the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the corresponding apparatus, electronic device, and non-volatile computer storage medium will not be repeated here.

[0061] Those skilled in the art will also know that, besides implementing the controller in the form of purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller take the form of logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices included within it for implementing various functions can also be considered structures within that hardware component. Alternatively, the devices for implementing various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0062] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0063] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0064] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects.

[0065] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0066] 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.

[0067] 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.

[0068] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0069] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0070] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0071] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside on local and remote computer storage media, including storage devices.

[0072] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0073] The above description is merely an embodiment of this specification and is not intended to limit the scope of one or more embodiments of this specification. Various modifications and variations can be made to one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of one or more embodiments of this specification.

Claims

1. A method for predicting the strength of ceramic grinding wheel green blanks based on images, characterized in that, Specifically, the following steps are included: Acquire raw images of the surface of the green ceramic grinding wheel to be tested; wherein, the raw digital images include coded areas for identifying the formula, abrasive particle areas reflecting the internal composition, and pore areas characterizing structural defects; The formula identification information of the ceramic grinding wheel green body is identified and extracted from the coded region of the original digital image; The strength prediction model pre-associated with the ceramic grinding wheel green body is invoked based on the formula identification information; The original image is processed to segment and extract the abrasive particle region and the pore region in the original image; Based on the binarized image of the abrasive particle region, a uniformity index characterizing the uniformity of the abrasive distribution is calculated. Based on the binarized image of the pore region, a defect density index characterizing the density of the pores is calculated. The uniformity index and the defect density index are fused to generate green body structural characteristic parameter information; The structural characteristic parameters of the green body are input into the strength prediction model for calculation, so as to output the strength prediction value of the ceramic grinding wheel green body.

2. The image-based method for predicting the strength of ceramic grinding wheel green blanks according to claim 1, characterized in that, The intensity prediction model is constructed by the following steps: Selecting a variety of different ceramic grinding wheel formulations; According to the formula, multiple standard green test strips with the same formula and process parameters were prepared. Acquire standard digital images of the surface of each of the aforementioned standard green test strips; Obtain the green structure characteristic parameter information of the standard green test strip based on the standard digital image; Each of the standard green test strips is subjected to a destructive mechanical strength test to obtain the measured strength value of the standard green test strip; A training dataset is constructed based on the green structure characteristic parameters and measured strength values ​​of all the standard green test strips for each formulation. Using the training dataset, a regression analysis algorithm is used to train the intensity prediction model for each formula. The strength prediction model corresponding to each formula is associated with its formula identification information and stored in the model database.

3. The image-based method for predicting the strength of ceramic grinding wheel green bodies according to claim 2, characterized in that, The destructive mechanical strength test is performed on each of the aforementioned standard green test strips, including the following steps: Place the standard green billet horizontally on the support platform of the testing equipment, aligning one end of the billet with the edge of the platform; and ensure that one side of the standard green billet contacts the pushing mechanism of the testing equipment. A constraint force perpendicular to the support platform is applied to the standard green test strip by the pressing component; The propulsion mechanism is driven to push the standard green test strip at a constant speed until the test strip breaks. Measure and record the length of the remaining portion of the test strip that remains on the support platform after breakage; Determine whether the length meets the preset threshold length; When the length meets the threshold length, the ratio of the remaining length to the original total length of the test strip is calculated and used as the measured strength value of the standard green test strip; When the length does not meet the threshold length, a new standard green test strip that meets the preset specifications is prepared or selected, and the destructive mechanical strength test steps are repeated.

4. The image-based method for predicting the strength of ceramic grinding wheel green bodies according to claim 1, characterized in that, The original image is processed to segment and extract the abrasive particle region and pore region, including the following steps: The original image is converted to grayscale to obtain a grayscale image; The grayscale image is filtered to obtain a filtered image; A threshold-based image segmentation algorithm is used to divide the filtered image into an abrasive particle region, a pore region, and a background region.

5. The image-based method for predicting the strength of ceramic grinding wheel green bodies according to claim 1, characterized in that, The calculation of the uniformity index characterizing the uniformity of the abrasive distribution includes the following steps: The image of the abrasive particle region is divided into multiple grid units; Calculate the area ratio of the abrasive particle region in each of the grid cells; Calculate the statistical coefficient of variation of the area ratio based on the area ratio of all the grid cells; The reciprocal of the statistical coefficient of variation is used as the uniformity index.

6. The image-based method for predicting the strength of ceramic grinding wheel green bodies according to claim 1, characterized in that, The calculation of the defect density index, which characterizes the porosity, includes the following steps: Calculate the total number and total area of ​​the pore regions; The ratio of the total area of ​​the pore region to the total area of ​​the original digital image region is defined as the area density; The ratio of the total number of pore regions to the total area of ​​the original digital image region is defined as the number density; The defect density index is obtained by weighted summation of the area density and the number density.

7. The image-based method for predicting the strength of ceramic grinding wheel green bodies according to claim 1, characterized in that, The process of fusing the uniformity index and the defect density index to generate green body structural characteristic parameter information includes the following steps: The uniformity index and the defect density index are respectively normalized. Assign preset weighting coefficients to the normalized uniformity index and the defect density index; The weighted product of the uniformity index and the defect density index is calculated according to the preset weighting coefficients to generate the green body structural characteristic parameter information.

8. An image-based system for predicting the strength of green ceramic grinding wheels, characterized in that, Includes the following modules: An image acquisition module is used to acquire raw digital images of the surface of the green ceramic grinding wheel to be inspected; wherein, the raw digital image includes a coding area for identifying the formula, an abrasive particle area reflecting the internal composition, and a pore area characterizing structural defects; The image processing module is used to identify and extract the formula identification information of the ceramic grinding wheel green body from the coded area of ​​the original digital image; call the strength prediction model pre-associated with the ceramic grinding wheel green body according to the formula identification information; process the original image, segment and extract the abrasive particle area and the pore area in the original image; The data processing module is used to calculate a uniformity index representing the uniformity of the abrasive distribution based on the binarized image of the abrasive particle region; calculate a defect density index representing the density of the pores based on the binarized image of the pore region; and fuse the uniformity index and the defect density index to generate green body structural feature parameter information. The strength prediction module is used to input the structural characteristic parameter information of the green body into the strength prediction model for calculation, so as to output the strength prediction value of the ceramic grinding wheel green body.

9. The image-based ceramic grinding wheel green strength prediction system according to claim 8, characterized in that, It also includes the following modules: The model building module is used to select a variety of different ceramic grinding wheel formulations; according to the formulations, multiple standard green body test strips with the same formulation and process parameters are prepared; standard digital images of the surface of each standard green body test strip are acquired; and green body structural feature parameter information of the standard green body test strip is obtained based on the standard digital images. Each of the standard green test strips is subjected to a destructive mechanical strength test to obtain the measured strength value of the standard green test strip; A training dataset is constructed based on the green structure characteristic parameters and measured strength values ​​of all the standard green test strips for each formulation. Using the training dataset, a regression analysis algorithm is used to train and obtain the intensity prediction model corresponding to each formulation; the intensity prediction model corresponding to each formulation is associated with its formulation identification information and stored in the model database.

10. A readable storage medium having a computer program stored thereon, characterized in that, The computer program is executed by a processor using the image-based method for predicting the green strength of ceramic grinding wheels as described in any one of claims 1-7.

Images