Fabricated building teaching system and method based on AR live-action simulation training

Abstract

The invention discloses a fabricated building teaching system and method based on AR live-action simulation practical training, relates to the technical field of AR teaching systems, and aims to solve the technical problems that the theory and practice of students are disjointed and the requirement of the industry for the practical ability of professionals is difficult to meet due to the fact that an existing fabricated building teaching system lacks a practical training link. Comprising a live-action teaching module, an AR experience module and a data interaction module. The data interaction module is used for constructing a data interaction path between the live-action teaching module and the AR experience module; the live-action teaching module comprises a teaching object model unit used for obtaining a teaching object model; the building module data unit is used for obtaining assembled module data information associated with the teaching object; and the live-action teaching unit is used for realizing practical operation teaching of the teaching object according to the teaching object model and the assembled module data information. The system has the advantages that the AR technology and live-action teaching are fused, multi-module collaboration is realized, and the practical ability of students in fabricated buildings is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of AR teaching systems, and more specifically, to an assembly building teaching system and method based on AR real-scene simulation training. Background Art

[0002] At present, with the rapid development of the construction industry, prefabricated buildings have gradually become an important development direction in the construction field due to their advantages such as high efficiency and environmental protection. This makes the cultivation of professionals related to prefabricated buildings crucial and puts forward higher requirements for teaching quality and teaching methods.

[0003] Currently, the progress of educational technology has brought new ideas to the teaching of prefabricated buildings. Many teaching systems have introduced AR / VR technology and achieved certain results in the teaching of theoretical knowledge, enabling students to more intuitively understand the theoretical concepts of prefabricated buildings through virtual scenarios and improving learning efficiency to a certain extent. However, the existing technology still has significant defects. Most of the existing teaching systems using AR / VR technology only focus on teaching theoretical knowledge and test questions and lack a real-scene simulation training module. This means that students cannot practice in a nearly real scenario and it is difficult for them to truly master the operation skills in the actual construction of prefabricated buildings and the ability to handle unexpected problems. In view of this, we propose an assembly building teaching system and method based on AR real-scene simulation training. Summary of the Invention

[0004] The purpose of the present invention is to provide an assembly building teaching system and method based on AR real-scene simulation training to solve the technical problem that the existing assembly building teaching system lacks a training link, resulting in the disconnection between students' theory and practice and being difficult to meet the industry's demand for the practical ability of professional talents.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: An assembly building teaching system and method based on AR real-scene simulation training, including a real-scene teaching module, an AR experience module, and a data interaction module; The data interaction module is used to build a data interaction path between the real-scene teaching module and the AR experience module; The real-scene teaching module includes: A teaching object model unit for obtaining a teaching object model; A building module data unit for obtaining the prefabricated module data information associated with the teaching object; A real-scene teaching unit for implementing practical teaching of the teaching object according to the teaching object model and the prefabricated module data information; The AR experience module includes: A perception module for capturing gesture information; An image tracking module for obtaining real - scene image information associated with a teaching object; A virtual matching unit for obtaining a virtual teaching object model based on the real - scene image information and the teaching object model; A visual enhancement unit for realizing the enhanced display of the virtual teaching object model and the real - scene image information.

[0006] Preferably, the building module data unit includes: A pre - processing sub - unit for pre - processing the prefabricated module data information associated with the teaching object, specifically including unifying the data format and cleaning the noise data. The original data is , and after mean filtering to remove noise, the filtering formula is , where is the radius of the filtering window, is the data point index; An extraction sub - unit for extracting the pre - processed prefabricated module data information to obtain corresponding prefabricated module characteristic parameters, extracting for building structure and material property data, and the extracted characteristic parameters include size and weight; A fusion sub - unit for fusing all prefabricated module characteristic parameters to obtain fusion characteristic parameters, using a weighted average algorithm to integrate each parameter, and then realizing the construction of the database storage module of the prefabricated building teaching system. The fusion formula is , where is the th characteristic parameter, is the corresponding weight, and .

