A method and system for identifying inclusions in steel

Abstract

The present application provides a method and system for identifying inclusions in steel, and relates to the technical field of inclusion detection. The method comprises: polishing the steel sample to be identified N times according to the target angle, and etching the steel sample to be identified after each polishing in a chemical corrosive agent at a preset temperature, where N is a positive integer greater than or equal to 1; obtaining a surface image of the etched surface of the steel sample to be identified after each etching; inputting each surface image into a corrosion feature model, and outputting the corrosion features of the corresponding surface image; and determining the contour morphology of the inclusions in the steel sample to be identified based on the corrosion features of each surface image. The present application can more accurately identify the contour morphology of specific inclusions in steel samples.

Description

Technical Field

[0001] The present application relates to the technical field of inclusion detection, and in particular to a method and system for identifying inclusions in steel. Background Art

[0002] At present, in the manufacturing of medium manganese steel, some defects often occur in the smelting stage, and inclusions are one of them. Non-metallic inclusions in steel destroy the continuity and uniformity of steel, and affect the processing performance, mechanical properties and fatigue performance of steel. Therefore, it is necessary to accurately control inclusions, and accurate detection and characterization of inclusions in steel is the premise for accurate control of inclusions. Summary of the invention

[0003] The present application provides a method and system for identifying inclusions in steel, so as to solve the problem of inaccurate identification of inclusions in steel in the prior art.

[0004] In a first aspect, the present application provides a method for identifying inclusions in steel, comprising:

[0005] The steel sample to be identified is polished N times according to the target angle, and the steel sample to be identified after each polishing is placed in a chemical corrosive agent at a preset temperature for etching, where N is a positive integer greater than or equal to 1;

[0006] Acquire a surface image of the etched surface of the steel sample to be identified after each etching;

[0007] Input each surface image into the corrosion feature model respectively, and output the corrosion features of the corresponding surface image;

[0008] Based on the corrosion characteristics of each surface image, the contour morphology of the inclusions in the steel sample to be identified is determined.

[0009] In a second aspect, the present application provides a system for identifying inclusions in steel, the system comprising a polishing device, a chemical etching device, an image collector and a data processor;

[0010] The polishing device is used to polish the steel sample to be identified N times according to the target angle, where N is a positive integer greater than or equal to 1;

[0011] The chemical etching device is used to place the steel sample to be identified after each polishing into a chemical etching agent at a preset temperature for etching;

[0012] The image collector is used to collect the surface image of the etched surface of the steel sample to be identified after each etching, and send each surface image to the data processor;

[0013] The data processor is used to input each surface image into the corrosion feature model, output the corrosion feature of the corresponding surface image, and determine the contour morphology of the inclusions in the steel sample to be identified based on the corrosion feature of each surface image.

[0014] The present application provides a method and system for identifying inclusions in steel, which polishes the steel sample to be identified N times according to the target angle, and puts the steel sample to be identified after each polishing into a chemical corrosive agent at a preset temperature for etching, where N is a positive integer greater than or equal to 1; obtains the surface image of the etched surface of the steel sample to be identified after each etching; inputs each surface image into a corrosion feature model, and outputs the corrosion features of the corresponding surface image; and determines the contour morphology of the inclusions in the steel sample to be identified based on the corrosion features of each surface image. The present application can gradually reveal the microstructure inside the steel sample through multiple polishing and chemical etching treatments, making the inclusions more clearly visible, and then combines the corrosion feature model to perform feature recognition on the surface image, which can more accurately identify the contour morphology of specific inclusions in the steel sample, avoiding errors caused by traditional human judgment or equipment limitations. 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 method for identifying inclusions in steel provided in the embodiment of the present application;

[0017] Figure 2 is a schematic diagram of the target angle of a steel sample provided in an embodiment of the present application;

[0018] Figure 3 It is a schematic diagram of the structure of the steel inclusion identification system provided in the 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] Since non-metallic inclusions in steel will destroy the continuity and uniformity of the steel and affect the processing performance, mechanical properties and fatigue properties of the steel, it is necessary to accurately control the inclusions in the steel. Accurate detection and characterization of inclusions in steel is the prerequisite for accurate control of inclusions.

