Classroom teaching intelligent auxiliary system based on large language model

Abstract

The invention discloses a classroom teaching intelligent auxiliary system based on a large language model, and the system obtains student thinking concentration degree demonstration parameters based on collected video data, obtains classroom activity demonstration parameters based on collected audio data, obtains classroom adaptability demonstration parameters based on collected environment data, and carries out the classroom teaching. According to the method, deep understanding and real-time feedback of the classroom teaching process are realized through an innovative data acquisition and processing technology, the accuracy and comprehensiveness of teaching auxiliary feedback are improved, and the teaching auxiliary feedback efficiency is improved. A feedback score of a classroom is quantified according to data calculation, and the problems that an existing classroom teaching auxiliary system is lack of deep understanding and real-time feedback capability on teaching content on one hand, has limitations on data acquisition and processing on the other hand, cannot comprehensively capture classroom interaction and student learning states, and improves teaching efficiency are solved. The analysis result is not accurate and comprehensive enough.

Description

Technical Field

[0001] The present invention relates to the technical field of teaching aids, and particularly to an intelligent teaching assistant system for classroom teaching based on large language models. Background Art

[0002] Existing classroom teaching assistant systems mostly rely on traditional video monitoring and simple data analysis, lacking in-depth understanding of the teaching process and real-time feedback capabilities. In addition, these systems have limitations in data collection and processing, unable to comprehensively capture classroom interactions and students' learning status, resulting in inaccurate and incomplete analysis results. Summary of the Invention

[0003] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract, and the title. However, such simplifications or omissions shall not be used to limit the scope of the present invention.

[0004] In view of the problems existing in the above-mentioned existing classroom teaching assistant systems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is to address the issues that existing classroom teaching assistant systems lack in-depth understanding of the teaching process and real-time feedback capabilities on the one hand, and have limitations in data collection and processing on the other hand, unable to comprehensively capture classroom interactions and students' learning status, resulting in inaccurate and incomplete analysis results.

[0006] To solve the above technical problem, the present invention provides the following technical solution: An intelligent teaching assistant system for classroom teaching based on large language models, including the following components: A data collection module that collects classroom videos, audio, and environmental data through a configured plurality of high-definition cameras, microphone arrays, and sensor networks; A data processing module that is wirelessly data-connected to the data collection module, and performs real-time processing and analysis after obtaining the collected data; A feedback module that is data-connected to the data processing module, generates real-time feedback based on the analysis results, and transmits it to the teacher's terminal for display.

[0007] As a preferred solution of the intelligent teaching assistant system for classroom teaching based on large language models of the present invention, wherein: A data preprocessing unit is also embedded in the data collection module, and data preprocessing is sequentially performed after obtaining the collected data; Among them, the data preprocessing steps specifically include: data cleaning, data augmentation, and data standardization.

[0008] As a preferred solution of the classroom teaching intelligent assistance system based on large language models according to the present invention, wherein: when the data acquisition module acquires various types of data, it specifically includes: Video data acquisition: Real-time capture of classroom images through 3 high-definition cameras evenly distributed in the front of the classroom to ensure full frontal coverage of the classroom space; Audio data acquisition: Real-time acquisition of classroom voices using a set of microphone arrays evenly arranged in the front of the classroom, and extraction of sound pressure fluctuations by combining sound pressure recognition technology; Environmental data acquisition: Acquisition of classroom temperature, humidity, and lighting environmental data through a sensor network to assist in analyzing students' learning status.

[0009] As a preferred solution of the classroom teaching intelligent assistance system based on large language models according to the present invention, wherein: after the data processing module obtains the acquired data, the real-time processing and analysis specifically include the following steps: S1: Obtaining a reference parameter for students' thinking concentration based on the acquired video data; S2: Generating a sound pressure fluctuation curve based on the acquired audio data, and obtaining a reference parameter for classroom activity based on the sound pressure fluctuation curve; S3: Obtaining a reference parameter for classroom adaptability based on the acquired environmental data; S4: Establishing an analysis management model, inputting the reference parameter for students' thinking concentration, the reference parameter for classroom activity, and the reference parameter for classroom adaptability, and outputting an analysis management parameter value; S5: Conducting a feedback score based on the analysis management parameter value and transmitting the analysis result to the feedback module in real time.