[0007] Preferably, the perception module includes: A sensing generator with a built - in depth sensor, a camera sensor, and an infrared sensor, which are respectively used to obtain the position data of the user's limb parts, gesture information, and action information; An interaction sub - unit for sending the position data of the user's limb parts, gesture information, and action information to the database storage module; A database storage module for matching corresponding teaching action data and their corresponding prefabricated module characteristic parameters according to the position data of the limb parts, gesture information, and action information, providing a basis for constructing a virtual teaching object model or providing a real - scene operation guide for the AR experience module. Let the position data of the limb parts be , the gesture information be , the action information be , using a fuzzy matching algorithm, defining a fuzzy similarity function , where , , The similarities of the limb part position, gesture, and movement respectively , , are the corresponding weight coefficients, and , the similarity calculation uses the Euclidean distance method , where , is the test data is the reference data; A speech recognition unit, used to recognize speech commands, and uses the deep neural network algorithm DNN for speech recognition

[0008] Preferably, the image tracking module includes: A tracking processing sub-unit, used to collect image information related to the teaching object and input it into the reinforcement learning model. Parameters such as the frequency and resolution of image collection can be dynamically adjusted according to the actual teaching scenario; A reinforcement learning model, used to obtain the corresponding key feature points from the processed image information and generate position information, and uses the scale-invariant feature transform algorithm to extract key feature points. First, construct a difference-of-Gaussians pyramid , where is the Gaussian function is the scale factor, then find the extreme points in the difference pyramid, accurately determine the position and scale of the key points by fitting a three-dimensional quadratic function, and finally calculate the main direction and feature descriptor of the key points; An association sub-unit, used to associate the image information with the key feature points and store them in the database storage module. Let the image information be , and the key feature point be , and use the hash algorithm for associated storage, map the descriptor of the key feature point to the hash table, and the association formula is , where is the hash value

[0009] Preferably, the virtual matching unit includes: A data extraction unit, used to extract the fusion feature parameters obtained by the fusion sub-unit ; A position matching unit, used to obtain the key feature points and their position information obtained by the reinforcement learning model; A matching unit, used to obtain the matching feature parameters of the virtual teaching object model and the real-scene image information according to the position information of the key feature points, and uses the random sample consensus algorithm for matching. Let the position of the key feature point in the virtual model be , and the position in the real-scene image be , and the random sample consensus algorithm randomly selects a set of sample points and calculates the transformation matrix such that , then calculate the error of all points to the transformed positions, and select the transformation matrix with the minimum error as the final result; A data storage unit for storing the virtual teaching object model according to the matching feature parameters.

[0010] Preferably, the visual enhancement unit includes: An enhancement processing subunit for obtaining the virtual teaching object model and performing enhancement processing, including optimizing the lighting and texture of the virtual model, and using a physically based rendering algorithm for lighting optimization. The lighting optimization formula is , where is the outgoing radiance, is the incident radiance, is the bidirectional reflectance distribution function, is the surface normal, and are the incident and outgoing directions respectively; An enhanced display subunit for overlapping the enhanced virtual teaching object model with the real scene image information to obtain enhanced display data; A display subunit for displaying the fused image through AR glasses.

[0011] An assembly building teaching method based on AR real-scene simulation training, including the following steps: S1. Model acquisition: Obtain the assembly teaching object model; S2. Information acquisition: Obtain the assembly module data information associated with the assembly teaching object; S3. Data preprocessing and storage construction: Preprocess the obtained assembly module data information, extract feature parameters and fuse them to implement the construction of the database storage module of the assembly building teaching system; S4. Real scene image information acquisition: Obtain the real scene image information associated with the teaching object, collect the image and perform processing, extract key feature points and position information, and then associate and store them; S5. Virtual teaching object model construction: Extract the assembly module feature parameters associated with the assembly module data information, combine the key feature points and their position information obtained by the reinforcement learning model, calculate the matching feature parameters and store the virtual teaching object model; S6. Virtual model enhancement processing: Perform enhancement processing on the virtual teaching object model; S7. Virtual-real image fusion display: Perform enhanced display on the enhanced virtual teaching object model and the real scene image information to obtain enhanced display data; S8. Enhanced data display: Display the enhanced display data; S9. Practical teaching implementation: Implement practical teaching for the teaching object according to the prefabricated teaching object model and the prefabricated module data information, and augmented display data can be used to assist teaching during the practical process; S10. Model data association: Associate the virtual teaching object model with the assembly module data information to facilitate data interaction and feedback.

