A steel surface defect analysis method, device, terminal and storage medium

Abstract

The present application provides a steel surface defect analysis method, device, terminal and storage medium, and relates to the technical field of steel analysis. The method includes: obtaining a first image and a second image of the surface of each layer of a target steel sample, the first image being the initial image of the surface of each layer of the target steel sample, and the second image being the image of each layer of the surface of the target steel sample after corrosion; inputting the first image and the second image of each layer into a defect boundary point generation model respectively, and outputting the defect boundary points of the corresponding layer; based on all defect boundary points, determining the defect three-dimensional boundary point data image of the target steel sample, and inputting the defect three-dimensional boundary point data image of the target steel sample into a defect type discrimination model to determine the defect type of the target steel sample, and the defect type discrimination model is constructed based on a machine learning algorithm. The present application can more accurately capture and identify defects on the steel surface and improve detection efficiency.

Description

Technical Field

[0001] The present application relates to the technical field of steel analysis, and in particular to a steel surface defect analysis method, device, terminal and storage medium. Background Art

[0002] Due to the addition of a certain Mn content (4%~12%), medium manganese steel will cause a multi-phase composite state of martensite, austenite and ferrite to appear in the room temperature organization. Due to the uneven mechanical properties of each phase, defects such as grooves and shrinkage holes will appear, increasing the degree of confusion of surface defect types, which is not conducive to the feature extraction and analysis of surface defects. And with the in-depth study of medium manganese steel, elements such as Ni, Al, Ti, and V will be added to obtain more excellent performance. However, due to the effect of Mn, it will promote the precipitation of NiAl phase in the ferrite matrix, and non-metallic inclusions such as MnS will be generated. It will also promote the formation of oxides at higher temperatures, resulting in defects such as inclusions and scales on the surface. Summary of the invention

[0003] The present application provides a steel surface defect analysis method, device, terminal and storage medium to solve the problem of inaccurate steel surface defect identification in the prior art.

[0004] In a first aspect, the present application provides a method for analyzing steel surface defects, comprising:

[0005] Acquire a first image and a second image of each surface layer of a target steel sample, wherein the first image is an initial image of each surface layer of the target steel sample, and the second image is an image of each surface layer of the target steel sample after corrosion;

[0006] Inputting the first image and the second image of each layer into the defect boundary point generation model respectively, and outputting the defect boundary points of the corresponding layer;

[0007] Based on all defect boundary points, the defect three-dimensional boundary point data image of the target steel sample is determined, and the defect three-dimensional boundary point data image of the target steel sample is input into a defect type discrimination model to determine the defect type of the target steel sample. The defect type discrimination model is constructed based on a machine learning algorithm.

[0008] In a second aspect, the present application provides a steel surface defect analysis device, comprising:

[0009] An image acquisition module, used to acquire a first image and a second image of each surface layer of a target steel sample, wherein the first image is an initial image of each surface layer of the target steel sample, and the second image is an image of each surface layer of the target steel sample after corrosion;

[0010] A boundary point determination module, used to input the first image and the second image of each layer into the defect boundary point generation model respectively, and output the defect boundary points of the corresponding layer;

[0011] A defect type determination module is used to determine the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points, and input the defect three-dimensional boundary point data image of the target steel sample into a defect type discrimination model to obtain the inclusion type in the target steel sample. The defect type discrimination model is constructed based on a machine learning algorithm.

[0012] In a third aspect, the present application provides a terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the method described in the first aspect or any possible implementation of the first aspect are implemented.

[0013] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the method described in the first aspect or any possible implementation of the first aspect.

[0014] The present application provides a steel surface defect analysis method, device, terminal and storage medium, which obtains a first image and a second image of the surface of each layer of a target steel sample, wherein the first image is an initial image of the surface of each layer of the target steel sample, and the second image is an image of the surface of each layer of the target steel sample after corrosion; the first image and the second image of each layer are respectively input into a defect boundary point generation model, and the defect boundary points of the corresponding layer are output; based on all defect boundary points, a defect three-dimensional boundary point data image of the target steel sample is determined, and the defect three-dimensional boundary point data image of the target steel sample is input into a defect type discrimination model to determine the defect type of the target steel sample, wherein the defect type discrimination model is constructed based on a machine learning algorithm. The present application can more accurately capture and identify defects on the steel surface by acquiring the initial image and the image after corrosion of each layer of the target steel sample. The corrosion treatment helps to highlight the defect characteristics, making the defects more obvious in the image, thereby improving the accuracy of defect identification. The defect boundary point generation model is used to output the defect boundary points of the corresponding layer, thereby achieving precise positioning of the defects. At the same time, by introducing a defect type discrimination model constructed based on a machine learning algorithm, automatic identification and classification of defect types are achieved, which not only greatly improves the detection efficiency, but also reduces the subjectivity and errors of human judgment, making the detection results more objective and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] Figure 1 It is a flow chart of the implementation of the steel surface defect analysis method provided in the embodiment of the present application;