[0010] As a preferred solution of the classroom teaching intelligent assistance system based on large language models according to the present invention, wherein: obtaining the reference parameter for students' thinking concentration based on the acquired video data specifically includes the following steps: S1: A set of high-definition cameras real-time capture classroom images; S2: During the acquisition time, reference points are taken at uniform time intervals, and the proportion of students' faces facing the blackboard at the current moment is obtained for the corresponding reference points; S3: Establishing a model for obtaining the reference parameter for students' thinking concentration, and sequentially inputting the face proportions at different reference point times in chronological order, and outputting the reference parameter for students' thinking concentration; Among them, the model for obtaining the reference parameter for students' thinking concentration is specifically: ;

[0011] Among them, δ is the reference parameter for students' thinking concentration; α1 is the face proportion at the first reference point time, %; α n is the nth, that is, the face proportion at the last reference point time, %; n is the number of reference points; α is the face proportion at the corresponding reference point time, %; -1.29 and 0.67 are adjustment constants.

[0012] As a preferred solution of the intelligent classroom teaching assistance system based on the large language model of the present invention, wherein: obtaining the classroom activity reference parameter based on the sound pressure fluctuation curve specifically includes the following steps: Q1: generating a sound pressure fluctuation curve based on the collected audio data; Q2: selecting the reference point moment in S2; Q3: obtaining the sound pressure value at the corresponding moment, establishing a classroom activity reference parameter model, and sequentially inputting the sound pressure values at different reference point moments in chronological order to output the classroom activity reference parameter; Wherein, the classroom activity reference parameter model is specifically: ;

[0013] Wherein, ε is the classroom activity reference parameter; D1 is the sound pressure value at the first reference point moment, dB; D n is the nth, that is, the sound pressure value at the last reference point moment, dB; n is the number of reference points; D is the sound pressure value at the corresponding reference point moment, dB; -0.87 and 1.33 are adjustment constants.

[0014] As a preferred solution of the intelligent classroom teaching assistance system based on the large language model of the present invention, wherein: obtaining the classroom adaptability reference parameter based on the collected environmental data specifically includes the following steps: H1: real-time collecting classroom temperature, humidity, and light environment data through a sensor network; H2: obtaining the average temperature, average humidity, and average light data during the collection period; H3: establishing a classroom adaptability reference parameter model, inputting the environmental data in H2, and outputting the classroom adaptability reference parameter; Wherein, the classroom adaptability reference parameter is specifically: ;

[0015] Wherein, η is the classroom adaptability reference parameter; n is the number of reference points; A is the average temperature, °C; B is the average humidity, %; C is the average light, lux; -1, 1.09, 1.45, and 3.70 are adjustment constants; dx is an integral operation.

[0016] As a preferred solution of the intelligent classroom teaching assistance system based on the large language model of the present invention, wherein: the established analysis and management model is specifically: ;

[0017] Wherein, ω is the analysis and management parameter value; δ is the student thinking concentration reference parameter; ε is the classroom activity reference parameter; η is the classroom adaptability reference parameter; ln2.09, 1, and 1.03 are adjustment constants.

[0018] As a preferred solution of the intelligent classroom teaching assistance system based on the large language model of the present invention, specifically: the feedback scoring based on the analysis management parameter value is specifically as follows: when the analysis management parameter value is higher than 2.7 or 2.72, it is feedback that the classroom teaching demonstration reaches the standard, and the analysis management parameter value and the scoring result are synchronously sent to the feedback module.

