X-ray flaw detection virtual simulation teaching system

Abstract

The invention discloses an X-ray flaw detection virtual simulation teaching system, and relates to the technical field of industrial flaw detection teaching systems, the system comprises an operation box, a computer all-in-one machine and an X-ray generator simulation module; the all-in-one computer and the X-ray generator simulation module are connected with the operation box; the X-ray generator simulation module sends a signal to the operation box to simulate the starting state of the X-ray generator; the operation box is used for starting the X-ray generator simulation module and setting exposure parameters; an X-ray flaw detection virtual simulation teaching model is deployed on the computer all-in-one machine; the computer all-in-one machine is used for interacting with the student user and carrying out flaw detection evaluation teaching on the student user based on the X-ray flaw detection virtual simulation teaching model, the X-ray flaw detection virtual simulation teaching model is built, multiple modules in the flaw detection teaching process can be integrated, and the automation level of the X-ray flaw detection virtual simulation teaching system is improved.

Description

Technical Field

[0001] The present application relates to the technical field of industrial flaw detection teaching systems, and in particular to an X-ray flaw detection virtual simulation teaching system. Background Art

[0002] X-ray flaw detection is a method of detecting internal defects in materials by using X-rays to penetrate metal materials and form images of different blackness on the film due to the different absorption and scattering effects of the materials on the rays. Currently, the application of X-ray flaw detection is relatively mature, but the identification of film defects still mainly relies on manpower. The identification of defects has extremely high requirements on the experience and concentration of the flaw detectors, and due to the limitation of human resources, it is difficult to carry out large-scale defect identification work in a short period of time.

[0003] Unity is a real-time 3D interactive content creation and operation platform. All creators, including game development, art, architecture, automotive design, and film and television, use Unity to turn their ideas into reality. The Unity platform provides a complete set of software solutions that can be used to create, operate, and monetize any real-time interactive 2D and 3D content, supporting platforms including mobile phones, tablets, PCs, game consoles, augmented reality, and virtual reality devices. Through Unity modeling technology, a three-dimensional, high-degree-of-freedom, virtual industrial flaw detection scene can be built to provide students with flaw detection learning and training. The entire industrial process of X-ray nondestructive flaw detection includes four major steps: bagging (X-ray film, sample plate loaded into dark bag), cloth (placing the dark bag in the X-ray generator), film washing (using the X-ray generator for exposure processing), and film evaluation (determining defects based on the exposed film). The steps of bagging cloth pieces need to be carried out in laboratories and darkrooms. The environment construction requirements are high and it is difficult to meet the learning and training venue needs when there are a large number of student users. In addition, X-rays have radiation, which poses certain safety hazards in the actual learning and training process. The existing X-ray flaw detection simulation equipment only simulates the specific step of film washing and cannot fully restore the flaw detection process. The existing flaw detection simulation system mainly conducts training by randomly selecting images from the database for student users to identify. The modules are relatively scattered and do not have a three-in-one process-based training program for teaching, training, and assessment. Summary of the invention

[0004] The purpose of this application is to provide an X-ray flaw detection virtual simulation teaching system that can integrate multiple modules in the flaw detection teaching process and improve the automation level of the X-ray flaw detection virtual simulation teaching system.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] The present application provides an X-ray flaw detection virtual simulation teaching system, comprising: an operation box, an all-in-one computer and an X-ray generator simulation module;

[0007] The all-in-one computer and the X-ray generator simulation module are both connected to the operation box;

[0008] The X-ray generator simulation module sends a signal to the operation box to simulate the on state of the X-ray generator;

[0009] The operation box is used to start the X-ray generator simulation module and set exposure parameters; the exposure parameters include exposure voltage and exposure time; the operation box is also used to train the machine; the operation box is also used to display the operation box status, X-ray generator start status display and training status display;

[0010] An X-ray flaw detection virtual simulation teaching model is deployed on the all-in-one computer; the all-in-one computer is used to interact with student users and conduct flaw detection evaluation teaching for student users based on the X-ray flaw detection virtual simulation teaching model.

[0011] Optionally, the X-ray flaw detection virtual simulation teaching system further includes a bracket;

[0012] The bracket is used for supporting the X-ray generator.

[0013] Optionally, the X-ray flaw detection virtual simulation teaching model includes: a usage mode selection module, a teaching component and an assessment component;

[0014] The teaching component and the assessment component are both connected to the usage mode selection module;

[0015] The usage mode selection module is used to display a mode selection interface, obtain a mode selection instruction, and when the mode selection instruction is a teaching mode instruction, perform flaw assessment teaching on student users based on the teaching component; when the mode selection instruction is an assessment mode instruction, perform flaw assessment on student users based on the assessment component.

[0016] Optionally, the teaching component includes: a process selection module, an experiment overview display module, an experiment preparation teaching module, an experiment implementation teaching module and a teaching experiment report module;

[0017] The experiment overview display module, the experiment preparation teaching module, the experiment implementation teaching module and the teaching experiment report module are all connected to the process selection module;

[0018] The process selection module is used to display the teaching mode interface and obtain the process selection instruction when receiving the teaching mode instruction; the process selection instruction includes the experiment overview display instruction, the experiment preparation teaching instruction, the experiment implementation teaching instruction and the teaching experiment report instruction;

[0019] The experiment overview display module is used to call and display the experiment overview and experiment outline when receiving the experiment overview display instruction, and call the process selection module after receiving the reading completion instruction;

[0020] The experiment preparation teaching module is used to call and display the experiment scene and the experiment equipment model after receiving the experiment preparation teaching instruction, and call the process selection module after receiving the return instruction; the experiment scene includes: laboratory scene, darkroom scene and exposure scene;

[0021] The experiment implementation teaching module is used to prompt the student user to perform bagging training, cloth piece training, film washing training and film evaluation training after receiving the experiment implementation teaching instruction, and call the process selection module after receiving the return instruction;

[0022] The teaching experiment report module is used to generate and display the student user teaching experiment report when receiving the teaching experiment report instruction, and call the process selection module after receiving the return instruction.

