Education course intelligent recording and broadcasting management method

Abstract

The invention belongs to the technical field of education data management, and particularly discloses an intelligent recording and broadcasting management method for education courses, which comprises the following steps of: acquiring multi-modal data in real time through a multi-sensor group, and comprehensively capturing teaching actions of teachers, limb actions of students, speech audios and running states of teaching equipment; further comparing the feature combination with a pre-stored template feature combination to identify different teaching stages; recording modes are automatically switched according to the identified teaching stage, a main and auxiliary picture polling switching rule is established, and videos recorded at different camera positions are integrated into a composite video with synchronous main and auxiliary pictures; key teaching links of teaching videos are automatically identified, index tags are established, the videos are classified and stored in combination with a preset classification rule, and accordingly video contents meeting user requirements are pushed to different users. Structured management and rapid retrieval of recorded and broadcast video content are realized, and personalized service and utilization efficiency of teaching resources are enhanced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of educational data management and relates to an intelligent recording and broadcasting management method for educational courses. BACKGROUND

[0002] Under the background of the rapid development of network video education informatization, the recording and broadcasting system for educational courses, as an important tool for accumulating and sharing teaching resources, directly affects the teaching quality and resource utilization effect. However, the existing recording and broadcasting management method for educational courses generally faces the technical bottleneck of fixed recording strategy, which cannot dynamically adjust the recording mode and picture layout according to the real-time situation of the classroom, resulting in important teaching links and interactive details being easily missed. Therefore, it is of great significance to research and develop an intelligent recording and broadcasting management method for educational courses that can intelligently identify teaching stages and automatically adjust recording strategies for improving teaching quality and optimizing teaching resource allocation.

[0003] The current recording and broadcasting system for educational courses has the following defects in the technical aspect: 1. Fixed camera positions are mostly used for recording, which cannot automatically adjust the recording strategy according to the actual situation of the classroom, such as teaching stage conversion and teacher-student interaction, resulting in important teaching links being missed, and the performance of capturing teacher-student interaction is not good, making it difficult to fully present the dynamic teaching scene in the classroom, thereby affecting the efficiency of the recorded course video in teaching resource utilization.

[0004] 2. The video processing work after recording highly depends on manual editing, labeling and classification, which not only consumes a lot of time cost, but also is easily affected by subjective factors, resulting in inconsistency and inaccuracy of the processing results. At the same time, there is a lack of intelligent classification and management of recorded courses, which makes it inconvenient to search and use recorded videos, and reduces the utilization efficiency of teaching resources. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the application proposes the following technical solution: an intelligent recording and broadcasting management method for educational courses, comprising the following contents: S1, real-time collection of multi-modal data by a multi-sensor group arranged in the classroom, the multi-modal data including teacher teaching action features and student body action features, teacher and student speaking audio features, and teaching equipment running state features, the sensor group including a visual sensor, an audio sensor and an Internet of Things device state monitor.

[0006] S2, based on the preset teaching template library and the real-time collected multi-modal data, automatically identifying the current teaching stage as one of theory explanation, classroom interaction, practical operation or group discussion.

[0007] S3, automatically switch the corresponding recording mode according to the identified teaching stage, and real-time synthesis the videos recorded by different positions according to the time sequence to generate a composite video with the main and auxiliary pictures synchronized.

[0008] S4, automatically identify each key teaching node in the teaching recording process, and add the corresponding time point to the composite video time axis to generate a structured course index.

[0009] S5, based on the structured course index, combined with the preset classification rule, the processed composite video is automatically stored in the corresponding classified directory and pushed to the preset related user group.

[0010] Compared with the prior art, the beneficial effects of the present application are as follows: 1, the present application realizes the comprehensive capture and recording of the teacher's teaching action, the student's body action, the speaking audio and the running state of the teaching equipment by real-time collection of multi-sensor group multi-modal data, and combines it with the pre-stored template feature to realize the accurate identification of different teaching stages, and provides the basis for automatic switching of recording mode and video synthesis.

[0011] 2, the present application automatically switches the corresponding recording mode according to the identified teaching stage, establishes the main and auxiliary picture rotation switching rule, and real-time synthesizes the videos recorded by different positions according to the time sequence to generate a composite video with the main and auxiliary pictures synchronized, realizes the dynamic capture and efficient recording of the teaching recording process, and improves the pertinence and applicability of the recording video.