[0019] Advantages of the present invention: The present invention provides an intelligent classroom teaching assistance system based on the large language model. The student thinking concentration reference parameter is obtained based on the collected video data, the classroom activity reference parameter is obtained based on the collected audio data, and the classroom adaptability reference parameter is obtained based on the collected environmental data. Then, the analysis management parameter value is output according to the fitted analysis management model, and feedback is carried out according to the analysis management value. Through innovative data collection and processing technologies, the present invention realizes in-depth understanding and real-time feedback of the classroom teaching process, improves the accuracy and comprehensiveness of teaching assistance feedback, quantifies the classroom feedback scoring based on data calculations, and solves the problems that the existing classroom teaching assistance system lacks the ability of in-depth understanding and real-time feedback of teaching content on the one hand, and has limitations in data collection and processing on the other hand, and cannot comprehensively capture classroom interactions and students' learning states, resulting in inaccurate and incomplete analysis results. Description of the Drawings

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. Among them: Figure 1 It is the system module diagram of the intelligent classroom teaching assistance system based on the large language model provided by the present invention.

[0021] Figure 2 It is the method flowchart for the data processing module to perform real-time processing and analysis after obtaining the collected data provided by the present invention.

[0022] Figure 3 It is the method flowchart for obtaining the student thinking concentration reference parameter provided by the present invention.

[0023] Figure 4

[0024] Figure 5 It is the method flowchart for obtaining the classroom adaptability reference parameter provided by the present invention. Detailed Embodiments

[0025] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0026] Existing classroom teaching assistance systems mostly rely on traditional video surveillance and simple data analysis, lacking the ability to deeply understand teaching content and provide real-time feedback. In addition, these systems have limitations in data collection and processing, unable to comprehensively capture classroom interactions and students' learning states, resulting in inaccurate and incomplete analysis results.

[0027] Therefore, referring to Figure 1 , the present invention provides a classroom teaching intelligent assistance system based on a large language model, including the following components: A data collection module 100 that collects classroom video, audio, and environmental data through a configured plurality of high-definition cameras, microphone arrays, and sensor networks; A data processing module 200 that is wirelessly connected to the data collection module 100 and performs real-time processing and analysis after obtaining the collected data; A feedback module 300 that is data-connected to the data processing module 200, generates real-time feedback based on the analysis results, and transmits it to the teacher terminal for display.

[0028] It should be noted that the video collection unit, audio collection unit, and sensor network in the present invention are all applications of existing conventional electronic components, and there is no difference in the hardware itself, so no redundant description will be made here.

[0029] Furthermore, a data preprocessing unit is also embedded in the data collection module 100 to perform data preprocessing on the collected data in sequence; Among them, the data preprocessing steps specifically include: data cleaning, data augmentation, and data normalization.

[0030] It should be noted that: 1. Data cleaning Remove noise: Filter the noise from the collected original data (such as video, audio, sensor data). For example, remove background noise in audio (such as the sound of fans, the movement of desks and chairs) or blurred frames in video.

[0031] Outlier processing: Detect and process abnormal data. For example, abnormal temperature or humidity values may appear in sensor data, which need to be corrected or removed through threshold detection or statistical methods.

[0032] Data Alignment: Perform time synchronization on multimodal data. For example, align video frames, audio segments, and sensor data according to timestamps to ensure data consistency.

[0033] 2. Data Augmentation Video Data Augmentation: Perform operations such as rotation, scaling, and cropping on video data to increase data diversity and improve the generalization ability of the model.

[0034] Audio Data Augmentation: Perform operations such as speed change, pitch change, and adding background noise to audio data to simulate different classroom environments.

[0035] Environmental Data Augmentation: Simulate different environmental conditions (such as high temperature, low temperature, high humidity) through interpolation or generative models to enhance the robustness of the model.

[0036] 6. Data Standardization Standardization: Convert the feature values of different data sources into a unified scale. For example, standardize video features (such as pixel values) and audio features (such as frequency values) to the same numerical range.

[0037] Furthermore, when the data acquisition module 100 performs various types of data acquisition, it specifically includes: Video Data Acquisition: Real-time capture of classroom images through 3 high-definition cameras evenly distributed in the front of the classroom to ensure full and forward coverage of the classroom space; Audio Data Acquisition: Use a set of microphone arrays evenly arranged in the front of the classroom to real-time collect classroom voices, and combine sound pressure recognition technology to extract sound pressure fluctuations; Environmental Data Acquisition: Collect classroom temperature, humidity, and light environment data through a sensor network to assist in analyzing students' learning status.

