Casting defect detection system based on data analysis

Abstract

The present invention relates to the field of defect detection technology, specifically to a casting defect detection system based on data analysis, the system comprising a surface defect analysis module, a defect size assessment module, an internal defect analysis module and a defect impact assessment module. The present invention improves the accuracy and efficiency of casting defect recognition by integrating image recognition with ultrasonic flaw detection data processing, accurately classifies defect categories by analyzing casting surface images, matches defect features with a preset defect library, and assesses defect sizes by analyzing defect edge contours and calculating defect areas, making defect detection more detailed and comprehensive, and uses ultrasonic flaw detection data to analyze the internal structure of castings, and can accurately assess the location and severity of internal defects. By comprehensively assessing the quality of castings, manufacturers are allowed to make timely repair or elimination decisions during the production process, thereby improving the efficiency of the production process and the final quality of casting products.

Description

Technical Field

[0001] The invention relates to the technical field of defect detection, and in particular to a casting defect detection system based on data analysis. Background Art

[0002] The field of defect detection technology involves the use of various automated and semi-automated methods to identify and classify defects in manufactured products. Defect detection technology is widely used in quality control processes to ensure that products meet design specifications and performance standards. The main technologies include image processing, machine learning, pattern recognition and signal processing technologies. Defect detection systems can effectively detect defects such as cracks, holes, deformations and embedded foreign objects on various materials and components, such as metals, plastics, textiles and semiconductors. Defect detection technology can significantly improve the speed and accuracy of detection, reduce the need for manual inspection, thereby reducing production costs and improving product quality.

[0003] Among them, the casting defect detection system is a system specially designed to detect defects in castings produced during the casting process. Casting defects include pores, slag inclusions, cold shuts and cracks. The system usually uses high-resolution cameras and image analysis software to detect and classify defects. Through automated detection, the system can quickly identify quality problems in castings, allowing manufacturers to repair or remove unqualified castings in time before further processing. The application of this technology not only improves production efficiency, but also helps to ensure the structural integrity and performance of the final product.

[0004] Traditional defect detection systems mainly rely on image processing and signal processing, and their accuracy and efficiency in dealing with complex or minor defects are still limited. These technologies are usually unable to effectively evaluate the specific impact of defects, such as the specific impact of defect size and depth on casting performance. This technical limitation will lead to reduced production efficiency in actual production, and a higher scrap rate or rework rate due to inaccurate defect judgment, thereby increasing production costs and affecting the overall quality and market competitiveness of products. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a casting defect detection system based on data analysis.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a casting defect detection system based on data analysis, the system comprising:

[0007] The surface defect analysis module is based on the casting surface image. Through image analysis, it identifies the surface defects of the parts. According to the surface defect characteristics of the parts, it evaluates the matching degree between the defect characteristics and the defects in the preset defect library, identifies the surface defect category, and obtains the surface defect category information.

[0008] The defect size assessment module identifies the edge contour of the surface defect based on the surface defect category information, calculates the area of ​​each surface defect, and assesses the size of the surface defect in combination with the shape of the surface defect to obtain surface defect size information;

[0009] The internal defect analysis module identifies the internal defects of the casting based on the ultrasonic flaw detection data of the casting, evaluates the internal defect depth index according to the depth of the internal defect and the thickness data of the casting, and evaluates the severity of the internal defect in combination with the location of the internal defect on the casting to obtain the internal defect severity information;

[0010] The defect impact assessment module assesses the overall impact of defects on castings based on the surface defect category information, surface defect size information and internal defect severity information, classifies castings by quality, and obtains casting defect detection results.

[0011] The present invention is improved in that the method for identifying surface defects of parts is:

[0012] Based on the casting surface image, pixel-level features are extracted, including grayscale value, edge intensity and texture directionality, to obtain pixel-level feature information;

[0013] Based on the pixel-level feature information, the formula is:

[0014] ,

[0015] Calculate the feature difference between the pixel and the surrounding pixels ,in, is the normalization factor, It is The weight of the dimension feature, is the exponential coefficient, Indicates The feature difference of each pixel, It is The pixel dimensional eigenvalue, It is The pixel The range average of the dimension feature;

[0016] Based on the feature difference , compared with the preset feature difference threshold, the pixel points exceeding the preset feature difference threshold are marked as defect points to obtain defect marking information.