[0012] 3, the present application accurately locates the key teaching link of the teaching video through the structured course index, combines the preset classification rule to finely divide the video content, and automatically stores the classified video in the corresponding directory, thereby greatly improving the organization and retrievability of the video resource, and based on the user group information, the video content meeting the user demand is pushed, and the personalized service and utilization efficiency of the teaching resource are enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used for the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0014] Figure 1 The method embodiment step flowchart of the present application. DETAILED DESCRIPTION

[0015] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present application.

[0016] Please refer to Figure 1 As shown in the figure, the present application provides an education course intelligent recording and broadcasting management method, which includes the following contents: S1, real-time collection of multi-modal data by a multi-sensor group arranged in a classroom, the multi-modal data including teacher teaching action features and student body action features, teacher and student speaking audio features, and teaching equipment running state features, the sensor group including a visual sensor, an audio sensor, and an Internet of Things device state monitor.

[0017] Among them, the teacher teaching action features include questioning actions, blackboard writing actions, or demonstration actions, etc., the student body action features include raising hands actions, standing actions, experiment actions, discussion actions, etc., the teacher and student speaking audio features include the speaking sound source positions and audio volumes of the teachers and students, and the teaching equipment running state features include the existing page turning actions or blackboard writing actions of the electronic whiteboard display screen, and the existing standby states or running states of the experimental instrument power supply.

[0018] In a preferred embodiment, the multi-modal data is collected in real time by a multi-sensor group arranged in a classroom, wherein: the visual sensor includes a wide-angle camera deployed on the front / back wall of the classroom, and a main camera and auxiliary cameras for tracking teacher teaching actions and student body actions, the wide-angle camera is used to collect a panoramic picture of the classroom, and the main camera and auxiliary cameras are respectively used to collect teacher teaching action features or student body action features within the display range of the main picture and auxiliary pictures.

[0019] The audio sensor includes a microphone array uniformly distributed on the top of the classroom, which is used to collect the teacher and student speaking audio features.

[0020] The Internet of Things device state monitor is connected to the electronic whiteboard and the experimental instrument, and is used to collect the teaching equipment running state features.

[0021] The present application realizes comprehensive capture and recording of teacher teaching actions, student body actions, speaking audio, and teaching equipment running states by real-time collection of multi-modal data by a multi-sensor group, provides a rich data basis for subsequent teaching stage recognition and recording mode adjustment, and improves the completeness and accuracy of the recording and broadcasting content.

[0022] S2, automatically identifying the current teaching stage as one of theory teaching, classroom interaction, practice operation or group discussion based on the preset teaching template library and the real-time collected multi-modal data.

[0023] In a preferred embodiment, the S2 content includes: retrieving the pre-stored corresponding template feature combination of each teaching stage in the teaching template library, and the template feature combination is generated by the multi-modal data in different feature combinations.

[0024] The multi-modal data collected by the multi-sensor group in real time is compared with the pre-stored corresponding template feature combination of each teaching stage, and when the corresponding template feature combination of a certain teaching stage matches the data feature combination collected by the sensor in real time, the teaching stage is triggered.

[0025] In a further preferred embodiment, the corresponding template feature combination of each teaching stage includes: the corresponding template feature combination of the theory teaching stage is generated by the teacher position concentration, the high frequency of teaching equipment operation, and the student attention concentration.

[0026] Specifically, the corresponding template feature of the teacher position concentration refers to the duration of the teacher position concentrated in the podium area exceeding the preset concentration duration, the corresponding template feature of the high frequency of teaching equipment operation refers to the operation frequency of the page turning action or the board writing action of the electronic whiteboard display screen within the set period being greater than the preset operation frequency threshold, and the corresponding template feature of the student attention concentration refers to the student attention distribution concentration degree exceeding the preset concentration degree threshold. Wherein, the ratio of the total duration of the students keeping the state of facing the podium area to the total duration of the set observation period is defined as the average duration persistence rate, and then the student attention distribution concentration degree is calculated by multiplying the proportion of the number of students facing the podium area and the average duration persistence rate, and the student attention distribution concentration degree is used to quantify the degree of attention of the students to the teaching content in the theory teaching stage.

[0027] The corresponding template feature combination of the classroom interaction stage is generated by the teacher moving trajectory range expansion and the student speaking event triggering.

[0028] Specifically, the corresponding template feature of the teacher moving trajectory range expansion refers to the teacher moving trajectory exceeding the podium area and its maintenance duration exceeding the preset moving duration, and the corresponding template feature of the student speaking event triggering refers to the student having a hand raising action or a standing action.

[0029] The corresponding template feature combination of the practice operation stage is generated by the experimental equipment running state, the student experimental action high frequency, and the multi-sound source concurrency.