[0012] Preferably, the S2 further includes: S21. Data preprocessing: Preprocess the prefabricated module data information associated with the teaching object; S22. Feature parameter extraction: Extract the preprocessed prefabricated module data information to obtain the corresponding prefabricated module feature parameters; S23. Data fusion storage: Fusion all prefabricated module feature parameters to realize the construction of the database storage module of the prefabricated building teaching system.

[0013] Preferably, the S4 further includes: S41. Image acquisition and processing: Collect the image information associated with the teaching object and input it into the reinforcement learning model for image processing; S42. Feature point extraction and positioning: Obtain the corresponding key feature points from the processed image information and generate position information; S42. Image feature association storage: Associate the image information with the key feature points and store them.

[0014] Preferably, the S5 further includes: S51. Module feature parameter extraction: Extract the prefabricated module feature parameters associated with the assembly module data information; S52. Feature point information acquisition: Obtain the key feature points and their position information obtained by the reinforcement learning model; S53. Matching parameter calculation: Obtain the matching feature parameters of the virtual teaching object model and the real scene image information according to the position information of the key feature points; S54. Virtual model storage: Realize the storage of the virtual teaching object model according to the matching feature parameters.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. The prefabricated building teaching system based on AR real-scene simulation training provided by the present invention deeply integrates AR technology with real-scene teaching. Through the collaborative work of the real-scene teaching module, the AR experience module, and the data interaction module, a teaching environment that is safe, controllable, and highly restored to the real scene is constructed for students. In this environment, students can conduct practical training, making up for the deficiencies of relying solely on theoretical teaching in the past, enabling students to better combine theoretical knowledge with practice, improving their practical operation ability, meeting the industry's demand for the practical skills of prefabricated building professionals, and solving the problem that the existing prefabricated building teaching system lacks a training link, resulting in the disconnection between students' theory and practice and the difficulty in meeting the industry's demand for the practical ability of professionals.

[0016] 2. In the building module data unit of the present invention, a preprocessing subunit, an extraction subunit, and a fusion subunit are set. These subunits comprehensively process the data information of the prefabricated modules, making the characteristic parameter information of the prefabricated modules more abundant and accurate. The comprehensive learning mode can help students more deeply understand the characteristics, parameters, and the relationships between the various modules of the prefabricated building, further improving the operation accuracy of students during the training process and their ability to solve complex assembly problems, strengthening students' mastery of the knowledge system of prefabricated buildings, and laying a solid foundation for students to handle various assembly scenarios in actual work.

[0017] 3. By setting a perception module, the present invention can capture the gestures, facial expressions, and body movements of users in real time. Based on the characteristic parameters of the prefabricated modules, the AR experience module can achieve more accurate real-scene teaching. This not only enhances the immersion and interactivity of students during the training process but also enables the system to give timely feedback and guidance according to the real-time actions of students. When students are performing virtual assembly operations, the system can judge whether the operations are correct based on the perceived gesture actions, improving the learning efficiency and further enhancing the practical skill level of students. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a schematic diagram of the system framework of the present invention; Figure 2 is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0019] To facilitate the understanding of the technical solution of the present invention by those skilled in the art, the technical solution of the present invention will be further described below in conjunction with the accompanying drawings of the specification.