[0038] It should be noted that: The high-definition cameras real-time capture classroom images and obtain the facial orientations of students at different times. Specifically, obtaining the facial orientations can be achieved through conventional facial recognition technology, and there is no need to elaborate here.

[0039] Sound pressure recognition technology is also the application of existing conventional means, and there is no need to elaborate here.

[0040] Furthermore, refer to Figure 2 , after the data processing module 200 obtains the collected data, the real-time processing and analysis specifically include the following steps: S1: Obtain the reference parameter of students' thinking concentration based on the collected video data; S2: Generate a sound pressure fluctuation curve based on the collected audio data, and obtain the reference parameter of classroom activity based on the sound pressure fluctuation curve; S3: Obtain the reference parameter of classroom adaptability based on the collected environmental data; S4: Establish an analysis management model, input the reference parameters of students' thinking concentration, classroom activity, and classroom adaptability, and output the analysis management parameter values; S5: Conduct feedback scoring based on the analysis management parameter values and transmit the analysis results to the feedback module 300 in real time.

[0041] Furthermore, refer to Figure 3 , obtaining the reference parameters of students' thinking concentration based on the collected video data specifically includes the following steps: S1: A group of high-definition cameras capture classroom images in real time; S2: During the collection time, take reference points at uniform time intervals, and obtain the proportion of students' faces facing the blackboard at the current moment for the corresponding reference points; S3: Establish a model for obtaining the reference parameters of students' thinking concentration, and input the face proportions at different reference point times in chronological order to output the reference parameters of students' thinking concentration; Among them, the model for obtaining the reference parameters of students' thinking concentration is specifically: ;

[0042] Among them, δ is the reference parameter of students' thinking concentration; α1 is the face proportion at the first reference point time, %; α n is the nth, that is, the face proportion at the last reference point time, %; n is the number of reference points; α is the face proportion at the corresponding reference point time, %; -1.29 and 0.67 are adjustment constants.

[0043] It should be noted that: when specifically generating the above model, it is first necessary to consider that the thinking concentration of students at a specific moment is mainly generalized and statistically analyzed through the face orientation. When the face orientation of students is concentrated on the blackboard, it is generally defined as concentrated thinking. Therefore, this model incorporates the face proportion as a basic parameter for discussion. The first item of the model discusses the differences between the basic parameters at different times (only by incorporating the basic parameters at different times as a whole can the overall trend during the overall monitoring time be maximally reflected), taking one unit as the measurement standard, and generally obtaining the influence of the overall face orientation trend; the second item of the model discusses the average face orientation influence at each moment; the two are combined, and correction parameters are given through a simulator to increase the robustness and optimize the calculation fluency of the model, which can reflect the reference parameter values during the overall monitoring period as much as possible while saving computing power.

[0044] Furthermore, refer to Figure 4 , obtaining the reference parameters of classroom activity based on the sound pressure fluctuation curve specifically includes the following steps: Q1: Generate a sound pressure fluctuation curve based on the collected audio data; Q2: Select the reference point time in S2; Q3: Obtain the sound pressure value at the corresponding moment, establish a reference model for classroom activity, and input the sound pressure values at different reference point moments in chronological order to output the reference for classroom activity; Among them, the reference model for classroom activity is specifically: ;

[0045] Among them, ε is the reference for classroom activity; D1 is the sound pressure value at the first reference point moment, dB; D n is the nth, that is, the sound pressure value at the last reference point moment, dB; n is the number of reference points; D is the sound pressure value at the corresponding reference point moment, dB; -0.87 and 1.33 are adjustment constants.

[0046] It should be noted that: For the discussion of the specific process of generating the above model, please refer to the process of obtaining the reference for students' thinking concentration for the same reason.