[0017] Figure 2 It is a structural schematic diagram of a steel surface defect analysis device provided in an embodiment of the present application;

[0018] Figure 3 It is a schematic diagram of a terminal provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0020] In order to make the purpose, technical solutions and advantages of the present application clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0021] Due to the uneven mechanical properties of each phase, defects such as grooves and shrinkage holes will appear in steel during production, which increases the complexity of the types of surface defects and is not conducive to the feature extraction and analysis of surface defects. Based on this, the design concept of the embodiment of the present application is: different from the two-dimensional machine vision detection method of the traditional architecture, the surface of the steel sample is processed for data, then polished, and the surface data after polishing is sorted, and the process of polishing and data sorting is repeated until the surface is flat and has no defects, and then the surface data obtained from each layer is fitted to obtain three-dimensional data, and the three-dimensional data missing from the plane vision is obtained through the multi-layer corrosion method. At the same time, it can also improve the detection accuracy, reduce the detection difficulty, and ensure the flexibility and scalability of the detection process.

[0022] Based on the above design concept, the embodiment of the present application proposes a steel surface defect analysis method for concave-convex surface defects such as shrinkage cavities and inclusions that often appear in steel (these defects are not easy to directly construct three-dimensional data), including multi-layer corrosion data collection, preprocessing, data reconstruction and other designs. Among them, multi-layer corrosion data collection is introduced to obtain multi-layer data on the surface of the target object, and data preprocessing is performed at the same time to remove noise points for subsequent processing; the plane data collected from multi-layer corrosion is reconstructed into three-dimensional data of defects for easy observation and processing.

[0023] Among them, the steel in the embodiment of the present application is medium manganese steel.

[0024] Figure 1 The implementation flow chart of the steel surface defect analysis method provided in the embodiment of the present application is detailed as follows:

[0025] In step 101, a first image and a second image of each surface layer of a target steel sample are obtained, wherein the first image is an initial image of each surface layer of the target steel sample, and the second image is an image of each surface layer of the target steel sample after corrosion.

[0026] In the embodiment of the present application, the first image and the second image of the surface of the target steel sample are obtained layer by layer using an imaging device, wherein the first image is the initial image of each layer surface, that is, the initial image of the surface that has not been corroded, and the second image is the surface image of each layer surface after corrosion.

[0027] In a possible implementation, obtaining the first image and the second image of each layer surface of the target steel sample may include:

[0028] For each surface layer of the target steel sample, perform the following steps:

[0029] Acquire an initial image of the target steel sample on the surface of the layer, and use the initial image as the first image;

[0030] After the surface layer of the target steel sample is continuously corroded by the preset corrosive liquid for a preset time, a surface image of the target steel sample after the surface layer is corroded is obtained, and the surface image of the surface layer after the corrosion is used as the second image.

[0031] Optionally, since the processing of each surface layer is the same, the following operations are performed on each surface layer:

[0032] An initial image of the layer surface before being corroded is obtained, and the initial image is used as the first image. Then, the layer surface is corroded for a preset time using a preset corrosive liquid, and after the corrosion, a surface image of the layer surface after corrosion is obtained, and the surface image is used as the second image.