[0020] Example 1, as Figure 1 shown, a prefabricated building teaching system based on AR real-scene simulation training includes a real-scene teaching module, an AR experience module, and a data interaction module; The data interaction module is used to construct a data interaction path between the real - scene teaching module and the AR experience module; The real - scene teaching module includes: A teaching object model unit, which is used to obtain a teaching object model; A building module data unit, which is used to obtain the prefabricated module data information related to the teaching object; A real - scene teaching unit, which is used to implement the practical teaching of the teaching object according to the teaching object model and the prefabricated module data information; The AR experience module includes: A perception module, which is used to capture gesture information; An image tracking module, which is used to obtain the real - scene image information related to the teaching object; A virtual matching unit, which is used to obtain a virtual teaching object model according to the real - scene image information and the teaching object model; A visual enhancement unit, which is used to realize the enhanced display of the virtual teaching object model and the real - scene image information.

[0021] In an embodiment of the present invention, the building module data unit includes: A pre - processing sub - unit, which is used to pre - process the prefabricated module data information related to the teaching object, specifically including the unification of data formats and the cleaning of noise data. The original data is , and the noise is removed through mean - filtering processing. The filtering formula is , where is the radius of the filtering window, is the data point index; An extraction sub - unit, which is used to extract the corresponding prefabricated module characteristic parameters from the pre - processed prefabricated module data information. For the building structure and material property data, the extracted characteristic parameters include size and weight. The principal component analysis algorithm is used to calculate the covariance matrix , where is the th sample, is the sample mean, and then the covariance matrix is subjected to eigen - decomposition . The first eigen - vectors corresponding to the largest eigenvalues are selected to form a projection matrix . The sample is projected into the new space , where is the eigen - vector matrix, is a diagonal matrix, as a whole represents the transpose matrix of the matrix ; A fusion subunit, which is used to fuse all the characteristic parameters of the prefabricated modules to obtain the fused characteristic parameters, integrates each parameter by using the weighted average algorithm, and then realizes the construction of the database storage module of the prefabricated building teaching system. The fusion formula is , where is the th characteristic parameter, is the corresponding weight, and .

[0022] In an embodiment of the present invention, the sensing module includes: A sensing generator, which is built-in with a depth sensor, a camera sensor, and an infrared sensor, and is respectively used to obtain the position data of the user's limb parts, gesture information, and action information; An interaction subunit, which is used to send the position data of the user's limb parts, gesture information, and action information to the database storage module; A database storage module, which is used to match the corresponding teaching action data and its corresponding prefabricated module characteristic parameters according to the position data of the limb parts, gesture information, and action information, and provide a basis for constructing a virtual teaching object model or providing a real-scene operation guide for the AR experience module. Let the position data of the limb parts be , the gesture information be , and the action information be . Using the fuzzy matching algorithm, define the fuzzy similarity function , where , , are the similarities of the limb part position, gesture, and action respectively, , , are the corresponding weight coefficients, and . The similarity calculation uses the Euclidean distance method, , where , is the test data, is the reference data; A voice recognition unit, which is used to recognize voice commands, and uses the deep neural network algorithm DNN for voice recognition. The forward propagation formula of DNN is , where is the output of the th layer, is the weight matrix of the th layer, is the bias vector of the th layer, is the activation function.

[0023] In an embodiment of the present invention, the image tracking module includes: A tracking processing subunit, configured to collect image information associated with teaching objects and input it into a reinforcement learning model. Parameters such as the frequency and resolution of image collection can be dynamically adjusted according to the actual teaching scenario; A reinforcement learning model, configured to obtain corresponding key feature points from the processed image information and generate position information. The scale-invariant feature transform algorithm is used to extract key feature points. First, a difference-of-Gaussians pyramid is constructed , where is a Gaussian function, is a scale factor. Then, extreme points are searched for in the difference pyramid, and the position and scale of the key points are accurately determined by fitting a three-dimensional quadratic function. Finally, the main direction and feature descriptor of the key points are calculated; An association subunit, configured to associate the image information with the key feature points and store them in the database storage module. Let the image information be , and the key feature points be . The hash algorithm is used for associated storage, and the descriptors of the key feature points are mapped into a hash table. The association formula is , where is the hash value.