[0022] The present application provides a method for identifying inclusions in steel, including corrosion and three-dimensional characterization design of CuO or Cu2O inclusions in medium manganese steel, that is, three-dimensional reconstruction is performed by an automatic cycle process of corrosion surface-recognition image-grinding. Among them, pits are left by chemically corroding CuO or Cu2O, and the pits are three-dimensionally reconstructed to achieve identification of CuO or Cu2O inclusions.

[0023] Figure 1 The implementation flow chart of the method for identifying inclusions in steel provided in the embodiment of the present application is described in detail as follows:

[0024] In step 101, the steel sample to be identified is polished N times according to the target angle, and the steel sample to be identified after each polishing is placed in a chemical corrosive agent at a preset temperature for etching, where N is a positive integer greater than or equal to 1.

[0025] In this embodiment, the polishing times N can be adjusted according to actual needs to meet the needs of identifying inclusions of different types and sizes.

[0026] In an embodiment of the present application, the steel sample to be identified is polished N times according to the target angle, and then, for each polishing, the steel sample to be identified after the polishing is placed in a chemical corrosive agent and etched at a preset temperature to obtain the steel sample to be identified after each etching.

[0027] Etching is to put the steel sample to be identified after each polishing into a chemical corrosive agent, so that the inclusions in the steel sample to be identified are corroded by the chemical corrosive agent, resulting in pits on the surface of the steel sample to be identified.

[0028] The chemical corrosive agent at the preset temperature in the embodiment of the present application can also be optimized and selected according to the specific composition and inclusion type of the steel sample, thereby improving flexibility and applicability.

[0029] For example, when the steel sample to be identified is medium manganese steel, and the inclusions in the steel sample to be identified are CuO or Cu2O, the chemical corrosive agent used is an ethanol reagent containing perovskite-type composite oxides and spinel-type composite oxides, that is, perovskite-type composite oxides (LaMnO3, LaFeO3) and spinel-type composite oxides (MgAl2O4, ZnFe2O4) are added to ethanol as catalysts. The active sites on the surface of the two composite oxides can adsorb ethanol molecules and CuO or Cu2O, increasing their concentration on the catalyst surface, thereby increasing the reaction probability. At the same time, electron transfer may occur between the catalyst and the reactant, promoting the dehydrogenation process of the hydroxyl group in the ethanol molecule and the reduction process of CuO or Cu2O, and accelerating the entire reaction. In addition, the lattice oxygen of the catalyst may also participate in the reaction, undergoing redox reactions with the reactants, further promoting the reaction.

[0030] At the same time, during the etching process, a constant temperature space with the optimal reaction temperature required for the chemical corrosion reaction needs to be provided so that the etching can be carried out efficiently and stably.

[0031] When the preset temperature of the etching setting is less than 250°C, the chemical corrosive agent may not react completely or not react with CuO or Cu2O, and the etching effect cannot be achieved. When the preset temperature of the etching setting is greater than 350°C, ethanol will react with iron oxide. For example, at 350°C-500°C, ethanol can reduce Fe2O3 to Fe3O4; at 500°C-600°C, it will be further reduced to Fe3O4 / FeO; when the temperature reaches 700°C, it will be reduced to Fe. This will make the surface image of the target steel sample after etching too much compared with the actual demand, and it is impossible to accurately obtain the contour morphology of CuO or Cu2O required later.

[0032] Therefore, in the embodiment of the present application, the preset temperature range for the inclusions CuO or Cu2O is 300°C. 50℃, that is, 250℃~350℃.

[0033] In the embodiment of the present application, a constant temperature space can also be built based on the etching requirements, and the construction process is as follows:

[0034] (1) Shell design: Choose high temperature resistant materials, such as stainless steel or high temperature alloys, to ensure that they can withstand high temperature environments. It is also necessary to design a sealing structure to ensure the sealing of the reaction space and prevent heat loss.

[0035] (2) Placing platform: The placing platform has an angle control function. According to the inclination angle of the target steel sample during polishing, the placing platform is controlled to tilt according to the inclination angle of the target steel sample during polishing, and then the target steel sample is placed on the tilted placing platform to etch the polished surface of the target steel sample.