[0023] Optionally, the experiment implementation teaching module includes: a bagging training unit, a cloth piece training unit, a film washing training unit and a film evaluation training unit;

[0024] After receiving the teaching instruction for implementing the experiment, the bagging training unit calls the darkroom scene and guides the student user to select experimental equipment for bagging operation training, and then calls the cloth piece training unit;

[0025] The film training unit is used to call the exposure scene and guide the student user to set the exposure parameters and perform film operation training through the buttons on the operation box, and then call the film processing training unit;

[0026] The film processing training unit is used to call the darkroom scene and guide the student user to perform film processing operation training when receiving the film processing training instruction, and then call the film evaluation training unit;

[0027] The film evaluation training unit is used to randomly select a film image from a film image database as a film to be evaluated, obtain the student user's evaluation result on the film to be evaluated as the student user's evaluation result, input the film to be evaluated into the evaluation model, obtain the system evaluation result, and correct the student user's evaluation result based on the system evaluation result; the film evaluation training unit is also used to obtain the auditor user's evaluation result on the film to be evaluated as the auditor user's evaluation result when receiving a manual evaluation instruction, and correct the student user's evaluation result based on the auditor user's evaluation result; the manual evaluation instruction is input by the student user when he does not agree with the system evaluation result; the film evaluation training unit is also used to call the process selection module after receiving a return instruction; the evaluation model is obtained by training a convolutional neural network using multiple film images marked with evaluation results.

[0028] Optionally, the convolutional neural network includes: a ResNet101 residual structure and an FPN feature pyramid network connected in sequence.

[0029] Optionally, the ResNet101 residual structure includes a convolutional layer, a maximum pooling downsampling layer, a residual network and a fully connected layer connected in sequence.

[0030] Optionally, before training the convolutional neural network, the film images used for training are subjected to average weighted grayscale processing.

[0031] Optionally, the assessment component includes: an objective question assessment module, an experiment preparation assessment module, an experiment implementation assessment module, and an assessment experiment report module connected in sequence;

[0032] The objective question assessment module is used to display the objective question assessment interface upon receiving the assessment mode instruction, select and display objective questions from the objective question database, obtain the student user's answer result to each objective question and then enter the experiment preparation assessment module; the assessment content of the objective question is the experiment overview and experiment outline;

[0033] The experiment preparation assessment module is used to obtain the scene selection results and experimental equipment selection results of the student user, and then enter the experiment implementation assessment module;

[0034] The experiment implementation assessment module is used to obtain the user's bagging operation results, cloth piece operation results, film washing operation results and film evaluation operation results, and then enter the assessment experiment report module;

[0035] The assessment experiment report module is used to generate and display a student user assessment experiment report based on the scene selection result, the experimental equipment selection result, the bagging operation result, the cloth operation result, the film washing operation result and the film evaluation operation result.

[0036] Optionally, the experiment implementation assessment module includes: an implementation selection unit, a bagging assessment unit, a cloth piece assessment unit, a film washing assessment unit and a film evaluation assessment unit;

[0037] The bagging assessment unit, the cloth piece assessment unit, the film washing assessment unit and the film evaluation assessment unit are all connected to the implementation selection unit;

[0038] The implementation selection unit is used to display the experiment implementation assessment interface and obtain implementation selection instructions; the implementation selection instructions include bagging assessment instructions, cloth piece assessment instructions, film washing assessment instructions and film evaluation assessment instructions;

[0039] The bagging assessment unit is used to call the scene selection result and the experimental equipment selection result when receiving the bagging assessment instruction, and obtain the bagging operation result, and call the implementation selection unit when receiving the return instruction;

[0040] The cloth piece assessment unit is used to call the scene selection result, the experimental equipment selection result and the bagging operation result when receiving the cloth piece assessment instruction, and obtain the cloth piece operation result, and call the implementation selection unit when receiving the return instruction;

[0041] The film processing and assessment unit is used to call the scene selection result, the experimental equipment selection result and the cloth operation result when receiving the film processing and assessment instruction, and obtain the film processing operation result, and call the implementation selection unit when receiving the return instruction;

[0042] The film evaluation unit is used to obtain a film evaluation type instruction when receiving a film processing evaluation instruction; the film evaluation type instruction includes a system film evaluation instruction and a manual film evaluation instruction;

[0043] The film evaluation and assessment unit randomly selects a film image from the film image database as the film to be evaluated upon receiving the system film evaluation instruction, and calls the assessment experiment report module after obtaining the student user's evaluation result on the film to be evaluated as the film evaluation operation result;

[0044] The film evaluation and assessment unit obtains the film image uploaded by the staff user as the film to be evaluated when receiving the manual film evaluation instruction, and obtains the evaluation result of the student user on the film to be evaluated as the film evaluation operation result, and then calls the assessment experiment report module.

[0045] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0046] The present application provides an X-ray flaw detection virtual simulation teaching system, which integrates the teaching, training and assessment of the industrial X-ray flaw detection process. It allows student users to complete X-ray flaw detection learning without radiation hazards in the form of Unity simulation modeling. At the same time, it is equipped with physical objects such as X-ray generators for student users to practice, which not only reduces the learning cost of student users but also standardizes the training process of X-ray flaw detection.

[0047] This application introduces a convolutional neural network algorithm to realize the recognition of X-ray films, and thereby assist student users in flaw detection identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments 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 work.

[0049] Figure 1 This is a schematic diagram of the structure of an X-ray flaw detection virtual simulation teaching system in one embodiment of the present application.

[0050] Figure 2 This is a circuit diagram of an operating box in one embodiment of the present application.

[0051] Figure 3 1. Three views of an X-ray generator in one embodiment of the present application.

[0052] Figure 4 This is a top view of the operating box in one embodiment of the present application.

[0053] Figure 5 This is a flow chart of the X-ray flaw detection teaching mode in one embodiment of the present application.

[0054] Figure 6 This is a flow chart of the X-ray flaw detection assessment mode in one embodiment of the present application.

[0055] Figure 7 Schematic diagram of the FPN feature pyramid network in one embodiment of the present application.

[0056] Figure 8 This is a schematic diagram of the experimental scene and equipment selection interface in one embodiment of the present application. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0058] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0059] X-ray flaw detection uses the difference in the degree of X-ray absorption due to different material thicknesses, and uses X-ray fluoroscopy and industrial television real-time imaging to show the internal defects of materials, parts and welds from film and imaging. Such as cracks, shrinkage cavities, pores, slag inclusions, unfused, incomplete penetration, etc., to determine the location and size. The quality of materials or products is evaluated by observing the nature, size and location of the defects, thereby preventing major accidents caused by internal defects of materials and poor processing. Ship tilt simulation experiment: Use Unity3D to establish a ship simulation model, and calculate the true center of gravity of the ship by adjusting the position of the ship's weights and measuring the ship's draft and other data under different conditions. With the development of deep learning technology, it has become a reality to use computers to replace human labor for image recognition. In recent years, many scholars have applied image recognition technology to the recognition of X-ray film defects of welds, which can achieve recognition accuracy close to or even beyond that of manual labor, greatly improving the efficiency of X-ray non-destructive flaw detection. Image features are extracted through convolutional neural networks, and the features are processed at different levels to identify image content. Common feature extraction algorithms include VGG, ResNet, etc., and common image recognition algorithms include YOLO, RCNN, Faster-RCNN, etc. Convolutional neural network image recognition is a branch of artificial intelligence. The convolutional neural network algorithm is used to extract the feature values ​​of the image and generate the corresponding feature layer. On the basis of the feature layer, a series of algorithms are used for training, and finally a model is generated. The model can recognize the input image and output which category the image belongs to. In X-ray weld image recognition, there are preset training models for defect types such as cracks, pores, and slag inclusions, so that the resulting model has the ability to recognize the above defects. A three-dimensional, high-degree-of-freedom, full-process virtual X-ray industrial flaw detection environment is established through Unity3d, and convolutional neural network image recognition technology is introduced to identify defects in X-ray films, which helps students complete the learning of X-ray industrial flaw detection and X-ray film defect recognition without the need to build a laboratory or darkroom.