[0017] The present invention is improved in that the step of obtaining the surface defect category information is:

[0018] Based on the defect marking information, the defect area is compared with the preset defect library, and the formula is:

[0019] ,

[0020] Calculate the similarity between the features of the defect area and the features in the defect library to obtain a similarity score ,in, is the similarity score, For the The weight of the feature, The defect area eigenvalues, The first eigenvalues, is the sensitivity coefficient, is the total number of features, is the base of natural logarithms;

[0021] Based on the similarity score , and compared with the preset similarity threshold, the defect types exceeding the preset similarity threshold are marked as corresponding categories to obtain surface defect category information.

[0022] The present invention is improved in that the method for calculating the area of ​​each surface defect is:

[0023] Based on the surface defect category information, the formula:

[0024] ,

[0025] Calculate the edge intensity response of a defect , the edge intensity response Compare with the preset edge threshold, mark the pixel points exceeding the preset edge threshold as defective edges, and obtain defective edge information, where: It is at coordinates The edge intensity response at and The images are and Intensity gradient in direction;

[0026] Based on the defect edge information, the formula:

[0027] ,

[0028] Calculate defect area , get the defect area information, where represents the defect area, For the The number of pixels in the defect area, is the total number of defect areas of the same type, is the image scaling factor, It is The shape complexity parameter of the defect area.

[0029] The present invention is improved in that the step of obtaining the surface defect size information is:

[0030] Based on the defect area , combined with the shape complexity of surface defects, through the formula:

[0031] ,

[0032] Calculate the rating of surface defect size ,in, is the defect area, is the average shape complexity parameter of surface defects, and is the adjustment parameter, is a rating of the size of surface defects;

[0033] Rating based on the size of the surface defects ,according to The size of the value is used to evaluate the size of surface defects and obtain surface defect size information.

[0034] The present invention is improved in that the method for evaluating the internal defect depth index is:

[0035] Based on the ultrasonic flaw detection data of castings, the depth data of internal defects of castings are extracted, and the thickness data of castings are obtained to obtain depth correlation data;

[0036] Based on the deep correlation data, by formula:

[0037] ,

[0038] Calculate the depth index of internal defects ,in, is the depth of the internal defect, is the thickness of the casting, is the attenuation coefficient, is the adjustment parameter, is the internal defect depth index.

[0039] The present invention is improved in that the method for evaluating the severity of internal defects is:

[0040] Based on the ultrasonic flaw detection data of castings, the distance between internal defects and key positions of castings is analyzed to obtain the location information of internal defects;

[0041] Depth index based on the internal defects And internal defect location information, through the formula:

[0042] ,

[0043] Calculate the severity index of internal defects ,in, is the position sensitivity adjustment factor, is the critical position threshold, is the internal defect depth index, is the location of the defect in the casting, is the position influence coefficient, is the severity index of internal defects, is the base of natural logarithms;

[0044] Based on the severity index of the internal defects ,according to The value is used to evaluate the severity of internal defects and obtain information on the severity of internal defects.

[0045] The present invention is improved in that the steps of obtaining the casting defect detection result are:

[0046] Based on the surface defect category information, surface defect size information and internal defect severity information, the formula is:

[0047] ,

[0048] Calculate the impact of defects ,in, is the defect type influencing factor, is the surface defect size rating, is the severity index of internal defects, and is the adjustment parameter, is the degree of defect impact, is the base of natural logarithms;

[0049] Based on the impact of the defect , according to the degree of defect impact The castings are divided into different quality categories based on the value and the casting defect detection results are obtained.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are:

[0051] In the present invention, the accuracy and efficiency of casting defect recognition are improved by integrating image recognition and ultrasonic flaw detection data processing. By analyzing the surface image of the casting, the defect characteristics are matched with the preset defect library to accurately classify the defect category. By analyzing the edge contour of the defect and calculating the defect area, the size of the defect is evaluated, making the defect detection more detailed and comprehensive. By using ultrasonic flaw detection data to analyze the internal structure of the casting, the location and severity of the internal defects can be accurately evaluated. By comprehensively evaluating the quality of the casting, the manufacturer is allowed to make repair or rejection decisions in a timely manner during the production process, thereby improving the efficiency of the production process and the final quality of the casting product. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a system flow chart of the present invention;

[0053] Figure 2 A flow chart for identifying surface defects of a part for the present invention;

[0054] Figure 3 A flow chart for obtaining surface defect category information of the present invention;

[0055] Figure 4 A flow chart for calculating the area of ​​each surface defect for the present invention;

[0056] Figure 5 A flow chart for obtaining surface defect size information of the present invention;

[0057] Figure 6 A flow chart for evaluating an internal defect depth index for the present invention;

[0058] Figure 7 A flow chart for obtaining internal defect severity information for the present invention;

[0059] Figure 8 The present invention is a flow chart for obtaining casting defect detection results. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0062] See also Figure 1 The present invention provides a technical solution: a casting defect detection system based on data analysis, the system comprising:

[0063] The surface defect analysis module identifies the surface defects of parts through image analysis based on the surface image of the casting. According to the surface defect characteristics of the parts, it evaluates the matching degree between the defect characteristics and the defects in the preset defect library, identifies the surface defect categories, including pores, slag inclusions and surface depressions, and obtains the surface defect category information;

[0064] The defect size assessment module identifies the edge contour of the surface defect based on the surface defect category information, calculates the area of ​​each surface defect, and assesses the size of the surface defect in combination with the shape of the surface defect to obtain the surface defect size information;

[0065] The internal defect analysis module identifies the internal defects of the casting based on the ultrasonic flaw detection data of the casting, evaluates the internal defect depth index according to the depth of the internal defect and the thickness data of the casting, and evaluates the severity of the internal defect in combination with the location of the internal defect on the casting to obtain the internal defect severity information;

[0066] The defect impact assessment module assesses the overall impact of defects on castings based on surface defect category information, surface defect size information and internal defect severity information, divides castings into quality categories, and obtains casting defect detection results.

[0067] Surface defect category information includes defect type and defect location, surface defect size information includes defect area, defect shape information and defect area ratio, internal defect severity information includes defect depth index, defect location coordinates and defect severity level, and casting defect detection results include casting quality grade and key defect influencing factors.

[0068] See also Figure 2 , the method of identifying the surface defects of parts is:

[0069] Based on the casting surface image, pixel-level features are extracted, including grayscale value, edge intensity and texture directionality, to obtain pixel-level feature information;

[0070] Based on the pixel-level feature information, the formula is:

[0071] ,

[0072] Calculate the feature difference between the pixel and the surrounding pixels ,in, is the normalization factor, It is The weight of the dimension feature, is the exponential coefficient, Indicates The feature difference of each pixel, It is The pixel dimensional eigenvalue, It is The pixel The range average of the dimension feature;

[0073] Based on the feature difference , compared with the preset feature difference threshold, the pixel points exceeding the preset feature difference threshold are marked as defect points to obtain defect marking information.

[0074] formula: ,

[0075] Parameter details and how to obtain them:

[0076] Indicates The pixel at The value of the dimensional feature. The value can be obtained by processing the grayscale, texture, color, etc. of the pixel points of the image. For example, for the grayscale feature, It can be extracted directly from the grayscale data of the image.

[0077] It is The area around the pixel is The average value of the dimensional feature. Usually calculated by All pixels in a small window around the pixel The average of the values ​​is obtained.

[0078] It is The weight coefficient of the dimension feature when calculating the difference. This weight can be set according to the importance of different features, usually assigned by experts based on experience or obtained through optimization algorithms.