[0024] In an embodiment of the present invention, the virtual matching unit includes: A data extraction unit, configured to extract the fusion feature parameters obtained by the fusion subunit ; A position matching unit, configured to obtain the key feature points and their position information obtained by the reinforcement learning model; A matching unit, configured to obtain the matching feature parameters between the virtual teaching object model and the real-scene image information according to the position information of the key feature points, and perform matching using the random sample consensus algorithm. Let the position of the key feature point in the virtual model be , and the position in the real-scene image be . The random sample consensus algorithm randomly selects a set of sample points and calculates the transformation matrix such that . Then, the error of all points to the transformed position is calculated, and the transformation matrix with the minimum error is selected as the final result; A data storage unit, configured to store the virtual teaching object model according to the matching feature parameters.

[0025] In an embodiment of the present invention, the visual enhancement unit includes: An enhancement processing subunit, configured to obtain the virtual teaching object model and perform enhancement processing, including optimizing the lighting and texture of the virtual model. The physically based rendering algorithm is used for lighting optimization. The lighting optimization formula is , where is the outgoing radiance, is the incident radiation rate, is the bidirectional reflectance distribution function, is the surface normal, and are the incident and outgoing directions respectively; An enhanced display subunit, which is used to overlap the enhanced virtual teaching object model with the real-scene image information to obtain enhanced display data. The Laplacian pyramid image fusion algorithm is adopted to decompose the virtual model and the real-scene image into Laplacian pyramids respectively, then fuse them at different scales, and finally reconstruct the fused image; A display subunit, which is used to display the fused image.

[0026] Embodiment 2, as Figure 2 shown, a prefabricated building teaching method based on AR real-scene simulation training includes the following steps: S1. Model acquisition: Obtain a prefabricated teaching object model; S2. Information acquisition: Obtain the prefabricated module data information associated with the prefabricated teaching object; S3. Data preprocessing and storage construction: Preprocess the obtained prefabricated module data information, extract feature parameters and fuse them to realize the construction of the database storage module of the prefabricated building teaching system; S4. Real-scene image information acquisition: Obtain the real-scene image information associated with the teaching object, collect and process the image, extract key feature points and position information, and then associate and store them; S5. Virtual teaching object model construction: Extract the prefabricated module feature parameters associated with the assembly module data information, combine the key feature points and their position information obtained by the reinforcement learning model, calculate the matching feature parameters and store the virtual teaching object model; S6. Virtual model enhancement processing: Perform enhancement processing on the virtual teaching object model; S7. Virtual-real image fusion display: Perform enhanced display on the enhanced virtual teaching object model and the real-scene image information to obtain enhanced display data; S8. Enhanced data display: Display the enhanced display data; S9. Practical teaching implementation: Implement the practical teaching of the teaching object according to the prefabricated teaching object model and the prefabricated module data information, and the enhanced display data can be used to assist teaching during the practical process; S10. Model data association: Associate the virtual teaching object model with the assembly module data information to facilitate data interaction and feedback.

[0027] In the embodiment of the present invention, the S2 further includes: S21. Data preprocessing: Preprocess the prefabricated module data information associated with the teaching object; S22. Feature parameter extraction: Extract the data information of the pre-processed prefabricated module to obtain the corresponding feature parameters of the prefabricated module; S23. Data fusion and storage: Integrate all the feature parameters of the prefabricated modules to construct the database storage module of the prefabricated building teaching system.

[0028] In the embodiment of the present invention, the S4 further includes: S41. Image acquisition and processing: Collect the image information related to the teaching object and input it into the reinforcement learning model for image processing; S42. Feature point extraction and positioning: Obtain the corresponding key feature points from the processed image information and generate the position information; S42. Image feature association storage: Associate the image information with the key feature points and store them.

[0029] In the embodiment of the present invention, the S5 further includes: S51. Module feature parameter extraction: Extract the feature parameters of the prefabricated module associated with the data information of the assembly module; S52. Feature point information acquisition: Obtain the key feature points and their position information obtained by the reinforcement learning model; S53. Matching parameter calculation: Obtain the matching feature parameters of the virtual teaching object model and the real scene image information according to the position information of the key feature points; S54. Virtual model storage: Realize the storage of the virtual teaching object model according to the matching feature parameters.