[0060] In an exemplary embodiment, Figure 1As shown, a virtual simulation teaching system for X-ray flaw detection is provided, including: an operation box, an all-in-one computer and an X-ray generator simulation module; the all-in-one computer and the X-ray generator simulation module are both connected to the operation box; the X-ray generator simulation module sends a signal to the operation box to simulate the on state of the X-ray generator; the operation box is used to start the X-ray generator simulation module and set exposure parameters; the exposure parameters include exposure voltage and exposure time; the operation box is also used for training the machine; the operation box is also used to display the operation box status, the X-ray generator on state display and the training state display; an X-ray flaw detection virtual simulation teaching model is deployed on the all-in-one computer; the all-in-one computer is used to interact with student users, and flaw detection teaching is conducted for student users based on the X-ray flaw detection virtual simulation teaching model. The virtual simulation teaching system for X-ray flaw detection also includes a bracket; the bracket is used to support the X-ray generator.

[0061] The X-ray flaw detection virtual simulation teaching system provided in this embodiment allows student users to learn relevant knowledge of the entire process of X-ray flaw detection without radiation risks, without relying on laboratory and darkroom environments, greatly reducing the cost of learning and practice; student users can switch between practice and assessment modes, and can immediately conduct self-tests after completing learning, which can greatly improve the learning efficiency of student users; the convolutional neural network algorithm is used to realize the recognition of weld defects in X-ray films, which can greatly save manpower and improve the accuracy of film identification; the configuration of an all-in-one computer helps student users learn the complete process of X-ray flaw detection through a combination of software and hardware, and provides student users with a three-dimensional, highly simulated virtual environment at the operation box end to complete the operations of bagging, cloth and other links, cooperate with the X-ray generator to realize the film washing step, and finally complete the student users' learning and training on film evaluation through the film evaluation system. First, operating in a virtual system not only reduces the cost of environmental configuration but also avoids the potential safety hazards of X-ray radiation. Second, this application can realize the complete industrial process of X-ray flaw detection, and can truly restore the industrial flaw detection process down to each step of the operation. Third, this application has a complete teaching, training and assessment plan and is equipped with a special learning guide, so the student user training process is very smooth.

[0062] Among them, the hardware part includes X-ray generator, operation box (including STM32 development board, signal light, button and safety lock), all-in-one computer (including external mouse and keyboard), bottom bracket (during flaw detection operation, the X-ray generator needs to be placed horizontally on the bracket with its ray emission port facing downward to expose the film) and related cables: a thick cable with nine-hole female and male connectors is used to connect the X-ray generator and the operation box, and a power cord connects the power supply and the control box.

[0063] Three views of X-ray generator Figure 3The X-ray generator is used to radiate X-ray exposure film. The XXG-2505 model X-ray generator is selected, with a tube current of 5mA and a tube voltage of 250kVp. The power cord and data cable are connected to the X-ray operation box.

[0064] The top view of the operation box is as follows Figure 4 As shown in the figure, the X-ray operation box uses the STM32 development board as the main control unit, equipped with multiple control buttons and a digital tube display for setting and adjusting exposure parameters and controlling the exposure process. It is connected to the built-in all-in-one computer to run the virtual simulation program for student users to operate and learn, and the data interaction between the STM32 development board and the all-in-one computer is realized through series connection.

[0065] like Figure 1 and Figure 2 As shown, the buttons, signal lights, digital tubes and other components inside the operation box are connected to the STM32 development board using Dupont wires, the X-ray generator is connected to the external interface of the operation box through a cable with a nine-hole female-female connector, and the all-in-one computer is connected to the development board through a serial port to complete the communication connection of the entire system. The X-ray flaw detection virtual simulation teaching model includes: a use mode selection module, a teaching component and an assessment component; both the teaching component and the assessment component are connected to the use mode selection module; the use mode selection module is used to display the mode selection interface and obtain the mode selection instruction. When the mode selection instruction is a teaching mode instruction, the flaw detection teaching is evaluated for the student user based on the teaching component, and when the mode selection instruction is an assessment mode instruction, the flaw detection assessment is evaluated for the student user based on the assessment component.

[0066] The teaching components include: a process selection module, an experiment overview display module, an experiment preparation teaching module, an experiment implementation teaching module and a teaching experiment report module; the experiment overview display module, the experiment preparation teaching module, the experiment implementation teaching module and the teaching experiment report module are all connected to the process selection module; the process selection module is used to display the teaching mode interface when receiving the teaching mode instruction and obtain the process selection instruction; the process selection instruction includes the experiment overview display instruction, the experiment preparation teaching instruction, the experiment implementation teaching instruction and the teaching experiment report instruction; the experiment overview display module is used to call and display the experiment overview and experiment outline when receiving the experiment overview display instruction, and The process selection module is called after the instruction is completed; the experiment preparation teaching module is used to call and display the experiment scene and the experiment equipment model after receiving the experiment preparation teaching instruction, and call the process selection module after receiving the return instruction; the experiment scene includes: laboratory scene, darkroom scene and exposure scene; the experiment implementation teaching module is used to prompt the student user to perform bagging training, cloth training, film washing training and film evaluation training after receiving the experiment implementation teaching instruction, and call the process selection module after receiving the return instruction; the teaching experiment report module is used to generate and display the student user teaching experiment report when receiving the teaching experiment report instruction, and call the process selection module after receiving the return instruction.