[0036] (3) Heating system: Select an electric heating rod or heating wire of appropriate power, and use a temperature controller to accurately control the operation of the heating element to maintain a stable temperature.

[0037] (4) Insulation layer design: Use high-quality insulation materials, such as aluminum silicate fiber, to reduce heat loss and ensure insulation effect.

[0038] (5) Temperature monitoring system: Install a temperature sensor to monitor the temperature in the reaction space in real time and connect it to the temperature controller for real-time feedback and adjustment.

[0039] (6) Safety measures: Install an over-temperature protection device to prevent danger caused by excessive temperature, and install a pressure relief device to ensure safety.

[0040] Based on the above-mentioned chemical etchant and preset temperature design, the steel sample to be identified after each polishing is etched at a preset temperature of 250°C to 350°C using an ethanol reagent containing perovskite-type composite oxides and spinel-type composite oxides. The reaction equation is as follows:

[0041]

[0042] The red color of the Cu element generated by the reaction product can be used as another proof to identify that the inclusion is CuO or Cu2O.

[0043] The embodiment of the present application can gradually expose different layers of inclusions in the steel through multiple polishing and etching processes, so that the contour morphology of the inclusions is clearer after multiple etchings, which helps to more accurately identify the type and morphology of the inclusions, thereby improving the accuracy of identification.

[0044] Regarding polishing, in a possible implementation, the steel sample to be identified is polished N times according to the target angle, which may include:

[0045] Tilt the steel sample to be identified at the target angle;

[0046] The steel sample to be identified is ground N times with the target sandpaper at the target tilt angle;

[0047] After each grinding, the steel sample was polished using diamond paste on a mechanical polishing machine.

[0048] Since there is an oxide film on the surface of the steel sample to be identified, it is first necessary to use coarse sandpaper to polish it to remove the oxide film on the surface. The rough polishing process is as follows:

[0049] The embodiments of the present application are mainly for the observation of metallographic surface pits, and due to the influence of factors such as material properties and observation purpose, the actual grinding angle needs to be adjusted according to the specific situation, that is, the steel sample to be identified is first tilted according to the target angle, and then the target sandpaper of the target specification is used to grind the steel sample to be identified after being tilted at the target angle N times.

[0050] The target angle may range from 15° to 30°.

[0051] The target angle of the embodiment of the present application is set within this angle range, which can better highlight the shape and characteristics of the pits, while also ensuring a certain degree of flatness and clarity, which helps to more accurately observe and analyze the pits on the metallographic surface.

[0052] The target sandpaper is aluminum oxide sandpaper, and the specification range of the target sandpaper is 400~2000.

[0053] For example, refer to Figure 2 , tilt the steel sample Then use the target sandpaper to adjust the tilt The edge A of the angled steel sample is ground N times.

[0054] Then after each grinding, the steel sample to be identified after each grinding is placed in a mechanical polishing machine, and the grinding surface of the steel sample to be identified after each grinding is polished with diamond grinding paste to remove surface scratches and obtain a smooth mirror surface for subsequent observation and imaging.

[0055] In step 102, a surface image of the etched surface of the steel sample to be identified after each etching is obtained.

[0056] In the embodiment of the present application, a high-resolution optical microscope or electron microscope is used to observe the etched surface of the steel sample to be identified after each etching, and a corresponding two-dimensional image, ie, a surface image, is collected.

[0057] In step 103, each surface image is input into the corrosion feature model, and the corrosion feature of the corresponding surface image is output.

[0058] In the embodiment of the present application, each surface image is input into the constructed corrosion feature model to obtain the corrosion features of the corresponding surface image, wherein the corrosion features may include corrosion pits, cracks, and the like.

[0059] In a possible implementation, before inputting each surface image into the corrosion feature model and outputting the corrosion feature of the corresponding surface image, the method may further include:

[0060] After the surface images are collected, each surface image is preprocessed with the help of image processing software, such as denoising, contrast enhancement, etc., to improve the image quality of the surface image.