[0030] The embodiments disclosed in the present invention are preferred embodiments, but not limited thereto. Those of ordinary skill in the art can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. However, as long as they do not depart from the spirit of the present invention, they are within the protection scope of the present invention.

Claims

1. An AR-based prefabricated building teaching system, characterized in that: Including real-life teaching module, AR experience module and data interaction module; The data interaction module is used to build a data interaction channel between the real-life teaching module and the AR experience module; The real-life teaching module includes: A teaching object model unit, used for obtaining a teaching object model; Building module data unit, used to obtain prefabricated module data information associated with the teaching object; A real-life teaching unit, used to implement practical teaching of teaching objects based on teaching object models and assembled module data information; The AR experience module includes: Perception module, used to capture gesture information; An image tracking module is used to obtain real-scene image information associated with the teaching object; A virtual matching unit, used for obtaining a virtual teaching object model according to the real scene image information and the teaching object model; The visual enhancement unit is used to realize the enhanced display of virtual teaching object models and real-scene image information.

2. The prefabricated building teaching system based on AR real-scene simulation training according to claim 1 is characterized in that: The building module data unit includes: The preprocessing subunit is used to preprocess the prefabricated module data information associated with the teaching object, including the unification of data format and the cleaning of noise data. The original data is , after mean filtering to remove noise, the filtering formula is: ,in, is the filter window radius, Index of data points; The extraction subunit is used to extract the preprocessed prefabricated module data information to obtain the corresponding prefabricated module characteristic parameters, and extract the building structure and material attribute data. The extracted characteristic parameters include size and weight; The fusion subunit is used to fuse all the characteristic parameters of the prefabricated modules to obtain the fusion characteristic parameters, and use the weighted average algorithm to integrate the parameters, thereby realizing the construction of the database storage module of the prefabricated building teaching system. The fusion formula is: ,in, For the characteristic parameters, is the corresponding weight, and .

3. The prefabricated building teaching system based on AR real-scene simulation training according to claim 2 is characterized in that: The perception module comprises: A sensor generator, with a built-in depth sensor, camera sensor and infrared sensor, which are used to obtain the user's body position data, gesture information and movement information respectively; An interaction subunit, used to send the user's body part position data, gesture information, and action information to a database storage module; The database storage module is used to match the corresponding teaching action data and its corresponding assembly module characteristic parameters according to the limb position data, gesture information, and action information, so as to provide a basis for the AR experience module to build a virtual teaching object model or provide a realistic operation guidance. , the gesture information is , the action information is , using fuzzy matching algorithm, defining fuzzy similarity function ,in, , , are the similarities of body part positions, gestures, and actions, respectively. , , is the corresponding weight coefficient, and , the similarity is calculated using the Euclidean distance method, ,in, , For test data, For reference data; The speech recognition unit is used to recognize voice commands and uses a deep neural network algorithm DNN for speech recognition.

4. The prefabricated building teaching system based on AR real-scene simulation training according to claim 3 is characterized in that: The image tracking module comprises: The tracking processing subunit is used to collect image information related to the teaching object and input it into the reinforcement learning model. The frequency, resolution and other parameters of image acquisition can be dynamically adjusted according to the actual teaching scenario; The reinforcement learning model is used to obtain the corresponding key feature points from the processed image information and generate position information. The scale-invariant feature transformation algorithm is used to extract the key feature points. First, a Gaussian difference pyramid is constructed. ,in, is a Gaussian function, is the scale factor, and then finds the extreme points in the difference pyramid, accurately determines the position and scale of the key points by fitting a three-dimensional quadratic function, and finally calculates the main direction and feature descriptor of the key points; The association subunit is used to associate the image information with the key feature points and store them in the database storage module. Suppose the image information is The key feature points are , the hash algorithm is used for associative storage, and the descriptors of key feature points are mapped into the hash table. The association formula is ,in, Is the hash value.