[0079] is the exponential coefficient, which is a parameter that can be adjusted according to actual application requirements. For example, increasing This can make the model more sensitive to subtle differences.

[0080] is a normalization factor to ensure The value of is within a reasonable range. This factor can be obtained by The value is calculated by determining its maximum possible value after statistical analysis.

[0081] Calculation example:

[0082] Assume the following assumed parameter values ​​and data:

[0083] Pixels The values ​​in the three feature dimensions are Grayscale value, texture intensity, color intensity),

[0084] The corresponding feature mean is ,

[0085] The weight is set to ,

[0086] Adjustment Index =2

[0087] Normalization factor =10,

[0088] Calculate the variance contribution of each feature:

[0089] ,

[0090] ,

[0091] ;

[0092] Calculate feature difference :

[0093]

[0094] Calculation results =1.323 represents pixel The feature difference is used to determine whether there is a defect. The higher the value, the more obvious the difference between the pixel and the surrounding area, and the more likely it is a defect.

[0095] See also Figure 3 , the steps to obtain surface defect category information are:

[0096] Based on the defect marking information, the defect area is compared with the preset defect library, using the formula:

[0097] ,

[0098] Calculate the similarity between the features of the defect area and the features in the defect library to obtain a similarity score ,in, is the similarity score, For the The weight of the feature, The defect area eigenvalues, The first eigenvalues, is the sensitivity coefficient, is the total number of features, is the base of natural logarithms;

[0099] Based on the similarity score , and compared with the preset similarity threshold, the defect types exceeding the preset similarity threshold are marked as corresponding categories to obtain surface defect category information.

[0100] formula:

[0101] ,

[0102] Parameter details and how to obtain them:

[0103] : Weight coefficient, which indicates the importance of each feature in the defect matching process. The weight can be obtained through expert scoring or historical data analysis. Experts score the features based on their effectiveness in identifying defects, or the weight of each feature is calculated from the historical identification success rate.

[0104] : It is the characteristic value of the defect area identified in the image, such as color or texture characteristics. The value is usually extracted through image processing algorithms such as edge detection or texture analysis.

[0105] : is the value of the corresponding feature in the defect library, which is obtained by feature extraction through the image dataset of historical known defects.

[0106] : It is a coefficient that adjusts the sensitivity of the exponential function, obtained through preliminary experiments or tuning, and affects the calculation sensitivity of the similarity score.

[0107] : Indicates the total number of features involved in the calculation. The value is based on the image feature set being analyzed. The number of features that need to be considered is usually determined by the image analysis stage.

[0108] Calculation example:

[0109] Assume there are three features, namely =3, feature weight =0.5, =0.3, =0.2, adjustment factor =0.1, image region eigenvalue =100, =150, =120, defect library characteristic value =90, =140, =110.

[0110] The calculation is as follows:

[0111] ,

[0112] for =1:

[0113] ,

[0114] for =2:

[0115] ,

[0116] for =3:

[0117] ,

[0118] The total similarity score for

[0119] ,

[0120] In this case, the total score =0.731 represents the similarity score of defect recognition. This score reflects the degree of match between the image area and the defects recorded in the defect library. The higher the score, the higher the degree of match between the image area and a known defect, thus identifying the corresponding defect category. This score is used to ultimately determine the defect type in the image.

[0121] See also Figure 4 , the method for calculating the area of ​​each surface defect is:

[0122] Based on the surface defect category information, the formula is:

[0123] ,

[0124] Calculate the edge intensity response of a defect , the edge intensity response Compare with the preset edge threshold, mark the pixel points exceeding the preset edge threshold as defective edges, and obtain defective edge information, where: It is at coordinates The edge intensity response at , the edge response intensity at , is mainly used to identify obvious edges in the image. This response value is obtained by calculating the gradient size of the image at this point, reflecting the local change intensity of the image brightness. Larger response values ​​usually correspond to the boundaries of objects or other significant graphic features. and The images are and Intensity gradient in direction;

[0125] Based on the defect edge information, the formula is:

[0126] ,

[0127] Calculate defect area , get the defect area information, where represents the defect area, For the The number of pixels in the defect area, is the total number of defect areas of the same type, is the image scaling factor, It is The shape complexity parameter of the defect area.