[0033] It should be noted that, due to the addition of 4%-12% Mn in medium manganese steel, MnS inclusions are easily precipitated in the matrix. MnS inclusions induce corrosion at the junction with the matrix. Once inclusions appear on the surface of the matrix, they will cause serious damage to the steel matrix. Especially pitting, because the corrosion products often cover the pitting pits and are not easy to be found, it is very easy to cause sudden accidents, which is one of the most dangerous and destructive corrosion types. The pitting resistance of steel decreases with the increase of chloride ion concentration, especially the strong corrosion of sodium chloride solution to MnS, so the process of using the preset corrosive liquid to corrode the surface for a preset time in the embodiment of the present application is divided into two steps, that is, the preset corrosive liquid includes a first preset corrosive liquid and a second preset corrosive liquid, the first preset corrosive liquid is a mixed solution of 4% nitric acid alcohol and ferric chloride solution, the second preset corrosive liquid is a solution of 100ml pure water and 3.5g sodium chloride, the preset time includes a first preset time and a second preset time, the time range of the first preset time is 10-30s, and the time range of the second preset time is 10-20s. The specific corrosion process is as follows:

[0034] For each surface layer of the target steel sample, a first preset corrosion liquid, a mixed solution of 4% nitric acid alcohol and ferric chloride solution, is used to corrode the surface layer for a first preset time, and then a second preset corrosion liquid, a solution of 100 ml pure water and 3.5 g sodium chloride is used to corrode the surface layer after the first corrosion for a second preset time.

[0035] For example, when the target steel sample is medium manganese steel, for the initial surface, first obtain the first image of the initial surface, then use a mixed solution of 4% nitric acid alcohol and ferric chloride solution as the preset corrosion solution to corrode the initial surface, and the corrosion time is between 10-30s, and then use 100ml pure water and 3.5g sodium chloride solution to corrode the corroded surface again, and the corrosion time is between 10-20s. The defective part becomes a weak link of stress concentration and the root cause of corrosion, which will cause the defective part to have more obvious changes after corrosion, and obtain the second image of the initial surface. According to the actual specifications of the target steel sample, determine the number of times corrosion is required.

[0036] It should be noted that after each etching to obtain the second image of the corresponding layer, the etching surface needs to be polished to ensure that the surface of the first image obtained at each layer is flat. The polishing thickness ranges from 2 to 3 μm, and the specific polishing thickness is determined according to the actual situation.

[0037] The embodiment of the present application can more accurately capture and identify defects on the steel surface by acquiring the initial image and the image after corrosion of each layer of the target steel sample. Corrosion treatment helps to highlight the defect features, making the defects more obvious in the image, thereby improving the accuracy of defect identification.

[0038] In step 102, the first image and the second image of each layer are respectively input into a defect boundary point generation model, and the defect boundary points of the corresponding layer are output.

[0039] In the embodiment of the present application, for the first image and the second image of each layer, the first image and the second image of the layer are input into the constructed defect boundary point generation model to obtain the defect boundary points of the corresponding layer.

[0040] The defect boundary point generation model compares the first image with the second image, identifies the boundary of the part with a large difference between the before and after changes, and uses it as the defect boundary point.

[0041] The construction process of the defect boundary point generation model is as follows:

[0042] A first image and a second image of a historical steel sample are obtained, wherein the first image is an initial image of a surface layer of the historical steel sample, and the second image is a surface image of the historical steel sample after corrosion of a surface layer; and a historical defect boundary point between the first image and the second image is obtained.

[0043] Construct defect boundary point generation model.

[0044] The first image and the second image of the historical steel sample are taken as input, and the corresponding historical defect boundary points are taken as output to train the defect boundary point generation model.

[0045] Wherein, the first image and the second image are both grayscale images.

[0046] In an embodiment of the present application, the defect boundary point generation model can be constructed based on the YOLOv5 model or based on a deep learning model, which is not limited in the embodiment of the present application.

[0047] The embodiment of the present application utilizes a defect boundary point generation model to output defect boundary points of corresponding layers, thereby achieving accurate positioning of defects.

[0048] In step 103, based on all defect boundary points, the defect three-dimensional boundary point data image of the target steel sample is determined, and the defect three-dimensional boundary point data image of the target steel sample is input into a defect type discrimination model to determine the defect type of the target steel sample. The defect type discrimination model is constructed based on a machine learning algorithm.

[0049] In the embodiment of the present application, the defect three-dimensional boundary point data image of the target steel sample is reconstructed using all defect boundary points obtained in step 102. Then, the defect three-dimensional boundary point data image is input into the constructed defect type discrimination model to obtain the defect type of the target steel sample.