[0067] The experimental implementation teaching module includes: a bagging training unit, a cloth piece training unit, a film developing training unit and a film evaluation training unit; after receiving the experimental implementation teaching instruction, the bagging training unit calls the darkroom scene and guides the student user to select the experimental equipment for bagging operation training, and then calls the cloth piece training unit; the cloth piece training unit is used to call the exposure scene and guide the student user to set the exposure parameters through the buttons on the operation box and perform cloth piece operation training, and then call the film developing training unit; the film developing training unit is used to call the darkroom scene and guide the student user to perform film developing operation training when receiving the film developing training instruction, and then call the film evaluation training unit; the film evaluation training unit is used to randomly select a negative film picture from the negative film picture database as the negative film picture to be selected. Evaluate the film, obtain the student user's evaluation result on the film to be evaluated as the student user's evaluation result, input the film to be evaluated into the evaluation model, obtain the system evaluation result, and correct the student user's evaluation result based on the system evaluation result; the film evaluation training unit is also used to obtain the auditor user's evaluation result on the film to be evaluated as the auditor user's evaluation result when receiving a manual evaluation instruction, and correct the student user's evaluation result based on the auditor user's evaluation result; the manual evaluation instruction is input by the student user when he does not agree with the system evaluation result; the film evaluation training unit is also used to call the process selection module after receiving the return instruction; the evaluation model is obtained by training the convolutional neural network with multiple film images marked with evaluation results.

[0068] The convolutional neural network includes: a ResNet101 residual structure and an FPN feature pyramid network connected in sequence. The ResNet101 residual structure includes a convolutional layer, a maximum pooling downsampling layer, a residual network, and a fully connected layer connected in sequence. Before training the convolutional neural network, the negative images used for training are subjected to average weighted grayscale processing.

[0069] The assessment components include: an objective question assessment module, an experiment preparation assessment module, an experiment implementation assessment module, and an assessment experiment report module, which are connected in sequence; the objective question assessment module is used to display the objective question assessment interface when receiving the assessment mode instruction, select and display objective questions from the objective question database, and enter the experiment preparation assessment module after obtaining the student user's answer results to each objective question; the assessment content of the objective questions is the experiment overview and experiment outline; the experiment preparation assessment module is used to obtain the student user's scene selection results and experimental equipment selection results, and then enter the experiment implementation assessment module; the experiment implementation assessment module is used to obtain the user's bagging operation results, cloth operation results, film washing operation results and film evaluation operation results, and then enter the assessment experiment report module; the assessment experiment report module is used to generate and display the student user's assessment experiment report based on the scene selection results, experimental equipment selection results, bagging operation results, cloth operation results, film washing operation results and film evaluation operation results.

[0070] The experiment implementation assessment module includes: an implementation selection unit, a bagging assessment unit, a cloth piece assessment unit, a film washing assessment unit and a film evaluation assessment unit; the bagging assessment unit, the cloth piece assessment unit, the film washing assessment unit and the film evaluation assessment unit are all connected to the implementation selection unit; the implementation selection unit is used to display the experiment implementation assessment interface and obtain implementation selection instructions; the implementation selection instructions include bagging assessment instructions, cloth piece assessment instructions, film washing assessment instructions and film evaluation assessment instructions; the bagging assessment unit is used to call the scene selection result and the experimental equipment selection result when receiving the bagging assessment instruction, and obtain the bagging operation result, and call the implementation selection unit when receiving the return instruction; the cloth piece assessment unit is used to call the scene selection result, the experimental equipment selection result and the bagging operation result when receiving the cloth piece assessment instruction, and obtain the cloth piece operation result, and call the implementation when receiving the return instruction selection unit; the film processing and assessment unit is used to call the scene selection result, the experimental equipment selection result and the cloth operation result when receiving the film processing and assessment instruction, and obtain the film processing operation result, and call the implementation selection unit when receiving the return instruction; the film evaluation and assessment unit is used to obtain the film evaluation type instruction when receiving the film processing and assessment instruction; the film evaluation type instruction includes the system film evaluation instruction and the manual film evaluation instruction; the film evaluation and assessment unit randomly selects a film image from the film image database as the film to be evaluated when receiving the system film evaluation instruction, and obtains the student user's evaluation result of the film to be evaluated as the film evaluation operation result, and then calls the assessment experiment report module; the film evaluation and assessment unit obtains the film image uploaded by the student user as the film to be evaluated when receiving the manual film evaluation instruction, and obtains the student user's evaluation result of the film to be evaluated as the film evaluation operation result, and then calls the assessment experiment report module.

[0071] In the software part, the X-ray flaw detection process is carried out in the form of experiments. This application consists of two major modules: one is the teaching mode, such as Figure 5 , mainly used for teacher teaching and student user self-practice. In this mode, student users need to learn to place the X-ray generator on the bottom bracket under the guidance of the teacher, use thick cables to connect the operation box and the X-ray generator, and use thin cables to connect the power supply. They can also open the X-ray flaw detection virtual simulation system on the all-in-one machine for learning and operation, and be familiar with the safety lock switch, training button, exposure voltage, exposure time on the operation box, and how to control the exposure through button adjustment and high-voltage switch key; the second is the assessment mode, such as Figure 6, mainly used for students' self-assessment. Students need to operate the all-in-one machine, control box and X-ray generator, complete the flaw detection process independently and finally generate a pdf file of the experimental report to summarize the student user's learning or assessment. The experiment simulates the real flaw detection process, which is roughly divided into three stages: experimental preparation, experimental implementation and generating experimental reports. The experimental overview and experimental outline are fixed options, and students can click to view them at any time. Experimental preparation stage Teaching mode Experimental scene: Provide four scenes of laboratory, darkroom, X-ray operation room and shipyard for students to conduct experiments, and provide text descriptions, operation requirements and operations that need to be paid attention to in each scene; experimental equipment displays pictures or models of instruments and equipment required in the entire flaw detection process, with text descriptions of the functions and models of the equipment. In the experimental preparation stage assessment mode: student users need to take objective test. Experimental implementation stage: student users need to complete steps such as bagging, cloth, washing and film evaluation under system prompts. Control through the mouse and keyboard, select the equipment to be used, and operate in sequence according to the steps of real flaw detection. During the film processing, the hardware system and the software system are used together. The student user needs to adjust the X-ray exposure time and exposure voltage and other parameters at the operating box end, and perform exposure experiments through the X-ray generator. After the exposure is completed, return to the software system for subsequent operations. In the film evaluation stage, the system provides two methods: images are randomly selected from the system's built-in database for student users to identify defects, and the student user's identification results are compared with the system's standard answers to give the student user's flaw detection results; the system provides an upload interface, and student users can upload X-ray defect films by themselves, which will be evaluated by the system's convolutional network module to help student users determine film defects. Generate experimental report stage: The system will record the various operations of student users during the practice process and generate a PDF document with personal marks for student users to summarize and learn.