[0061] In a possible implementation, the process of constructing the corrosion characteristic model may be:

[0062] Obtaining a surface image of the etched surface of the historical steel sample after etching, and obtaining corrosion characteristics corresponding to the surface image of the etched surface of the historical steel sample after etching;

[0063] Construct corrosion characteristic models;

[0064] The surface image of the etched surface of the historical steel sample after etching is taken as input, and the corrosion features corresponding to the surface image of the etched surface of the historical steel sample after etching are taken as output to train the corrosion feature model.

[0065] Optionally, the construction process of the corrosion characteristic model constructed in this embodiment is:

[0066] A large number of surface images of the etched surfaces of historical steel samples after etching and the corrosion features corresponding to the surface images of each historical steel sample are obtained. Then, a corrosion feature model is built, and the surface image of each historical steel sample is used as input and the corresponding corrosion features are used as output to train the corrosion feature model.

[0067] Among them, the corrosion feature model can be constructed based on an image recognition algorithm, based on a feature recognition method, or based on other recognition methods, depending on the specific situation.

[0068] The embodiment of the present application combines chemical etching and image processing technology, and by inputting the surface image after each etching into the corrosion feature model, feature information related to inclusions can be extracted. This multi-step, multi-feature analysis method enhances the reliability of identification and reduces the possibility of misjudgment.

[0069] In step 104, the contour morphology of the inclusions in the steel sample to be identified is determined based on the corrosion features of each surface image.

[0070] In the embodiment of the present application, the corrosion features of each surface image obtained in step 103 are used to determine the contour morphology of the inclusions in the steel sample to be identified.

[0071] In a possible implementation, determining the contour morphology of inclusions in the steel sample to be identified based on the corrosion characteristics of each surface image may include:

[0072] According to the polishing order, the corrosion characteristics of each surface image are input into the three-dimensional reconstruction model in turn, and the three-dimensional model of the inclusions in the steel sample to be identified is output;

[0073] The three-dimensional model of inclusions is used to determine the contour morphology of inclusions in the steel sample to be identified.

[0074] Optionally, the corrosion features of each surface image are input into the three-dimensional reconstruction model in turn according to the polishing order to obtain a three-dimensional model of the inclusions in the steel sample to be identified, that is, a three-dimensional model of the corrosion morphology is constructed in three-dimensional space. Then, the three-dimensional model of the inclusions is used to determine the contour morphology of the inclusions in the steel sample to be identified.

[0075] Among them, the construction of the three-dimensional reconstruction model can be based on a variety of methods, such as NeRF or other structures.

[0076] In a possible implementation, the three-dimensional reconstruction model is constructed based on NeRF, and the construction process of the three-dimensional reconstruction model is:

[0077] sequentially obtaining corrosion characteristics of the surface image of each etched surface of the historical steel sample and obtaining a three-dimensional model of inclusions in the historical steel sample;

[0078] Build a 3D reconstruction model based on NeRF;

[0079] The corrosion features of the surface images of each etched surface of the historical steel specimens are taken as input, and the 3D models of inclusions in the historical steel specimens are taken as output to train the 3D reconstruction model.

[0080] Optionally, the corrosion features of the surface image of each etched surface of the historical steel sample are obtained in sequence, and the three-dimensional model of the inclusions in the historical steel sample is obtained, and then the three-dimensional reconstruction model is constructed using NeRF. The corrosion features of the surface image of each etched surface of the historical steel sample are used as input, and the three-dimensional model of the inclusions in the historical steel sample is used as output to train the three-dimensional reconstruction model.

[0081] In a possible implementation, before using the three-dimensional model of the inclusion to determine the contour morphology of the inclusion in the steel sample to be identified, the method may further include:

[0082] The three-dimensional model of the inclusions is post-processed, including smoothing, optimizing the structure, etc., to improve the accuracy and authenticity of the three-dimensional model.