[0128] formula:

[0129] and ,

[0130] Parameter details and how to obtain them:

[0131] : is the coordinate in the image The edge response strength at is mainly used to identify obvious edges in the image. This response value is obtained by calculating the gradient of the image at this point, reflecting the local change intensity of the image brightness. A larger response value usually corresponds to the boundary of an object or other significant graphic features.

[0132] and :The parameters represent the image in and The intensity gradient in the direction. Obtained by applying a gradient filter (such as the Sobel filter) to the image, the filter measures the speed of change of brightness by differentiating the brightness values ​​of adjacent pixels, which can highlight the location of edges in the image.

[0133] :Indicates the The parameter is usually obtained by counting the number of pixels in a given area through image processing software or algorithms, and is often used to quantify the size of the defect area.

[0134] : The scaling factor of the image, used to convert pixel area to actual area units (such as square millimeters). The factor is set according to the image resolution and the known scale. For example, if one pixel represents an actual area of ​​0.1 mm × 0.1 mm, then the scaling factor will be 0.1.

[0135] : Represents the shape complexity parameter of each defect area. The parameter is calculated by analyzing the irregularity of the defect shape, for example, by calculating the curvature or rate of change of the defect boundary. Complex shapes may require more detailed processing to accurately calculate the area.

[0136] : The total number of areas with the same type of defects, indicating the total number of areas with the same type of defects.

[0137] Calculation example:

[0138] Assume that at a certain point in the image , use the Sobel operator to calculate the gradient. The following is a specific numerical setting:

[0139] exist Directional gradient =120 Directional gradient =80,

[0140] Use the edge detection formula:

[0141]

[0142] The edge response value obtained This value is high, indicating that the coordinate The image has obvious edges.

[0143] Assume that a specific defect is being analyzed with the following parameters:

[0144] =3: Three defect areas.

[0145] : The number of pixels in each area.

[0146] =0.05: The actual area represented by each pixel is 0.05×0.05 square millimeters.

[0147] : Shape complexity parameter of each defect.

[0148] Calculate the total area of ​​these defects :

[0149] ,

[0150] Calculate item by item:

[0151] for =1:

[0152] ,

[0153] for =2:

[0154] ,

[0155] for =3

[0156] ,

[0157] The total area obtained for:

[0158] ,

[0159] Calculated total area =0.8642 square millimeters, which represents the total area of ​​these three defect areas.

[0160] See also Figure 5 , the steps to obtain the surface defect size information are:

[0161] Based on the defect area , combined with the shape complexity of surface defects, through the formula:

[0162] ,

[0163] Calculate the rating of surface defect size ,in, is the defect area, is the average shape complexity parameter of surface defects, and is the adjustment parameter, is a rating of the size of surface defects;

[0164] Rating based on the size of the surface defects ,according to The size of the value is used to evaluate the size of surface defects and obtain surface defect size information.

[0165] formula:

[0166] ,

[0167] Parameter details and how to obtain them:

[0168] : The area of ​​the surface defect, calculated in the previous step.

[0169] : The average shape complexity parameter of the surface defect, which is quantified by analyzing the geometric characteristics of the defect contour, for example, by measuring the curvature of the contour or the irregularity of the defect boundary. The shape complexity parameter of each defect area) is averaged and calculated.

[0170] and : Parameters are used to adjust the sensitivity and threshold of defect size rating. In practical applications, Usually set as a factor that affects the slope of the curve, As thresholds for distinguishing defect size classification, the specific values ​​of the two can be obtained through historical data statistics and experimental optimization.

[0171] : The final defect size rating, which takes into account the area and shape complexity of the defect and is used to classify the defect more accurately.