5. The prefabricated building teaching system based on AR real-scene simulation training according to claim 4 is characterized in that: The virtual matching unit comprises: Data extraction unit, used to extract the fusion feature parameters obtained by the fusion sub-unit ; A position matching unit is used to obtain key feature points and their position information obtained by the reinforcement learning model; The matching unit is used to obtain the matching feature parameters of the virtual teaching object model and the real scene image information according to the position information of the key feature points, and use the random sampling consensus algorithm to match. Suppose the position of the key feature point in the virtual model is , and its position in the real scene image is The random sampling consensus algorithm randomly selects a set of sample points and calculates the transformation matrix , so that , then calculate the error from all points to the transformed position, and select the transformation matrix with the smallest error as the final result; The data storage unit is used to store the virtual teaching object model according to the matching feature parameters.

6. The prefabricated building teaching system based on AR real-scene simulation training according to claim 5 is characterized in that: The visual enhancement unit comprises: The enhancement processing subunit is used to obtain the virtual teaching object model and perform enhancement processing, including the lighting and texture optimization of the virtual model. The lighting optimization is performed using a physically based rendering algorithm. The lighting optimization formula is: ,in, is the outgoing radiance, is the incident radiance, is the bidirectional reflectance distribution function, is the surface normal, and are the incident and outgoing directions respectively; The enhanced display subunit is used to overlap the virtual teaching object model after the enhanced processing with the real scene image information to obtain enhanced display data; The display subunit is used to display the fused image.

7. A method applied to the prefabricated building teaching system based on AR real-scene simulation training as claimed in claim 6, characterized in that: The following steps are involved: S1. Model acquisition: obtaining the assembled teaching object model; S2, information acquisition: obtaining the prefabricated module data information associated with the prefabricated teaching object; S3, data preprocessing and storage construction: preprocess the acquired prefabricated module data information, extract and merge feature parameters, and realize the construction of the database storage module of the prefabricated building teaching system; S4, real scene image information acquisition: obtain the real scene image information associated with the teaching object, collect and process the image, extract key feature points and position information, and then associate and store them; S5, virtual teaching object model construction: extracting the assembly module feature parameters associated with the assembly module data information, combining the key feature points and their position information obtained by the reinforcement learning model, calculating the matching feature parameters and storing the virtual teaching object model; S6, virtual model enhancement processing: enhancing the virtual teaching object model; S7, virtual and real image fusion display: the virtual teaching object model after enhancement processing and the real scene image information are enhanced and displayed to obtain enhanced display data; S8, displaying enhanced display data: displaying enhanced display data; S9. Practical teaching implementation: Practical teaching of teaching objects is realized according to the assembled teaching object model and assembled module data information, and enhanced display data can be used to assist teaching during the practical operation; S10. Model data association: associate the virtual teaching object model with the assembly module data information to facilitate data interaction and feedback.

8. The method for teaching prefabricated buildings based on AR real-scene simulation training according to claim 7 is characterized in that: The S2 further includes: S21, data preprocessing: preprocessing the assembled module data information associated with the teaching object; S22, feature parameter extraction: extracting the preprocessed prefabricated module data information to obtain corresponding prefabricated module feature parameters; S23, data fusion storage: integrate all the characteristic parameters of the prefabricated modules to realize the construction of the database storage module of the prefabricated building teaching system.

9. The method for teaching prefabricated buildings based on AR real-scene simulation training according to claim 8 is characterized in that: The S4 further comprises: S41, image acquisition and processing: collecting image information associated with the teaching object and inputting it into the reinforcement learning model for image processing; S42, feature point extraction and positioning: obtaining corresponding key feature points from the processed image information and generating position information; S42, image feature association storage: associating the image information with key feature points and storing them.

10. The method for teaching prefabricated buildings based on AR real-scene simulation training according to claim 9 is characterized in that: The S5 further includes: S51, module characteristic parameter extraction: extracting the assembly module characteristic parameters associated with the assembly module data information; S52, feature point information acquisition: acquiring key feature points and their position information obtained by the reinforcement learning model; S53, matching parameter calculation: obtaining matching feature parameters between the virtual teaching object model and the real scene image information according to the position information of the key feature points; S54, virtual model storage: storing the virtual teaching object model according to matching feature parameters.

Images