[0083] The present application provides a method for identifying inclusions in steel, which includes polishing the steel sample to be identified N times according to the target angle, and etching the steel sample to be identified in a chemical corrosive agent at a preset temperature after each polishing, where N is a positive integer greater than or equal to 1; obtaining a surface image of the etched surface of the steel sample to be identified after each etching; inputting each surface image into a corrosion feature model, and outputting the corrosion features of the corresponding surface image; and determining the contour morphology of the inclusions in the steel sample to be identified based on the corrosion features of each surface image. The present application can gradually reveal the microstructure inside the steel sample through multiple polishing and chemical corrosion treatments, making the inclusions more clearly visible, and then combining the corrosion feature model to perform feature recognition on the surface image, which can more accurately identify the contour morphology of specific inclusions in the steel sample, avoiding errors caused by traditional human judgment or equipment limitations.

[0084] 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.

[0085] The following is a system embodiment of the present application. For details not described in detail therein, please refer to the corresponding method embodiment described above.

[0086] The embodiment of the present application can determine the contour morphology of inclusions in steel by manual operation or by automatic cycle operation.

[0087] Based on the above-mentioned steel inclusion identification method, refer to Figure 3 The embodiment of the present application also provides a system for identifying inclusions in steel, which includes a polishing device 31, a chemical etching device 32, an image collector 33 and a data processor 34, and the polishing device 31, the chemical etching device 32, the image collector 33 and the data processor 34 are connected in sequence.

[0088] The polishing device 31 is used to polish the steel sample to be identified N times according to the target angle, where N is a positive integer greater than or equal to 1.

[0089] The chemical etching device 32 is used to place the steel sample to be identified after each polishing into a chemical etching agent at a preset temperature for etching.

[0090] The image collector 33 is used to collect the surface image of the etched surface of the steel sample to be identified after each etching, and send each surface image to the data processor 34 .

[0091] The data processor 34 is used to input each surface image into the corrosion feature model, output the corrosion feature of the corresponding surface image, and determine the contour morphology of the inclusions in the steel sample to be identified based on the corrosion feature of each surface image.

[0092] To realize the functions of both manual and automatic processes, refer to Figure 3 In the embodiment of the present application, the inclusion identification system in steel further includes a control device 35 and a robot arm 36 , and the control device 35 includes a manual module 351 and an automatic module 352 .

[0093] The control device 35 is respectively connected to the polishing device 31, the chemical etching device 32, the image acquisition device 33, the data processor 34 and the robotic arm 36, and is used to obtain the work instructions triggered by the manual module 351 or the automatic module 352 input by the staff, and enter the manual mode or the automatic mode according to the work instructions, wherein the work instructions are manual instructions and automatic instructions.

[0094] The mechanical arm 36 is used to clamp the steel sample to be identified and move it in the polishing device 31, the chemical etching device 32 and the image collector 33 in sequence, so as to realize the automatic polishing, automatic etching and automatic image collection functions of the steel sample to be identified.

[0095] In order to realize the automatic cycle process, a corresponding automation script or program can be written to support the automatic mode. The script or program should be able to control the movement of the image collector, the acquisition and processing of images, the construction of the three-dimensional model and other steps to realize the automation of the process. During the automatic cycle, the appropriate number of cycles and stop conditions should be set to ensure the smooth progress of the process and the accuracy of the results.

[0096] Exemplarily, when the staff presses the manual module 351, the entire inclusion identification system in steel enters the manual mode, that is, the staff puts the steel sample to be identified into the polishing device 31 for polishing, puts it into the chemical corrosion device 32 for etching, and places the etched steel sample to be identified at the position of the image collector 33 to collect the surface image, and inputs the surface image into the data processor 34 to determine the contour morphology of the inclusion.

[0097] When the staff presses the automatic module 352, the entire inclusion identification system in steel enters the automatic mode, that is, the robot arm 36 is controlled to clamp the steel sample to be identified and pass it through the polishing device 31, the chemical corrosion device 32 and the image collector 33 in sequence to obtain the surface image of the steel sample to be identified, and the surface image is sent to the data processor 34 for analysis to obtain the contour morphology of the inclusions in the steel sample to be identified.