[0172] Calculation example:

[0173] Assuming the defect area , average shape complexity , shape adjustment coefficient = 0.5, area threshold .

[0174] Calculate the size rating of defects :

[0175] ,

[0176] The defect size rating is calculated as The value reflects the size of the defect, which is evaluated in combination with the area and shape complexity. A higher rating indicates that the defect is larger in overall size.

[0177] See also Figure 6 , the method for evaluating the internal defect depth index is:

[0178] Based on the ultrasonic flaw detection data of castings, the depth data of internal defects of castings are extracted, and the thickness data of castings are obtained to obtain depth correlation data;

[0179] Based on the deep correlation data, by formula:

[0180] ,

[0181] Calculate the depth index of internal defects ,in, is the depth of the internal defect, is the thickness of the casting, is the attenuation coefficient, is the adjustment parameter, is the internal defect depth index.

[0182] formula:

[0183] ,

[0184] Parameter details and how to obtain them:

[0185] : Indicates the vertical depth from the surface of the casting to the internal defect. The value is usually measured by an ultrasonic flaw detector and calculated based on the propagation time of the ultrasonic wave and the known sound velocity of the material. The propagation speed of the sound wave in the material is known, and the distance can be calculated by measuring the time it takes for the sound wave to travel from the emission point to the reception point.

[0186] : is the total thickness of the casting, this value is known by measurement or manufacturing specifications before flaw detection. It is one of the physical properties of the casting and is usually directly given by the design parameters in the production process.

[0187] : Attenuation coefficient, which is an experimentally determined constant used to adjust the effect of depth and thickness differences on the depth index. The setting of this coefficient depends on the characteristics of the material and the impact of the defect type being evaluated.

[0188] : are adjustment parameters used to introduce flexibility into the calculations, allowing the model to better adapt to different measurement conditions and material properties. The specific values ​​of the parameters can be obtained by adjusting the optimization algorithm or empirical data.

[0189] Calculation example:

[0190] Assumed parameters:

[0191] Defect depth =20mm, casting thickness =50mm, attenuation coefficient =0.1, adjust the parameters =0.5.

[0192] Calculate the depth index of internal defects :

[0193] ,

[0194] The calculated internal defect depth index is The value provides an adjusted measure of the defect relative to the casting thickness, taking into account both the absolute value of the depth and its relative importance to the structure. Values ​​may indicate that the defect location has a greater impact on the overall integrity of the casting and requires further evaluation and treatment. This evaluation helps prioritize those defects that may have the greatest impact on the safety of the structure.

[0195] See also Figure 7 , the method for evaluating the severity of internal defects is:

[0196] Based on the ultrasonic flaw detection data of castings, the distance between internal defects and key positions of castings is analyzed to obtain the location information of internal defects;

[0197] Depth index based on internal defects And internal defect location information, through the formula:

[0198] ,

[0199] Calculate the severity index of internal defects ,in, is the position sensitivity adjustment factor, is the critical position threshold, is the internal defect depth index, is the location of the defect in the casting, is the position influence coefficient, is the severity index of internal defects, is the base of natural logarithms;

[0200] Internal defect severity index ,according to The value is used to evaluate the severity of internal defects and obtain information on the severity of internal defects.

[0201] formula:

[0202] ,

[0203] Parameter details and how to obtain them:

[0204] : Internal defect depth index, calculated by the previous step.

[0205] : Location influence factor, evaluated based on the location of the defect relative to critical parts of the casting structure. The factor reflects the potential impact of the defect location on the overall function of the casting and is usually determined by the structural engineer based on experience or based on previous case studies.

[0206] : Position sensitivity adjustment factor, which adjusts the impact of defect position information on the final evaluation result. This is an empirical value, usually obtained through simulation or historical data analysis, to adjust the sensitivity of the formula to position changes.

[0207] : Critical location threshold sets a limit on the defect location to determine when the defect location begins to significantly affect its severity assessment. The value is usually set based on engineering judgment or safety standards.