[0047] Furthermore, referring to Figure 5 , obtaining the reference for classroom adaptability based on the collected environmental data specifically includes the following steps: H1: Real-time collect classroom temperature, humidity, and light environment data through a sensor network; H2: Obtain the average temperature, average humidity, and average light data during the collection period; H3: Establish a reference model for classroom adaptability, input the environmental data in H2, and output the reference for classroom adaptability; Among them, the reference for classroom adaptability is specifically: ;

[0048] Among them, η is the reference for classroom adaptability; n is the number of reference points; A is the average temperature, °C; B is the average humidity, %; C is the average light, lux; -1, 1.09, 1.45, and 3.70 are adjustment constants; dx is an integral operation.

[0049] It should be noted that: The upper limit of the integral in the above model is n -1 , when obtaining the average temperature, average humidity, and average light, when maximizing the influence degree during the monitoring period. Therefore, this model uses the basic expression of integration. When obtaining the upper limit of the integral, the first thing determined is to select the number of reference points itself (the number of reference points that appear during the monitoring period reflects the number of different environmental statistical points), and then the adjustment of -1 is considered. The actual change trend is already small towards the end of the statistical points, and the maximum influence value has been reflected in the early stage. Therefore, it is most appropriate to use the function expression of -1 for generalization.

[0050] Furthermore, the established analysis and management model is specifically: ;

[0051] Among them, ω is the value of the analysis management parameter; δ is the parameter indicating the concentration of students' thinking; ε is the parameter indicating the activity of the classroom; η is the parameter indicating the adaptability of the classroom; ln2.09, 1, and 1.03 are adjustment constants.

[0052] Specifically, the feedback scoring based on the value of the analysis management parameter is as follows: when the value of the analysis management parameter is higher than 2.7 or 2.72, it is considered that the demonstration of classroom teaching meets the standard, and the value of the analysis management parameter and the scoring result are synchronously sent to the feedback module 300.

[0053] Experimental verification process In order to verify the technical effect of the intelligent auxiliary system for classroom teaching based on the large language model, the following experimental verification process is designed and implemented. The experiment simulates a real classroom environment, collects video, audio, and environmental data, calculates the parameters indicating the concentration of students' thinking, the activity of the classroom, and the adaptability of the classroom, and outputs the value of the analysis management parameter, and finally evaluates the feedback accuracy and comprehensiveness of the system.

[0054] Experimental design 1. Experimental environment: Simulated classroom: Standard classroom layout, equipped with 3 high-definition cameras, 1 microphone array, and a sensor network (temperature, humidity, and light sensors).

[0055] Experimental subjects: 30 students and 1 teacher.

[0056] Experimental duration: 45 minutes (standard classroom duration).

[0057] 2. Data collection: Video data: Real-time capture of students' facial expressions, postures, and behaviors through 3 high-definition cameras.

[0058] Audio data: Collect classroom voices through the microphone array and extract sound pressure fluctuations in combination with sound pressure recognition technology.

[0059] Environmental data: Real-time collection of classroom temperature, humidity, and light data through the sensor network.

[0060] 3. Data processing: Parameter indicating the concentration of students' thinking (δ): Analyze the degree of students' attention concentration through video data.

[0061] Parameter indicating the activity of the classroom (ε): Analyze the frequency of classroom interaction and sound pressure fluctuations through audio data.

[0062] Parameter indicating the adaptability of the classroom (η): Analyze the degree of students' adaptation to the classroom environment through environmental data.

[0063] Analysis management model: 4. Obtain the analysis management parameter value.

[0064] 5. Feedback the score.

[0065] 6. Test data The following are the specific data collected during the test and the calculation results: 7. Data analysis Student thinking concentration parameter (δ): In the 15 - 45 minute interval, the δ value is relatively high (2.63 - 3.44), indicating that students are concentrated.

[0066] In the 5 - 10 minute and 40 - 45 minute intervals, the δ value is relatively low (0.84 - 2.27), indicating that students are distracted.