[0030] Specifically, the experiment equipment operation state corresponding template feature indicates that the power supply of the experiment instrument changes from standby state to operation state, the student experiment action high-frequency corresponding template feature indicates that the contact frequency of the student hand action and the experiment instrument in the set period exceeds the preset contact frequency, and the multi-sound source concurrent corresponding template feature indicates that multiple student speaking audio segments exist in the same time interval.

[0031] The group discussion stage is generated by the student group dispersion clustering and the multi-sound source concurrent corresponding template feature combination.

[0032] Specifically, the student group dispersion clustering corresponding template feature indicates that multiple independent audio clusters containing at least 3 people exist in the classroom, and the distance between the clusters is greater than the set distance.

[0033] The present application realizes accurate identification of the theoretical explanation, classroom interaction, practical operation and group discussion teaching stages by the technical feature that the real-time data are compared with the pre-stored template features, provides a basis for automatic switching of the recording mode and video synthesis, and enhances the pertinence and applicability of the recording and broadcasting content.

[0034] S3, according to the identified teaching stage, automatically switch the corresponding recording mode, and real-time synthesis the videos recorded by different positions according to the time sequence to generate a composite video with synchronized main and auxiliary pictures.

[0035] In a preferred embodiment, the automatic switching of the corresponding recording mode according to the identified teaching stage comprises: when the theoretical explanation stage is identified, the main camera is controlled to adjust to the medium focal length shooting mode to shoot the panoramic picture of the electronic whiteboard display screen, and each auxiliary camera is controlled to shoot the teacher's teaching picture or the student area panoramic picture.

[0036] When the classroom interaction stage is identified, the main camera is controlled to switch to the corresponding close-up picture of the teacher or the speaking student, and each auxiliary camera is controlled to switch to the wide-angle lens to cover the panoramic picture of the teacher and the student area or the corresponding close-up picture of the hand-raising student.

[0037] When the practical operation stage is identified, the main camera is controlled to switch to the near focal length shooting mode to shoot the experiment instrument parameter display picture, and each auxiliary camera is controlled to shoot the corresponding angle close-up picture of the teacher's demonstration action or the student's experiment action, and the experiment instrument is, for example, a microscope, a circuit board, etc.

[0038] When the group discussion stage is identified, the main camera and each auxiliary camera are controlled to rotate and patrol each group, and a main and auxiliary picture rotation switching rule is set, the main and auxiliary picture rotation switching rule comprises that the priority of the group in which the teacher participates is higher than that of the student spontaneous discussion group, and the priority of the group with the largest audio volume is higher than that of the group with high visual activity.

[0039] Specifically, the visual activity refers to the proportion of the posture conversion frequency of the limb action of the team member collected by the visual sensor within a certain time interval compared with the preset reference conversion frequency.

[0040] The shooting range of the medium focal length shooting mode, the wide-angle lens coverage, the short focal length shooting mode, and the classroom panoramic picture is calculated in advance, which can ensure that the corresponding target picture size of the teacher demonstration action or the student operation hand or the experimental instrument parameter display is appropriate, such as the focal length being in the interval of [X, Y] millimeters, X and Y being set constants.

[0041] When the recording mode is automatically switched, the speaking audio of the teacher or student speaking object in the main picture is used as the accompanying audio of the recording video, and the microphone beam direction and width of the microphone array in the speaking source collection position of the teacher or student speaking object in the main picture are adjusted synchronously to enhance the audio input of the directional microphone and further suppress the environmental noise.

[0042] Specifically, the corresponding adjustment method of the microphone beam direction and width is a mature existing technology. Since the existing technical solution is perfect and universally applicable, this part of the existing technology will not be described in detail in order to highlight the focus of the description and make it simple and clear.

[0043] In a further preferred embodiment, when the recording mode is automatically switched, the content further includes: defining a multi-picture conflict condition triggered by the same teaching stage or multiple teaching stages existing in the same period, the multi-picture conflict condition triggered by the same teaching stage, for example, two groups exist speaking audio at the same time, and the multiple teaching stages existing in the same period, for example, the combination of the corresponding template features of the practical operation stage and the group discussion stage exists at the same time in the same period.

[0044] If the multi-picture conflict condition triggered by the same teaching stage exists, all teaching pictures collected by multiple cameras are integrated, the priority of each teaching picture is sorted by analyzing the correlation degree of each teaching picture with the current teaching stage and the picture quality, and the main and auxiliary pictures are switched in the order of decreasing priority.