[0208] :The specific location of the defect in the casting is determined based on the relative distance between the defect and the key position of the casting.

[0209] Calculation example:

[0210] Assumptions:

[0211] Defect Depth Index =0.4,

[0212] Position influence coefficient =0.7,

[0213] Position sensitivity adjustment factor =0.3,

[0214] Key location threshold =10mm,

[0215] Defect location =15mm,

[0216] Calculate the severity of internal defects :

[0217] ,

[0218] The obtained internal defect severity index This value represents the combined effect of internal defects, combining the two key factors of depth and location. The values ​​indicate that despite the depth of the defect, its location has relatively little impact on the overall structural safety of the casting.

[0219] See also Figure 8 , the steps to obtain the casting defect detection results are:

[0220] Based on the surface defect category information, surface defect size information and internal defect severity information, the formula is:

[0221] ,

[0222] Calculate the impact of defects ,in, is the defect type influencing factor, is the surface defect size rating, is the severity index of internal defects, and is the adjustment parameter, is the degree of defect impact, is the base of natural logarithms;

[0223] Based on the impact of defects , according to the degree of defect impact The castings are divided into different quality categories based on the value and the casting defect detection results are obtained.

[0224] formula:

[0225] ,

[0226] Parameter details and how to obtain them:

[0227] : Surface defect type factor, which scores the severity of the defect type based on its potential impact on casting performance. The coefficients are obtained through historical data analysis and materials science research, and are intended to quantify the severity of different types of defects.

[0228] : Surface defect size rating, calculated in the previous step.

[0229] : The severity index of internal defects, calculated by the previous step.

[0230] and : Adjustment factor, used to adjust the sensitivity and reaction threshold of the defect assessment model. The parameters are set empirically through optimization algorithms or according to the standards of casting quality control to ensure the accuracy and applicability of the model.

[0231] Calculation example:

[0232] Assume that the specific parameter values ​​are as follows:

[0233] Defect type factor =1.5, reflecting a surface defect of moderate severity.

[0234] Surface defect size rating =0.8, indicating larger surface defects.

[0235] Internal Defect Severity Index =0.5, indicating that the defect is located in an important structural position of the casting.

[0236] Modulating Factor = 0.2 and threshold =2, used to adjust the calculation of the comprehensive defect score.

[0237] Use the formula to calculate the comprehensive impact of defects in castings :

[0238] ,

[0239] Comprehensive influence of casting defects The value indicates that the comprehensive impact of defects on the overall quality of castings is medium to high. Values ​​below 0.05 indicate that the overall properties of the casting may be significantly affected and require further repair or rejection.