[0050] In the embodiment of the present application, based on all defect boundary points, a three-dimensional boundary point data image of the defect of the target steel sample can be constructed. This three-dimensional visualization technology not only makes the morphology and distribution of the defects more intuitive, but also helps engineers to better understand the structure and characteristics of the defects, and provides strong support for subsequent analysis and processing. In addition, by introducing a defect type discrimination model built based on a machine learning algorithm, the method realizes automatic identification and classification of defect types. This not only greatly improves the detection efficiency, but also reduces the subjectivity and errors of human judgment, making the detection results more objective and reliable. At the same time, with the continuous optimization and training of the machine learning algorithm, the accuracy and intelligence level of the method will be further improved.

[0051] In a possible manner, before determining the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points, the method may further include:

[0052] Use filtering algorithm to remove noise points from defect boundary points of each layer;

[0053] Accordingly, based on the defect boundary points of each layer, determining the defect three-dimensional boundary point data image of the target steel sample may include:

[0054] Based on all defect boundary points after noise point removal, the defect three-dimensional boundary point data image of the target steel sample is determined.

[0055] Optionally, for the defect boundary points of each layer, a filtering algorithm is used to perform noise point removal processing on all defect boundary points to ensure that the obtained defect boundary points are accurate.

[0056] In addition, if the target steel sample is large, it needs to be scanned multiple times, and then the amount of data can be reduced through downsampling or data compression technology to improve the subsequent processing speed.

[0057] Then, the preprocessed defect boundary points are used to determine the defect three-dimensional boundary point data image of the target steel sample.

[0058] In a possible implementation, determining a defect three-dimensional boundary point data image of a target steel sample based on all defect boundary points may include:

[0059] Using the Poisson surface reconstruction algorithm, all defect boundary points are reconstructed into a three-dimensional defect boundary point data image of the target steel sample.

[0060] Optionally, this embodiment can use Poisson reconstruction algorithm, Delaunay triangulation algorithm, etc. to reconstruct all defect boundary points into a defect three-dimensional boundary point data image of the target steel sample. The embodiment of the present application uses algorithms such as Poisson reconstruction algorithm and Delaunay triangulation algorithm to construct a three-dimensional data model of defects more clearly and conveniently.

[0061] The embodiment of the present application can intuitively detect and identify the three-dimensional structural part of the defect by reconstructing the three-dimensional boundary point data image of the target steel sample, which is more intuitive and richer than the data obtained by traditional two-dimensional machine vision detection. Among them, the embodiment of the present application can obtain the volume, cross-section, surface area, geometric shape and other characteristics of the defect by using the three-dimensional boundary point data image of the defect.

[0062] In a possible implementation, the process of constructing the defect type discrimination model may be:

[0063] Obtaining defect three-dimensional boundary point data images of historical steel samples, and obtaining defect types of historical steel samples;

[0064] Use machine learning algorithms to build a defect type identification model;

[0065] The defect three-dimensional boundary point data image of the historical steel sample is taken as input, and the defect type of the historical steel sample is taken as output to train the defect type discrimination model.

[0066] Optionally, the embodiment of the present application can not only compare the reconstructed defect three-dimensional boundary point data image with the characteristic shape or pattern of the known defect, but also use template matching technology to identify the defect, such as the defect type discrimination model based on the machine learning algorithm provided in the embodiment of the present application. The machine learning algorithm may include but is not limited to support vector machine (SVM), random forest, convolutional neural network. Alternatively, it can also be compared with a defect-free model, analyze the difference, and identify the defect by calculating the deviation of the data point.

[0067] The construction process of the defect type discrimination model is as follows:

[0068] The defect three-dimensional boundary point data image of the historical steel sample and the defect type of the historical steel sample are obtained, and then a defect type discrimination model is constructed using a machine learning algorithm. The defect three-dimensional boundary point data image is used as input and the defect type is used as output to train the defect type discrimination model.

[0069] Compared with traditional detection methods, the embodiments of the present application can complete defect detection and classification tasks more quickly, thereby improving detection efficiency. In addition, since the embodiments of the present application realize automated and intelligent detection, the investment in manual intervention and detection equipment is reduced, which helps to reduce detection costs. In addition, the embodiments of the present application also provide strong technical support for steel production and quality control. By timely discovering and handling defects, it can ensure that the quality and performance of steel meet the requirements, and it also helps enterprises improve production processes and procedures, and improve product quality and market competitiveness.

[0070] In a possible implementation, after inputting the defect three-dimensional boundary point data image of the target steel sample into the defect type discrimination model to determine the defect type of the target steel sample, the method may further include:

[0071] Acquire the size parameters of the defect three-dimensional boundary point data image of the target steel sample, the size parameters including volume parameters and shape parameters;

[0072] The size parameters and defect types are input into the defect degree determination model, and the defect grade of the target steel sample is output to quantify the defect degree of the target steel sample.