[0067] Classroom activity parameter (ε): In the 10 - 35 minute interval, the ε value is relatively high (1.25 - 1.45), indicating frequent classroom interaction.

[0068] In the 5 minute and 40 - 45 minute intervals, the ε value is relatively low (1.18 - 1.22), indicating a decrease in classroom interaction.

[0069] Classroom adaptability parameter (η): In the 10 - 35 minute interval, the η value is relatively high (0.92 - 0.98), indicating that students adapt well to the environment.

[0070] In the 5 minute and 40 - 45 minute intervals, the η value is relatively low (0.89 - 0.91), indicating that students have poor adaptability to the environment.

[0071] Analysis management parameter value (ω): In the 15 - 35 minute interval, the ω value is higher than 2.7, indicating that the classroom teaching demonstration meets the standard.

[0072] In the 5 minute, 10 minute and 40 - 45 minute intervals, the ω value is lower than 2.7, indicating that the classroom teaching demonstration does not meet the standard.

[0073] 8. Conclusion The test results show that: This system can accurately capture students' thinking concentration, classroom activity and classroom adaptability, and output the analysis management parameter value through the analysis management model.

[0074] In the middle of the class (15 - 35 minutes), the system feedback shows that the classroom teaching demonstration meets the standard, which is consistent with the actual classroom performance.

[0075] At the beginning and end of the class (5 - 10 minutes and 40 - 45 minutes), the system feedback indicates that the demonstration of classroom teaching does not meet the standard, which is consistent with the phenomena of students' distracted attention and reduced interaction.

[0076] Through experimental verification, it is proved that the system has excellent technical effects, can achieve in - depth understanding and real - time feedback on the classroom teaching process, and improves the accuracy and comprehensiveness of teaching auxiliary feedback.

[0077] The present invention provides an intelligent auxiliary system for classroom teaching based on large - language models. It obtains the reference parameters of students' thinking concentration based on the collected video data, obtains the reference parameters of classroom activity based on the collected audio data, and obtains the reference parameters of classroom adaptability based on the collected environmental data. Then, it outputs the analysis and management parameter values according to the fitted analysis and management model, and conducts feedback based on the analysis and management values. Through innovative data collection and processing technologies, the present invention realizes in - depth understanding and real - time feedback on the classroom teaching process, improves the accuracy and comprehensiveness of teaching auxiliary feedback, quantifies the feedback score of the classroom based on data calculations, and solves the problems of the existing classroom teaching auxiliary systems. On the one hand, they lack the ability of in - depth understanding and real - time feedback on teaching content, and on the other hand, there are limitations in data collection and processing, unable to comprehensively capture classroom interactions and students' learning states, resulting in inaccurate and incomplete analysis results.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. The classroom teaching intelligent auxiliary system based on the large language model is characterized by: Includes the following components: A data collection module (100) collects classroom video, audio and environmental data through multiple configured high-definition cameras, microphone arrays and sensor networks; A data processing module (200) is wirelessly connected to the data acquisition module (100) and performs real-time processing and analysis after acquiring the acquired data; The feedback module (300) is data-connected to the data processing module (200), generates real-time feedback based on the analysis results, and transmits it to the teacher terminal for display.

2. The classroom teaching intelligent auxiliary system based on a large language model according to claim 1 is characterized in that: The data acquisition module (100) is also embedded with a data preprocessing unit, which performs data preprocessing in sequence after acquiring the collected data; Among them, the data preprocessing steps specifically include: data cleaning, data enhancement and data standardization.

3. The classroom teaching intelligent auxiliary system based on a large language model according to claim 2 is characterized in that: When the data collection module (100) collects various types of data, it specifically includes: Video data collection: The classroom images are captured in real time by three high-definition cameras evenly distributed in front of the classroom to ensure full and positive coverage of the classroom space; Audio data collection: A group of microphone arrays evenly arranged in front of the classroom are used to collect classroom speech in real time, and sound pressure fluctuations are extracted by combining sound pressure recognition technology; Environmental data collection: Collect classroom temperature, humidity, and lighting environment data through the sensor network to assist in analyzing students' learning status.