[0240] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A casting defect detection system based on data analysis, characterized in that: The system comprises: The surface defect analysis module is based on the casting surface image. Through image analysis, it identifies the surface defects of the parts. According to the surface defect characteristics of the parts, it evaluates the matching degree between the defect characteristics and the defects in the preset defect library, identifies the surface defect category, and obtains the surface defect category information. The defect size assessment module identifies the edge contour of the surface defect based on the surface defect category information, calculates the area of ​​each surface defect, and assesses the size of the surface defect in combination with the shape of the surface defect to obtain surface defect size information; The internal defect analysis module identifies the internal defects of the casting based on the ultrasonic flaw detection data of the casting, evaluates the internal defect depth index according to the depth of the internal defect and the thickness data of the casting, and evaluates the severity of the internal defect in combination with the location of the internal defect on the casting to obtain the internal defect severity information; The method for evaluating the internal defect depth index is: Based on the ultrasonic flaw detection data of castings, the depth data of internal defects of castings are extracted, and the thickness data of castings are obtained to obtain depth correlation data; Based on the deep correlation data, by formula: ， Calculate the depth index of internal defects ,in, is the depth of the internal defect, is the thickness of the casting, is the attenuation coefficient, is the adjustment parameter, is the internal defect depth index; The method for evaluating the severity of internal defects is: Based on the ultrasonic flaw detection data of castings, the distance between internal defects and key positions of castings is analyzed to obtain the location information of internal defects; Depth index based on the internal defects And internal defect location information, through the formula: ， Calculate the severity index of internal defects ,in, is the position sensitivity adjustment factor, which adjusts the influence of defect position information on the final evaluation result and is used to adjust the sensitivity of the formula to position changes. is the critical position threshold, the limit value of the defect location, which is used to determine when the location of the defect begins to significantly affect its severity assessment. is the internal defect depth index, is the location of the defect in the casting, is the position influence coefficient, which is evaluated based on the location of the defect relative to the critical part of the casting structure. The coefficient reflects the potential impact of the defect location on the overall function of the casting. is the severity index of internal defects, is the base of natural logarithms; Based on the severity index of the internal defects ,according to The value is used to evaluate the severity of internal defects and obtain information on the severity of internal defects; The defect impact assessment module assesses the overall impact of defects on castings based on the surface defect category information, surface defect size information and internal defect severity information, classifies castings by quality, and obtains casting defect detection results; The steps for obtaining the casting defect detection result are: Based on the surface defect category information, surface defect size information and internal defect severity information, the formula is: ， Calculate the impact of defects ,in, is the defect type influencing factor, is the surface defect size rating, is the severity index of internal defects, and is the adjustment parameter, is the degree of defect impact, is the base of natural logarithms; Based on the impact of the defect , according to the degree of defect impact The castings are divided into different quality categories based on the value and the casting defect detection results are obtained.

2. The casting defect detection system based on data analysis according to claim 1, characterized in that: The method for identifying surface defects of parts is: Based on the casting surface image, pixel-level features are extracted, including grayscale value, edge intensity and texture directionality, to obtain pixel-level feature information; Based on the pixel-level feature information, the formula is: ， Calculate the feature difference between the pixel and the surrounding pixels ,in, is the normalization factor, It is The weight of the dimension feature, is the exponential coefficient, Indicates The feature difference of each pixel, It is The pixel dimensional eigenvalue, It is The pixel The range average of the dimension feature; Based on the feature difference , compared with the preset feature difference threshold, the pixel points exceeding the preset feature difference threshold are marked as defect points to obtain defect marking information.

3. The casting defect detection system based on data analysis according to claim 2, characterized in that: The steps for obtaining the surface defect category information are as follows: Based on the defect marking information, the defect area is compared with the preset defect library, and the formula is: ， Calculate the similarity between the features of the defect area and the features in the defect library to obtain a similarity score ,in, is the similarity score, For the The weight of the feature, The defect area eigenvalues, The first eigenvalues, is the sensitivity coefficient, is the total number of features, is the base of natural logarithms; Based on the similarity score , and compared with the preset similarity threshold, the defect types exceeding the preset similarity threshold are marked as corresponding categories to obtain surface defect category information.

4. The casting defect detection system based on data analysis according to claim 1, characterized in that: The method for calculating the area of ​​each surface defect is: Based on the surface defect category information, the formula: ， Calculate the edge intensity response of a defect , the edge intensity response Compare with the preset edge threshold, mark the pixel points exceeding the preset edge threshold as defective edges, and obtain defective edge information, where: is at coordinates The edge intensity response at and The images are and Intensity gradient in direction; Based on the defect edge information, the formula: ， Calculate defect area , get the defect area information, where represents the defect area, For the The number of pixels in the defect area, is the total number of defect areas of the same type, is the image scaling factor, It is The shape complexity parameter of the defect area.

5. The casting defect detection system based on data analysis according to claim 4 is characterized in that: The steps for obtaining the surface defect size information are as follows: Based on the defect area , combined with the shape complexity of surface defects, through the formula: ， Calculate the rating of surface defect size ,in, is the defect area, is the average shape complexity parameter of surface defects, and is the adjustment parameter, is a rating of the size of surface defects; Rating based on the size of the surface defects ,according to The size of the value is used to evaluate the size of surface defects and obtain surface defect size information.

Images