[0073] Optionally, after obtaining the defect type of the target steel sample, the size parameters of the defect three-dimensional boundary point data image of the target steel sample are obtained, including volume parameters and shape parameters. Then, the size parameters and defect type are input into the constructed defect degree determination model to obtain the defect grade of the target steel sample, and the defect grade is used to quantify the defect degree of the target steel sample. Among them, the defect degree determination model can be constructed based on a neural network.

[0074] The embodiment of the present application can also quantify the severity of defects in the target steel sample by measuring the size of defects such as volume and shape parameters. According to the characteristics of the defects and the quantification results, the defects are classified into different categories to distinguish different types of defects. A detailed report containing the location, type, size and severity of the defects is generated for subsequent processing or quality control.

[0075] The present application provides a steel surface defect analysis method, which obtains a first image and a second image of the surface of each layer of a target steel sample, wherein the first image is an initial image of the surface of each layer of the target steel sample, and the second image is an image of the surface of each layer of the target steel sample after corrosion; the first image and the second image of each layer are respectively input into a defect boundary point generation model, and the defect boundary points of the corresponding layer are output; based on all defect boundary points, a defect three-dimensional boundary point data image of the target steel sample is determined, and the defect three-dimensional boundary point data image of the target steel sample is input into a defect type discrimination model to determine the defect type of the target steel sample, wherein the defect type discrimination model is constructed based on a machine learning algorithm. The present application can more accurately capture and identify defects on the steel surface by acquiring the initial image and the image after corrosion of each layer of the target steel sample. The corrosion treatment helps to highlight the defect characteristics, making the defects more obvious in the image, thereby improving the accuracy of defect identification. The defect boundary point generation model is used to output the defect boundary points of the corresponding layer, thereby achieving precise positioning of the defects. At the same time, by introducing a defect type discrimination model constructed based on a machine learning algorithm, automatic identification and classification of defect types are achieved, which not only greatly improves the detection efficiency, but also reduces the subjectivity and errors of human judgment, making the detection results more objective and reliable.

[0076] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0077] The following is an embodiment of the device of the present application. For details not described in detail, please refer to the corresponding method embodiment described above.

[0078] Figure 2 The schematic diagram of the structure of the surface defect analysis device provided in the embodiment of the present application is shown. For the convenience of explanation, only the part related to the embodiment of the present application is shown, which is described in detail as follows:

[0079] like Figure 2 As shown, the surface defect analysis device 2 includes:

[0080] An image acquisition module 21 is used to acquire a first image and a second image of each surface layer of the target steel sample, wherein the first image is an initial image of each surface layer of the target steel sample, and the second image is an image of each surface layer of the target steel sample after corrosion;

[0081] A boundary point determination module 22, used to input the first image and the second image of each layer into the defect boundary point generation model, and output the defect boundary points of the corresponding layer;

[0082] The defect type determination module 23 is used to determine the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points, and input the defect three-dimensional boundary point data image of the target steel sample into the defect type discrimination model to obtain the inclusion type in the target steel sample. The defect type discrimination model is constructed based on a machine learning algorithm.

[0083] The present application provides a surface defect analysis device, which obtains the first image and the second image of each layer of the target steel sample surface, the first image is the initial image of each layer of the target steel sample surface, and the second image is the image of each layer of the target steel sample surface after corrosion; the first image and the second image of each layer are respectively input into the defect boundary point generation model, and the defect boundary point of the corresponding layer is output; based on all defect boundary points, the defect three-dimensional boundary point data image of the target steel sample is determined, and the defect three-dimensional boundary point data image of the target steel sample is input into the defect type discrimination model to determine the defect type of the target steel sample, and the defect type discrimination model is constructed based on the machine learning algorithm. The present application can more accurately capture and identify the defects of the steel surface by obtaining the initial image and the image after corrosion of each layer of the target steel sample surface, and the corrosion treatment helps to highlight the defect characteristics, making the defects more obvious in the image, thereby improving the accuracy of defect recognition, and the defect boundary point generation model is used to output the defect boundary point of the corresponding layer, so as to realize the accurate positioning of the defects; at the same time, by introducing the defect type discrimination model constructed based on the machine learning algorithm, the automatic recognition and classification of the defect type is realized, which not only greatly improves the detection efficiency, but also reduces the subjectivity and error of human judgment, making the detection result more objective and reliable.