[0072] This embodiment provides an X-ray flaw detection virtual simulation teaching system, the working principle of which is as follows:

[0073] Step 1: The all-in-one computer receives a usage instruction, which includes: a teaching mode and an assessment mode; when the all-in-one computer receives a teaching mode usage instruction, it proceeds to step 2; when the all-in-one computer receives an assessment mode usage instruction, it proceeds to step 15.

[0074] The user issues instructions to the computer only by clicking the corresponding buttons with the mouse. The keyboard only serves to move the position and perspective of the virtual character in the scene. The buttons on the operation box only serve to input exposure time, exposure voltage, and high-voltage on and off data into the computer. The switching of UI interfaces and the display of pop-up windows are all automatically implemented through code writing in the computer system.

[0075] Step 2: When the all-in-one computer receives the instruction to use the teaching mode, the all-in-one computer switches the UI interface to the teaching mode interface by calling the scene switching method. Operation prompt information appears on the right side of the UI interface, and the display of the experimental overview, experimental outline, experimental preparation, experimental implementation, and experimental report buttons is realized; the system sets a click trigger event method for the above five buttons, that is, the system will continue to monitor whether the button is clicked. When the user clicks a button with the mouse, the system monitors the user's click operation and triggers the click event method bound to the button. This method will automatically issue the operation instruction corresponding to the button to the all-in-one computer. When the all-in-one computer receives the instruction to use the experimental overview, it goes to step 3; when the all-in-one computer receives the instruction to use the experimental outline, it goes to step 4; when the all-in-one computer receives the instruction to use the experimental preparation, it goes to step 5; when the all-in-one computer receives the instruction to use the experimental implementation, it goes to step 8;

[0076] Step 3: When the all-in-one computer receives the instruction to use the experiment overview, the all-in-one computer will automatically call the method set in the system to change the display property of the pop-up component related to the experiment overview from the default False to True. At this time, the content related to the experiment overview will be displayed in the all-in-one computer system interface, and the student user will be able to see and read it; the system sets a reading completion button in the experiment overview pop-up component. After the user finishes reading, he can click the button with the mouse to send a reading completion instruction to the all-in-one computer. When the all-in-one computer receives the instruction to use the reading completion, the system's internal method will change the display property of the pop-up component related to the experiment overview to False, and the content related to the experiment overview will be hidden.

[0077] Step 4: When the all-in-one computer receives the instruction to use the experiment outline, the all-in-one computer calls the click method bound to the experiment outline button, changes the display property of the pop-up component related to the experiment outline from the default False to True, realizes the display of the content related to the experiment outline and displays the instruction to use after reading below the content; when the all-in-one computer receives the instruction to use after reading, the display property of the pop-up component related to the experiment outline is changed to False, and the content related to the experiment outline is hidden. The above operations are all completed automatically by the computer and are implemented by the code written by the system. The user only needs to complete the operation of clicking the experiment outline button, and the system will automatically execute the above series of operations; the system sets a click event function for each button. The algorithm logic of this function is: the system monitors this button at all times. If the button is clicked by the mouse, this function will be triggered. Different functions are implemented according to the different codes written in the function, such as switching UI interfaces, popping up pop-ups, etc. If the button is not clicked by the mouse, the system will keep listening.

[0078] Step 5: The user uses the mouse to click on the experiment preparation button option. After the clicking action is completed, the all-in-one computer receives the experiment preparation usage instruction, calls the click method bound to the experiment preparation button, and displays sub-level usage instructions including: experiment scene, experiment equipment; the user uses the mouse to click on the experiment scene button option to send instructions to the computer. When the all-in-one computer receives the experiment scene usage instruction, it proceeds to step 6; the user uses the mouse to click on the experiment equipment button option to send instructions to the computer. When the all-in-one computer receives the experiment equipment usage instruction, it proceeds to step 7.

[0079] Step 6: After the all-in-one computer receives the experimental scene usage instruction issued by the user, the system will automatically call the click method bound to the experimental scene button. The function of this method is to display the sub-level usage instructions: laboratory, darkroom and exposure scene.

[0080] The experimental scenes are pre-built inside the all-in-one computer through Unity software, and the experimental scenes include: laboratory, darkroom, exposure scene (the exposure scene includes: X-ray operation room, shipyard operation scene). The technical means adopted for constructing the experimental scenes is to use Unity3D modeling to create different scenes; after the all-in-one computer receives the use instruction of the laboratory / darkroom / X-ray operation room / shipyard operation scene, the system interface background switches to the background picture of the corresponding instruction and displays the introduction text of the corresponding scene.

[0081] Step 7: After the computer receives the instruction to use the experimental equipment, Unity calls the click method bound to the experimental equipment button, and a pop-up display window contains: equipment model, equipment text introduction (name, function); the experimental equipment includes: dark bag, intensifying screen, film developer, densitometer, film viewing light, image quality meter, sample plate, lead door, warning line, X-ray generator, control box, lead mark. The above experimental equipment are all built through Blender and SolidWorks modeling software to create object models and imported into the Unity system for use.

[0082] Step 8: Bagging operation. The specific process is that when the all-in-one computer receives the instruction to implement the experiment, the system triggers the scene switching method and enters the darkroom scene by default; the user needs to use the mouse and keyboard to control the character (keyboard: W-forward, A-move left, S-backward, D-move right, mouse: move to control direction, right-click to trigger the model function, click to execute) approach the film and intensifying screen model, right-click the model, and a usage instruction pops up on the film and intensifying screen model: Load into dark bag. The user uses the mouse to click the instruction to complete the operation; right-click the dark bag model, and a usage instruction pops up: Move to the X-ray machine. After the user clicks, the dark bag model in the darkroom disappears; the bagging operation is completed, and go to step 9.

[0083] Step 9: Cloth piece operation. The specific process is that after the bagging operation is completed, the system switches the scene to the X-ray operation room by default, and a scene switching button appears on the top of the UI interface. After receiving the scene switching command, the computer can switch the scene to the real scene of the shipyard operation; the equipment used (dark bag, sample plate, lead mark, image quality meter, lead door (operation room scene) / warning line (shipyard operation scene), control box) appear on the left and right sides of the page UI in the two scenes. The user uses the mouse to right-click the sample plate, lead mark, and image quality meter respectively, and the use instruction pops up: move to the ray machine, and after clicking them, the three object models appear below the ray emission port of the X-ray generator model; click the lead door / warning line model, and the use instruction pops up: close the lead door / pull up the warning line. The user clicks the instruction to close the lead door / pull up the warning line; click the operation box model to pop up the operation box simulation panel and the system connects to the operation box entity. The user opens the safety lock of the operation box and increases by time. / Decrease button to set exposure time, set exposure voltage through voltage increase / decrease button, press high voltage start button to start exposure, and at the same time, the system UI interface operation box simulation panel is hidden, and a countdown window appears, which is synchronized with the countdown of the operation box entity; the above operation of the operation box entity will transmit the user's real-time operation information to the computer all-in-one through the STM32 development board in accordance with the RS232 standard serial port communication method, and the computer all-in-one scans the serial port in real time to obtain the corresponding data, while Unity reads the data scanned by the serial port in the background, and synchronously displays the obtained data to the operation box simulation panel, STM32 sends countdown data to the computer all-in-one in real time, and the exposure ends when the countdown data is 0; the user right-clicks the lead door / warning line, and clicks to pop up the use command: open the lead door / cancel the warning line; the user right-clicks the dark bag below the ray port of the X-ray generator, and clicks to pop up the move to darkroom command; the cloth piece is finished, and go to step 10.