[0098] The present application provides a system for identifying inclusions in steel, which includes a polishing device, a chemical etching device, an image collector and a data processor; the polishing device is used to polish the steel sample to be identified N times according to the target angle, N is a positive integer greater than or equal to 1; the chemical etching device is used to put the steel sample to be identified after each polishing into a chemical etchant at a preset temperature for etching; the image collector is used to collect the surface image of the etched surface of the steel sample to be identified after each etching, and send each surface image to the data processor; the data processor is used to input each surface image into a corrosion feature model, output the corrosion feature of the corresponding surface image, and determine the contour morphology of the inclusions in the steel sample to be identified based on the corrosion feature of each surface image. The present application can gradually reveal the microstructure inside the steel sample through multiple polishing and chemical etching treatments, making the inclusions more clearly visible, and then combine the corrosion feature model to perform feature recognition on the surface image, which can more accurately identify the contour morphology of specific inclusions in the steel sample, avoiding the errors caused by traditional human judgment or equipment limitations.

[0099] 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 identifying inclusions in steel, characterized in that: include: The steel sample to be identified is polished N times according to the target angle, and the steel sample to be identified after each polishing is placed in a chemical corrosive agent at a preset temperature for etching, where N is a positive integer greater than 1; Acquire a surface image of the etched surface of the steel sample to be identified after each etching; Input each surface image into the corrosion feature model respectively, and output the corrosion features of the corresponding surface image; Determining the contour morphology of the inclusions in the steel sample to be identified based on the corrosion characteristics of each surface image; The step of polishing the steel sample to be identified N times according to the target angle includes: Tilt the steel sample to be identified to the target angle; Grinding the steel sample to be identified which is inclined at the target angle N times using target sandpaper; After each grinding, the steel sample after grinding is polished on a mechanical polishing machine using diamond grinding paste; the angle range of the target angle is 15°~30°; When the steel sample to be identified is medium manganese steel and the inclusion is CuO or Cu2O, the preset temperature ranges from 250°C to 350°C, and the chemical corrosive agent is an ethanol reagent containing perovskite-type composite oxides and spinel-type composite oxides; Wherein, determining the contour morphology of the inclusions in the steel sample to be identified based on the corrosion characteristics of each surface image includes: Inputting the corrosion features of each surface image into the three-dimensional reconstruction model in turn according to the polishing order, and outputting the three-dimensional model of the inclusions in the steel sample to be identified; The three-dimensional model of the inclusions is used to determine the contour morphology of the inclusions in the steel sample to be identified.

2. The method for identifying inclusions in steel according to claim 1, characterized in that: The target sandpaper is aluminum oxide sandpaper, and the specification range of the target sandpaper is 400-2000.

3. The method for identifying inclusions in steel according to claim 1, characterized in that: The construction process of the corrosion characteristic model is as follows: Obtaining a surface image of the etched surface of the historical steel sample after etching, and obtaining corrosion features corresponding to the surface image of the etched surface of the historical steel sample after etching; Construct corrosion characteristic models; The surface image of the etched surface of the historical steel sample after etching is used as input, and the corrosion features corresponding to the surface image of the etched surface of the historical steel sample after etching are used as output to train the corrosion feature model.

4. The method for identifying inclusions in steel according to claim 1, characterized in that: The three-dimensional reconstruction model is constructed based on NeRF, and the construction process of the three-dimensional reconstruction model is as follows: sequentially acquiring corrosion characteristics of a surface image of each etched surface of a historical steel sample, and acquiring a three-dimensional model of inclusions in the historical steel sample; Build a 3D reconstruction model based on NeRF; The corrosion features of the surface image of each etched surface of the historical steel sample are used as input, and the three-dimensional model of inclusions in the historical steel sample is used as output to train the three-dimensional reconstruction model.

5. A system for identifying inclusions in steel, characterized in that: The system performs the method for identifying inclusions in steel as described in any one of claims 1 to 4 above, and the system comprises a polishing device, a chemical etching device, an image collector and a data processor; The polishing device is used to polish the steel sample to be identified N times according to the target angle, where N is a positive integer greater than 1; The chemical etching device is used to place the steel sample to be identified after each polishing into a chemical etching agent at a preset temperature for etching; The image collector is used to collect the surface image of the etched surface of the steel sample to be identified after each etching, and send each surface image to the data processor; The data processor is used to input each surface image into the corrosion feature model, output the corrosion feature of the corresponding surface image, and determine the contour morphology of the inclusions in the steel sample to be identified based on the corrosion feature of each surface image.

Images