[0084] In a possible implementation, the image acquisition module may be used to:

[0085] For each surface layer of the target steel sample, perform the following steps:

[0086] Acquire an initial image of the target steel sample on the surface of the layer, and use the initial image as the first image;

[0087] After the surface layer of the target steel sample is continuously corroded by the preset corrosive liquid for a preset time, a surface image of the target steel sample after the surface layer is corroded is obtained, and the surface image of the surface layer after the corrosion is used as the second image.

[0088] In a possible implementation, the preset corrosive liquid includes a first preset corrosive liquid and a second preset corrosive liquid, the first preset corrosive liquid is a mixed solution of 4% nitric acid alcohol and ferric chloride solution, and the second preset corrosive liquid is a solution of 100 ml of pure water and 3.5 g of sodium chloride;

[0089] The preset time includes a first preset time and a second preset time. The time range of the first preset time is 10-30s, and the time range of the second preset time is 10-20s.

[0090] In a possible implementation, the device may further include a preprocessing module, which may be used to:

[0091] Use filtering algorithm to remove noise points from defect boundary points of each layer;

[0092] Accordingly, the defect type determination model can be used to:

[0093] Based on all defect boundary points after noise point removal, the defect three-dimensional boundary point data image of the target steel sample is determined.

[0094] In one possible implementation, the defect type determination model can be used to:

[0095] Using the Poisson surface reconstruction algorithm, all defect boundary points are reconstructed into a three-dimensional defect boundary point data image of the target steel sample.

[0096] In a possible implementation, the process of building the defect type discrimination model is as follows:

[0097] Obtaining defect three-dimensional boundary point data images of historical steel samples, and obtaining defect types of historical steel samples;

[0098] Use machine learning algorithms to build a defect type identification model;

[0099] The defect three-dimensional boundary point data image of the historical steel sample is taken as input, and the defect type of the historical steel sample is taken as output to train the defect type discrimination model.

[0100] In a possible implementation, the device may further include a defect degree judgment module, which may be used to:

[0101] Acquire the size parameters of the defect three-dimensional boundary point data image of the target steel sample, the size parameters including volume parameters and shape parameters;

[0102] The size parameters and defect types are input into the defect degree determination model, and the defect grade of the target steel sample is output to quantify the defect degree of the target steel sample.

[0103] Figure 3 is a schematic diagram of a terminal provided in an embodiment of the present application. Figure 3 As shown, the terminal 3 of this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, the steps in the above-mentioned surface defect analysis method embodiments are implemented, for example Figure 1 Alternatively, when the processor 30 executes the computer program 32, the functions of each module / unit in the above-mentioned device embodiments are realized, for example, Figure 2 The functions of each module are shown.

[0104] Exemplarily, the computer program 32 may be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 30 to complete the present application. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program 32 in the terminal 3. For example, the computer program 32 may be divided into Figure 2 The modules shown.

[0105] The terminal 3 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal 3 may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art will appreciate that Figure 3 It is only an example of terminal 3 and does not constitute a limitation on terminal 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal may also include input and output devices, network access devices, buses, etc.

[0106] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0107] The memory 31 may be an internal storage unit of the terminal 3, such as a hard disk or memory of the terminal 3. The memory 31 may also be an external storage device of the terminal 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal 3. Further, the memory 31 may also include both an internal storage unit and an external storage device of the terminal 3. The memory 31 is used to store the computer program and other programs and data required by the terminal. The memory 31 may also be used to temporarily store data that has been output or is to be output.