[0084] Step 10: Film processing operation. The specific process is that after finishing step 9, the system detects the execution of the move to darkroom command, switches the scene to the darkroom, and the user controls the character to approach the dark bag through the mouse and keyboard. When the character approaches and right-clicks the dark bag, an execution command pops up on the dark bag: put it in the film processor. After the user clicks, the dark bag model on the experimental table disappears; move the character to the film processor, right-click the film processor to pop up the execution command: wash the film, click to start the film processor, after washing is completed, right-click the film processor, and a pop-up execution command: turn off the film processor. After clicking, the computer all-in-one receives the turn off the film processor click event command , call the scene switching method to the laboratory scene; the user moves the character to the film, right-clicks to pop up the execution command: Put in the densitometer, clicks to execute the command, and the film model on the laboratory table disappears; moves the character to the densitometer, right-clicks to pop up the execution command: Densitometer measurement, clicks to execute the command, after the densitometer measurement is completed, a film model that has completed the densitometer measurement appears on the laboratory table; moves the character to the film model, right-clicks to pop up the execution command: Put in the film viewing light, clicks to execute the command; moves the character to the film viewing light, right-clicks the film viewing light to pop up the execution command: View the film, clicks to execute the view film execution command, the film viewing is completed, and goes to step 11.

[0085] Step 11: Enter the film evaluation step. The specific steps are as follows: after the all-in-one computer receives the film evaluation execution instruction, it triggers the film evaluation instruction click event, and a pop-up window includes: system film evaluation, manual film evaluation; the system film evaluation operation process is that the user uploads the film image, and the system's built-in image recognition algorithm independently identifies the image defects. The user can directly view the identification results or identify the film by himself and compare it with the system results; the manual film evaluation is that the system randomly calls the film images containing defect types recorded in the database and identifies them by the user; the all-in-one computer receives the system film evaluation execution instruction and proceeds to step 12; the all-in-one computer receives the manual film evaluation instruction and proceeds to step 13.

[0086] Step 12: After the all-in-one computer receives the system evaluation instruction, a picture upload execution instruction pops up. After the user clicks it, he can upload the film picture from the Windows system's resource manager. After the picture is uploaded, the system calls the image recognition algorithm to automatically identify the picture defects. After the identification is completed, an execution command is provided: view the results and go to manual evaluation; the all-in-one computer receives the view result execution command, and a pop-up window displays the picture and the defect type determined by the system; the all-in-one computer receives the go to manual evaluation execution instruction, and a pop-up window displays the uploaded picture and provides a check box for the defect type. The user will conduct the evaluation and check the defects in the picture. After the identification is completed, click view results to compare with the system evaluation results.

[0087] The convolutional neural network model as a whole adopts the Faster-RCNN network to realize the recognition of weld X-ray film images, adopts the ResNet101 residual structure as the feature extraction network, and adopts the FPN feature pyramid network to integrate multi-scale features; as shown in Table 1, the ResNet101 network includes a 7*7 convolution kernel, a convolution layer conv1 with a stride of 2, a maximum pooling downsampling layer max pool with a stride of 2, and several layers of residual networks with different structures (conv2, conv3, conv4, conv5) and the last fully connected layer.

[0088] Table 1 ResNet101 network structure

[0089]

[0090] FPN feature pyramid network: The basic idea of ​​FPN feature pyramid network is to fuse feature maps of different scales by constructing feature pyramids to improve the model's detection ability for objects of different scales. This network structure achieves the fusion of multi-scale features through bottom-up and top-down paths combined with lateral connections, such as Figure 7 .

[0091] Convolutional Neural Network Training:

[0092] Dataset construction: X-ray professional imaging equipment was used to shoot, and 400 images of pores, slag inclusions, and cracks were selected. The image size was 224*224 format. The dataset was divided into a training set and a validation set at a ratio of 4:1. The images in the training set were first processed by average weighted grayscale using Python code to increase the image contrast and thus improve the recognition accuracy of the training model. The grayscaled images were used as images in the training set and input into the neural network for training.

[0093] Training environment: Use CenterOS7 operating system, python3.7, pytorch1.5, GPU accelerated training.

[0094] Convolutional neural network call: Training is performed under the CenterOS operating system, and the trained weight model is directly copied to the Python environment built in the all-in-one machine. The path information of the user uploaded picture is obtained through C# code in Unity, and the path information is passed to the image input end of the image prediction function of the convolutional neural network, so that the image recognition model can obtain the image data and perform recognition.

[0095] Step 13: After the computer receives the manual film evaluation execution instruction, a window pops up, which includes: the X-ray film image to be evaluated, multiple defect type check boxes, and an evaluation completed button; the user needs to select the defect type contained in the image in the check box according to the given film image, click the check box in front of the defect text with the mouse, and tick the box to complete the defect judgment. After the evaluation is completed, click the evaluation completed button; after the film evaluation is completed, go to step 14.

[0096] Description of multiple defect type identification: The Faster-RCNN network is a two-stage target detection network, that is, the first step is to generate a candidate box Proposal to locate the target in the image, and the second step is to identify it; in the first step of positioning, Faster-RCNN generates a region Proposal for an image through the RPN network, puts these candidate areas into the feature extraction network for calculation, and screens them through SVM classifiers, non-maximum suppression, corrected candidate areas, etc., and finally determines multiple targets to be detected, and then enters the second step to identify the targets to be detected, thereby completing the identification of multiple defect types in the image.

[0097] Step 14: Generate an experimental report. The system uses Python code to record the exposure time, exposure voltage, defect type and other data previously selected by the user, calls the reportlab function library to generate a PDF file, and generates a blank PDF document named report through canvas.Canvas("report.pdf"). The .drawString() method is called to write the user-selected data into it, and finally the save() method is used to save it. The system saves the generated PDF document to the specified directory of the computer. Student users can find their own experimental report PDF documents in the corresponding directory. The teaching mode ends and goes to step 1.