[0045] The priority of each teaching picture is sorted by analyzing the correlation degree of each teaching picture with the current teaching stage and the picture quality, which is specifically: obtaining a preset teaching element library of the current teaching stage, extracting the teaching elements belonging to the multi-modal data in each teaching picture through image recognition technology, comparing them with the preset teaching element library, counting the number of matched teaching elements in each teaching picture, and then taking the proportion of the number of matched teaching elements compared with the preset reference element number as the correlation degree of each teaching picture with the current teaching stage.

[0046] The image recognition technology is used to detect the definition and color restoration of the teaching pictures, and obtain various quality indexes of the teaching pictures, and then the cumulative quality of the teaching pictures is obtained, wherein the quality indexes include definition, color restoration, etc., and the better the quality index is, the better the picture quality is.

[0047] The correlation degree and picture quality of each teaching picture and the current teaching stage are weighted and solved according to the preset weight proportion, and the comprehensive score of each teaching picture is obtained, for example, the correlation degree and picture quality of a certain picture and the current teaching stage are p1 and p2 respectively, the correlation degree weight proportion is w1, and the picture quality weight proportion is w2, then the comprehensive score U of the picture is w1×p1+w2×p2, and the priority of each teaching picture is sorted according to the comprehensive score, and the teaching picture with the highest comprehensive score has the highest priority.

[0048] Through the built-in device state monitoring module, the combination of periodic detection and event triggered detection is used to continuously monitor various devices involved in the processing of the teaching pictures, including but not limited to cameras, microphones, display terminals, etc.

[0049] Periodic monitoring is to check the running parameters of various devices every fixed time, and record the monitoring results, such as the power state, network connection state, hardware working indicator state, etc. The event triggered monitoring starts immediately when the device sends an abnormal signal, such as device overheating alarm, signal loss prompt, etc., and conducts in-depth detection on the device to determine the fault type and severity.

[0050] The monitoring results are obtained, and whether there is a fault in various devices is judged according to the preset device fault judgment standard. For example, if the network connection state of a certain device shows that it is disconnected, and it has not been restored in the last three periodic monitoring, it is determined that the device has a network fault; if the hardware working indicator of the device shows an abnormality, and the event triggered monitoring confirms that the hardware is physically damaged, it is determined that the device has a hardware fault.

[0051] When it is determined that there is a faulty device, the teaching picture directly associated with the faulty device is obtained, which is recorded as the faulty picture, and then the main and auxiliary pictures are dynamically exchanged according to the following rules: when the faulty picture is the main picture, the corresponding priority of each auxiliary picture currently collected is obtained, and then the auxiliary picture with the highest priority is selected as the new main picture for full screen display, at the same time, the original main picture is reduced and placed in a specific area of the screen as an auxiliary picture display, and attempts are made to restore the normal display of the original main picture through a backup device or a repaired faulty device, and the specific area is, for example, the lower right corner area; when the faulty picture is an auxiliary picture, the next priority picture of the auxiliary picture in the priority order is obtained as the replacement picture of the auxiliary picture, at the same time, the auxiliary picture is removed from the display, so as to optimize the display effect of the overall picture.

[0052] By refining the device control logic and sensor trigger conditions of each stage, the self-switching smoothness of the main and auxiliary pictures can be enhanced.

[0053] In a further preferred embodiment, the real-time synthesis of videos recorded by different positions according to time sequence to generate a composite video with synchronized main and auxiliary pictures comprises: setting the corresponding recording end actions of each teaching stage, and triggering the recording picture end instruction of each teaching stage when the visual sensor marks the recording end action. For example, in the classroom interaction stage, after the student speaking event is triggered, the wide-angle camera marks the speaking action of the student in real time, and when the end action, i.e., the standing action, is marked as the sitting action, the corresponding main and auxiliary picture end instruction of the classroom interaction stage is triggered.

[0054] The videos recorded by different positions are time-aligned based on a unified clock source, and the display layout of the main and auxiliary pictures and the audio playback layout are dynamically adjusted according to the teaching stage.

[0055] According to the identified teaching stage, the corresponding recording mode is automatically switched, the main and auxiliary picture rotation switching rules are established, and the videos recorded by different positions are real-time synthesized according to time sequence to generate a composite video with synchronized main and auxiliary pictures, which realizes dynamic capture and efficient recording of the teaching recording process, and improves the viewing and practicality of the recording video.

[0056] S4, in the teaching recording process, automatically identify each key teaching node, and add the corresponding time point to the composite video time axis to generate a structured course index.

[0057] In a preferred embodiment, the S4 content comprises: based on the multi-modal data collected by the multi-sensor group in real time, identifying each key teaching node and its time point, including electronic whiteboard page turning action, teacher speech keywords, and student raising action.