[0108] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0109] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0110] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] In the embodiments provided in the present application, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0112] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0113] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0114] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned surface defect analysis method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0115] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for analyzing steel surface defects, characterized in that: include: Acquire a first image and a second image of each surface layer of a target steel sample, wherein the first image is an initial image of each surface layer of the target steel sample, and the second image is an image of each surface layer of the target steel sample after corrosion; Inputting the first image and the second image of each layer into the defect boundary point generation model respectively, and outputting the defect boundary points of the corresponding layer; Based on all defect boundary points, a three-dimensional defect boundary point data image of the target steel sample is determined, and the three-dimensional defect boundary point data image of the target steel sample is input into a defect type discrimination model to determine the defect type of the target steel sample, wherein the defect type discrimination model is constructed based on a machine learning algorithm; The step of acquiring the first image and the second image of each surface layer of the target steel sample comprises: For each surface layer of the target steel sample, perform the following steps: Acquire an initial image of the target steel sample on the surface of the layer, and use the initial image as the first image; After the surface of the target steel sample layer is continuously corroded by the preset corrosive liquid for a preset time, a surface image of the target steel sample layer after the corrosion is obtained, and the surface image of the layer after the corrosion is used as the second image, and after the second image of the layer is obtained after the corrosion is completed, the corroded surface of the layer is polished to ensure that the surface of each layer obtained in the first image is flat; The preset etching liquid includes a first preset etching liquid and a second preset etching liquid, the first preset etching liquid is a mixed solution of 4% nitric acid alcohol and ferric chloride solution, and the second preset etching liquid is a solution of 100 ml of pure water and 3.5 g of sodium chloride; The preset time includes a first preset time and a second preset time, the time range of the first preset time is 10-30s, and the time range of the second preset time is 10-20s.

2. The steel surface defect analysis method according to claim 1, characterized in that: Before determining the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points, the method further includes: Use filtering algorithm to remove noise points from defect boundary points of each layer; Accordingly, the method of determining the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points includes: Based on all defect boundary points after noise point removal, a three-dimensional defect boundary point data image of the target steel sample is determined.

3. The steel surface defect analysis method according to claim 1, characterized in that: The method of determining the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points includes: Using the Poisson surface reconstruction algorithm, all defect boundary points are reconstructed into a three-dimensional defect boundary point data image of the target steel sample.

4. The steel surface defect analysis method according to claim 1, characterized in that: The construction process of the defect type discrimination model is as follows: Obtaining a three-dimensional defect boundary point data image of a historical steel sample, and obtaining the defect type of the historical steel sample; Use machine learning algorithms to build a defect type identification model; The defect three-dimensional boundary point data image of the historical steel sample is used as input, and the defect type of the historical steel sample is used as output to train the defect type discrimination model.

5. The steel surface defect analysis method according to claim 1, characterized in that: After inputting the defect three-dimensional boundary point data image of the target steel sample into the defect type discrimination model to determine the defect type of the target steel sample, the method further includes: Acquire the size parameters of the defect three-dimensional boundary point data image of the target steel sample, wherein the size parameters include volume parameters and shape parameters; The size parameter and the defect type are input into a defect degree determination model, and the defect grade of the target steel sample is output to quantify the defect degree of the target steel sample.

6. A steel surface defect analysis device, characterized in that: include: An image acquisition module, used to acquire a first image and a second image of each surface layer of a target steel sample, wherein the first image is an initial image of each surface layer of the target steel sample, and the second image is an image of each surface layer of the target steel sample after corrosion; A boundary point determination module, used to input the first image and the second image of each layer into the defect boundary point generation model respectively, and output the defect boundary points of the corresponding layer; A defect type determination module is used to determine the defect three-dimensional boundary point data image of the target steel sample based on all defect boundary points, and input the defect three-dimensional boundary point data image of the target steel sample into a defect type discrimination model to obtain the inclusion type in the target steel sample, wherein the defect type discrimination model is constructed based on a machine learning algorithm; Wherein, the image acquisition module is used for: For each surface layer of the target steel sample, perform the following steps: Acquire an initial image of the target steel sample on the surface of the layer, and use the initial image as the first image; After the surface of the target steel sample layer is continuously corroded by the preset corrosive liquid for a preset time, a surface image of the target steel sample layer after the corrosion is obtained, and the surface image of the layer after the corrosion is used as the second image, and after the second image of the layer is obtained after the corrosion is completed, the corroded surface of the layer is polished to ensure that the surface of each layer obtained in the first image is flat; The preset etching liquid includes a first preset etching liquid and a second preset etching liquid, the first preset etching liquid is a mixed solution of 4% nitric acid alcohol and ferric chloride solution, and the second preset etching liquid is a solution of 100 ml of pure water and 3.5 g of sodium chloride; The preset time includes a first preset time and a second preset time, the time range of the first preset time is 10-30s, and the time range of the second preset time is 10-20s.

7. A terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the steel surface defect analysis method as described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the steel surface defect analysis method as described in any one of claims 1 to 5 are implemented.

Images