[0098] Step 15: When the all-in-one computer receives the instruction to use the assessment mode, the all-in-one computer switches the UI interface to the assessment mode interface by calling the scene switching method, and an objective question assessment window pops up in the center of the system UI. The upper part of the window is equipped with text for explaining the assessment to the user, and the lower part is a start answering button. After the user clicks the start answering button, the questions and options are displayed. The user needs to complete 10 single-choice, 5 multiple-choice and 5 true-or-false questions in the assessment window. The single-choice and true-or-false questions use single-choice boxes, that is, only one answer can be selected from the options, and the multiple-choice questions use check boxes, that is, multiple answers can be selected. The system is equipped with an objective question bank, which is stored on the computer in the form of an Excel table. When conducting an objective question assessment, 10 single-choice questions, 5 multiple-choice questions and 5 true-or-false questions are randomly selected through python code, including: question description, optional answers, and correct answers; the user needs to answer the questions by clicking the options corresponding to the answers. After completing the questions, click the submit button at the bottom of the submit pop-up window to complete the objective question assessment module and go to step 16.

[0099] Step 16: After the user clicks the Submit button, the system automatically jumps to the experimental scene and experimental equipment selection module. The computer all-in-one machine pops up a pop-up window "Experimental scene and experimental equipment selection" to prompt the user to make a selection. The left side of the pop-up window provides text descriptions, and the right side provides two options of X-ray operation room and shipyard operation scene in the form of pictures. The user clicks the option with the mouse to choose one of the two as the subsequent operation scene. After the mouse is clicked, the option color deepens, indicating that the option has been selected. At the same time, the pictures and names of all possible experimental equipment will be displayed. The user can choose by clicking the mouse to select it. The color of the selected equipment deepens. After the user selects the equipment, click the "OK" button below the pop-up window to complete the experimental scene and experimental equipment selection. Figure 8 , go to step 17.

[0100] Step 17: After the user clicks the "OK" button, the experimental scene and equipment selection module ends, and the system automatically enters the experimental implementation module. The left side of the system UI provides four option buttons from top to bottom: bagging, cloth pieces, film washing, and film evaluation. This module does not provide text information prompts, and all steps are operated by the user; the user clicks the "bagging" button with the mouse to proceed to step 18; clicks the "cloth pieces" button to proceed to step 19; clicks the "film washing" button to proceed to step 20; clicks the "film evaluation" button to proceed to step 21; Note: Flaw detection is a linear process operation, that is, bagging-cloth pieces-film washing-film evaluation steps are generally not disrupted, so the subsequent steps are often based on the operation of the previous step. In the experimental implementation module, only when the previous step completes the corresponding operation, will there be corresponding equipment in the next step. For example: in the "bagging" module, if the operation of "moving the film to the X-ray machine" is not completed, then there will be no film model in the cloth piece operation.

[0101] Step 18: The user clicks the "Bagging" button to enter this module. The scene automatically switches to the darkroom environment when entering this module. The darkroom scene is the same as in the teaching mode. The specific operation is the same as step 8. After the user clicks the "Move to X-ray Machine" command, the "Bagging" module ends. After the module ends, the system does not automatically switch the scene and remains in the suggestive environment. The user needs to manually click other function buttons to switch.

[0102] Step 19: After the user clicks the "Closure" button, the system will switch according to the scene selected by the user in step 16. For example, if the user selects the X-ray studio, the operating scene of this module is the X-ray studio. The scene model and operation are the same as in the teaching mode. Please refer to step 9. After the user clicks the "Move to Darkroom" operation, the "Closure" operation is completed, and the scene remains the same and will not switch automatically. The user needs to proceed to the next step manually.

[0103] Step 20: After the user clicks the "Develop" button, the system automatically switches the scene to the darkroom environment. The developing operation can refer to step 10. After the user clicks the "View" command, the "Develop" module ends. The scene switching in this step is consistent with the user operation teaching mode.

[0104] Step 21: After the user clicks the "Film Evaluation" button, he needs to choose system evaluation or manual film identification according to the requirements of the auditor. The flaw detection training institution adopts this system assessment mode when evaluating the learning outcomes of student users. At the same time, an auditor is assigned as an "invigilator" to manage the assessment process for each student user. The identification of the candidate's identity and the generation of the final experimental report all require the auditor to check and certify. If the system film evaluation is selected, the user only needs to operate on the all-in-one computer. The system will randomly provide negative film images with defects for the candidate to identify by checking the checkbox. If manual film evaluation is selected, the auditor needs to provide image resources and upload them to the system. The user identifies the defects. Unlike the teaching mode, this module will not provide answers in the assessment mode, and the test ends after the candidate selects.

[0105] Note: In steps 17 to 21, the scenes and equipment that appear are all those selected in step 16. If any of them are missed in step 16 and subsequent steps cannot be performed, the assessment will fail directly.

[0106] Step 22: When the user clicks the "Experimental Report" button, the system will read the user's operating data in the assessment mode through Python code, and make it into a PDF document through the reportlab function library. The document is directly stored in the specified file directory of the computer all-in-one machine. The directory can be specified through the system settings. After the assessment is completed, the PDF can be printed and the candidate can sign and authenticate to ensure the validity of the score; the assessment mode is completed, go to step 1.

[0107] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0108] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. An X-ray flaw detection virtual simulation teaching system, characterized in that: include: Operation box, all-in-one computer and X-ray generator simulation module; The all-in-one computer and the X-ray generator simulation module are both connected to the operation box; The X-ray generator simulation module sends a signal to the operation box to simulate the on state of the X-ray generator; The operation box is used to start the X-ray generator simulation module and set exposure parameters; the exposure parameters include exposure voltage and exposure time; the operation box is also used to train the machine; the operation box is also used to display the operation box status, X-ray generator start status display and training status display; An X-ray flaw detection virtual simulation teaching model is deployed on the all-in-one computer; the all-in-one computer is used to interact with student users and conduct flaw detection evaluation teaching for student users based on the X-ray flaw detection virtual simulation teaching model.

2. The X-ray flaw detection virtual simulation teaching system according to claim 1 is characterized in that: The X-ray flaw detection virtual simulation teaching system also includes a bracket; The bracket is used for supporting the X-ray generator.