4. The classroom teaching intelligent auxiliary system based on a large language model according to claim 3 is characterized in that: The data processing module (200) acquires the collected data and performs real-time processing and analysis, which specifically includes the following steps: S1: Obtain the students’ thinking concentration index based on the collected video data; S2: generating a sound pressure fluctuation curve based on the collected audio data, and obtaining a classroom activity indicator based on the sound pressure fluctuation curve; S3: Obtain classroom adaptability parameters based on the collected environmental data; S4: Establishing an analysis and management model, inputting the student's concentration parameter, the class activity parameter and the class adaptability parameter, and outputting an analysis and management parameter value; S5: Perform feedback scoring based on the analysis management parameter value, and transmit the analysis result to the feedback module (300) in real time.

5. The classroom teaching intelligent auxiliary system based on a large language model according to claim 4 is characterized in that: The steps of obtaining the student's thinking concentration indicator based on the collected video data specifically include the following steps: S1: A set of high-definition cameras capture classroom images in real time; S2: During the acquisition time, reference points are taken at uniform time intervals, and the corresponding reference points obtain the proportion of students' faces facing the blackboard at the current moment; S3: Establish a model for obtaining the parameter of student thinking concentration, input the facing ratios at different reference points in time sequence, and output the parameter of student thinking concentration; The model for obtaining the student thinking concentration parameter is as follows: ； Among them, δ is the parameter indicating the concentration of students' thinking; α1 is the facing ratio at the first reference point, %; α n is the facing ratio at the nth, i.e. the last reference point, %; n is the number of reference points; α is the facing ratio at the corresponding reference point, %; -1.29 and 0.67 are adjustment constants.

6. The classroom teaching intelligent auxiliary system based on a large language model according to claim 5 is characterized in that: Acquiring the classroom activity indicator based on the sound pressure fluctuation curve specifically includes the following steps: Q1: Generate a sound pressure fluctuation curve based on the collected audio data; Q2: Select the reference point moment in S2; Q3: Obtain the sound pressure value at the corresponding moment, establish a classroom activity parameter model, input the sound pressure values ​​at different reference points in time sequence, and output the classroom activity parameter; Among them, the classroom activity parameter model is specifically: ； Among them, ε is the classroom activity parameter; D1 is the sound pressure value at the first reference point, dB; D n is the sound pressure value at the nth, i.e. the last reference point, in dB; n is the number of reference points; D is the sound pressure value at the corresponding reference point, in dB; -0.87 and 1.33 are adjustment constants.

7. The classroom teaching intelligent auxiliary system based on a large language model according to claim 6 is characterized in that: Acquiring the classroom adaptability parameter based on the collected environmental data specifically includes the following steps: H1: Collect classroom temperature, humidity, and lighting environment data in real time through a sensor network; H2: Get the average temperature, average humidity and average light data during the collection period; H3: Establish a classroom adaptability parameter model, input various environmental data in H2, and output the classroom adaptability parameter; The classroom adaptability parameters are specifically: ； Among them, η is the classroom adaptability parameter; n is the number of reference points; A is the average temperature, ℃; B is the average humidity, %; C is the average light, lux; -1, 1.09, 1.45 and 3.70 are adjustment constants; dx is the integral operation.

8. The classroom teaching intelligent auxiliary system based on a large language model according to claim 7 is characterized in that: The analysis management model established is specifically: ； Among them, ω is the analysis and management parameter value; δ is the parameter indicating the concentration of students' thinking; ε is the parameter indicating the activity of the classroom; η is the parameter indicating the adaptability of the classroom; ln2.09, 1 and 1.03 are adjustment constants.

9. The classroom teaching intelligent auxiliary system based on a large language model according to claim 8 is characterized in that: The feedback scoring based on the analysis management parameter value is specifically as follows: when the analysis management parameter value is higher than 2.7 or 2.72, feedback is given that the classroom teaching has met the exemplary standards, and the analysis management parameter value and the scoring result are simultaneously sent to the feedback module (300).

Images