3. The X-ray flaw detection virtual simulation teaching system according to claim 1 is characterized in that: The X-ray flaw detection virtual simulation teaching model includes: a usage mode selection module, a teaching component and an assessment component; The teaching component and the assessment component are both connected to the usage mode selection module; The usage mode selection module is used to display a mode selection interface, obtain a mode selection instruction, and when the mode selection instruction is a teaching mode instruction, perform flaw assessment teaching on student users based on the teaching component; when the mode selection instruction is an assessment mode instruction, perform flaw assessment on student users based on the assessment component.

4. The X-ray flaw detection virtual simulation teaching system according to claim 3 is characterized in that: The teaching components include: a process selection module, an experiment overview display module, an experiment preparation teaching module, an experiment implementation teaching module and a teaching experiment report module; The experiment overview display module, the experiment preparation teaching module, the experiment implementation teaching module and the teaching experiment report module are all connected to the process selection module; The process selection module is used to display the teaching mode interface and obtain the process selection instruction when receiving the teaching mode instruction; the process selection instruction includes the experiment overview display instruction, the experiment preparation teaching instruction, the experiment implementation teaching instruction and the teaching experiment report instruction; The experiment overview display module is used to call and display the experiment overview and experiment outline when receiving the experiment overview display instruction, and call the process selection module after receiving the reading completion instruction; The experiment preparation teaching module is used to call and display the experiment scene and the experiment equipment model after receiving the experiment preparation teaching instruction, and call the process selection module after receiving the return instruction; the experiment scene includes: laboratory scene, darkroom scene and exposure scene; The experiment implementation teaching module is used to prompt the student user to perform bagging training, cloth piece training, film washing training and film evaluation training after receiving the experiment implementation teaching instruction, and call the process selection module after receiving the return instruction; The teaching experiment report module is used to generate and display the student user teaching experiment report when receiving the teaching experiment report instruction, and call the process selection module after receiving the return instruction.

5. The X-ray flaw detection virtual simulation teaching system according to claim 4 is characterized in that: The experimental implementation teaching module includes: bagging training unit, cloth piece training unit, film washing training unit and film evaluation training unit; After receiving the teaching instruction for implementing the experiment, the bagging training unit calls the darkroom scene and guides the student user to select experimental equipment for bagging operation training, and then calls the cloth piece training unit; The film training unit is used to call the exposure scene and guide the student user to set the exposure parameters and perform film operation training through the buttons on the operation box, and then call the film processing training unit; The film processing training unit is used to call the darkroom scene and guide the student user to perform film processing operation training when receiving the film processing training instruction, and then call the film evaluation training unit; The film evaluation training unit is used to randomly select a film image from a film image database as a film to be evaluated, obtain the student user's evaluation result on the film to be evaluated as the student user's evaluation result, input the film to be evaluated into the evaluation model, obtain the system evaluation result, and correct the student user's evaluation result based on the system evaluation result; the film evaluation training unit is also used to obtain the auditor user's evaluation result on the film to be evaluated as the auditor user's evaluation result when receiving a manual evaluation instruction, and correct the student user's evaluation result based on the auditor user's evaluation result; the manual evaluation instruction is input by the student user when he does not agree with the system evaluation result; the film evaluation training unit is also used to call the process selection module after receiving a return instruction; the evaluation model is obtained by training a convolutional neural network using multiple film images marked with evaluation results.

6. The X-ray flaw detection virtual simulation teaching system according to claim 5 is characterized in that: The convolutional neural network includes: a ResNet101 residual structure and an FPN feature pyramid network connected in sequence.

7. The X-ray flaw detection virtual simulation teaching system according to claim 6 is characterized in that: The ResNet101 residual structure includes a convolutional layer, a maximum pooling downsampling layer, a residual network and a fully connected layer connected in sequence.

8. The X-ray flaw detection virtual simulation teaching system according to claim 5 is characterized in that: Before training the convolutional neural network, the negative images used for training are processed by average weighted grayscale.

9. The X-ray flaw detection virtual simulation teaching system according to claim 3 is characterized in that: The assessment component includes: an objective question assessment module, an experiment preparation assessment module, an experiment implementation assessment module, and an assessment experiment report module connected in sequence; The objective question assessment module is used to display the objective question assessment interface upon receiving the assessment mode instruction, select and display objective questions from the objective question database, obtain the student user's answer result to each objective question and then enter the experiment preparation assessment module; the assessment content of the objective question is the experiment overview and experiment outline; The experiment preparation assessment module is used to obtain the scene selection results and experimental equipment selection results of the student user, and then enter the experiment implementation assessment module; The experiment implementation assessment module is used to obtain the user's bagging operation results, cloth piece operation results, film washing operation results and film evaluation operation results, and then enter the assessment experiment report module; The assessment experiment report module is used to generate and display a student user assessment experiment report based on the scene selection result, the experimental equipment selection result, the bagging operation result, the cloth operation result, the film washing operation result and the film evaluation operation result.

10. The X-ray flaw detection virtual simulation teaching system according to claim 9, characterized in that: The experiment implementation assessment module includes: an implementation selection unit, a bagging assessment unit, a cloth piece assessment unit, a film washing assessment unit and a film evaluation assessment unit; The bagging assessment unit, the cloth piece assessment unit, the film washing assessment unit and the film evaluation assessment unit are all connected to the implementation selection unit; The implementation selection unit is used to display the experiment implementation assessment interface and obtain implementation selection instructions; the implementation selection instructions include bagging assessment instructions, cloth piece assessment instructions, film washing assessment instructions and film evaluation assessment instructions; The bagging assessment unit is used to call the scene selection result and the experimental equipment selection result when receiving the bagging assessment instruction, and obtain the bagging operation result, and call the implementation selection unit when receiving the return instruction; The cloth piece assessment unit is used to call the scene selection result, the experimental equipment selection result and the bagging operation result when receiving the cloth piece assessment instruction, and obtain the cloth piece operation result, and call the implementation selection unit when receiving the return instruction; The film processing and assessment unit is used to call the scene selection result, the experimental equipment selection result and the cloth operation result when receiving the film processing and assessment instruction, and obtain the film processing operation result, and call the implementation selection unit when receiving the return instruction; The film evaluation unit is used to obtain a film evaluation type instruction when receiving a film processing evaluation instruction; the film evaluation type instruction includes a system film evaluation instruction and a manual film evaluation instruction; The film evaluation and assessment unit randomly selects a film image from the film image database as the film to be evaluated upon receiving the system film evaluation instruction, and calls the assessment experiment report module after obtaining the student user's evaluation result on the film to be evaluated as the film evaluation operation result; The film evaluation and assessment unit obtains the film image uploaded by the staff user as the film to be evaluated when receiving the manual film evaluation instruction, and obtains the evaluation result of the student user on the film to be evaluated as the film evaluation operation result, and then calls the assessment experiment report module.

Images