[0058] Index labels are added to the corresponding time points of the identified key teaching nodes, and then the index labels are added to the video time axis according to the time points, thereby generating a structured course index, and the index labels record the teaching content type of the corresponding key teaching node.

[0059] The video timeline is a timeline representing the video playback progress, and the addition of the mark point is actually adding an identification for a specific time point on the timeline. For example, for the electronic whiteboard page turning key node, the teaching stage index label can be marked as "teaching content switching"; for the teacher speech keyword triggering key node, the teaching stage index label can be marked as "key knowledge explanation"; and for the student raising hand key node, the teaching stage index label can be marked as "student questioning or interaction". Through such a structured course index, users can conveniently find and locate the key links in the teaching recording process, and improve the utilization efficiency of teaching resources.

[0060] In specific operation, a visual mark icon can be added for each mark point on the interface of the video playback software, and the user can quickly jump to the corresponding key teaching node by clicking the icon. At the same time, the position information of the mark point is recorded in the metadata of the video file, so that the retrieval and positioning of the mark point can be realized in different playback devices and software.

[0061] In a further preferred embodiment, the identification of each key teaching node containing the electronic whiteboard page turning action, the teacher speech keyword, and the student raising hand action and the time point thereof comprises: obtaining a course outline, continuously capturing the electronic whiteboard page turning action through an Internet of Things device state monitor, and then comparing the display screen courseware content after the electronic whiteboard page turning with the course outline. If the display screen courseware content matches the main knowledge point content or chapter title content in the course outline, the time point is determined to be marked, and the time point is marked as the key teaching node corresponding to the electronic whiteboard page turning action.

[0062] The teacher speech audio collected by the audio sensor in real time is converted into text, which is then compared with a preset term library. When a term in the term library is identified in the text, the time point at which the term appears is marked, and the time point is marked as the key teaching node corresponding to the teacher speech keyword.

[0063] The student body action recording picture is captured in real time through a visual sensor, and the raising hand action is set as the recognition action. When the recognition action is triggered, if the corresponding student body action in the recording picture has a standing action, the time point corresponding to the raising hand action is marked, and the time point is marked as the key teaching node corresponding to the student raising hand action.

[0064] S5, based on the structured course index, combining a preset classification rule, automatically storing the processed composite video to a corresponding classification directory, and pushing to a preset related user group.

[0065] In a preferred embodiment, the S5 content includes: based on the structured course index, the processed recorded video is divided into video segments corresponding to each key teaching node, and then based on the picture content of each video segment, the course content features of each key teaching node are determined. For example, for the key teaching node corresponding to the electronic whiteboard page turning action, the knowledge point title or example of the courseware is extracted; for the key teaching node corresponding to the teacher's speech keyword, the knowledge point keyword involved is extracted; for the key teaching node corresponding to the student's hand raising action, the knowledge point keyword involved in the student's speech audio is extracted. These features are extracted and quantified through natural language processing technology, semantic analysis algorithm and teaching knowledge graph and other tools.

[0066] Based on the knowledge system in the teaching field, course classification rules including different course content features are established, and then the course content features of each key teaching node are matched with the established course classification rules to determine the corresponding course classification category of the processed composite video.

[0067] The course classification rules include: dividing the teaching content into multiple dimensions, such as subject field, teaching stage, knowledge point type and teaching interaction degree, each dimension is further divided into multiple categories, such as mathematics, physics, chemistry and other categories corresponding to the subject field, primary school, junior high school, high school and other categories corresponding to the teaching stage, concept explanation, example analysis, experiment demonstration and other categories corresponding to the knowledge point type, and low interaction, medium interaction, high interaction and other categories corresponding to the teaching interaction degree, forming a multi-level classification system.

[0068] By analyzing the preset related user group information, the user group information includes user's role, historical browsing record, learning progress and other information, the personalized push video category is established for the user, and then the composite video stored in the corresponding classification directory is pushed to the corresponding user, the related user group includes teachers, students, teaching management personnel and the like, each user group has different teaching needs and focuses, for example, for the teacher user group of different subject fields, the video content related to the course design of the corresponding subject field is pushed; for the student user group of different teaching stages, the video content related to the knowledge point explanation and exercise solution of the corresponding teaching stage is pushed.

[0069] The present application accurately locates the key teaching link of the teaching video through the structured course index, combines the preset classification rules to finely divide the video content, and automatically stores the classified video in the corresponding directory, thereby greatly improving the organization and retrievability of the video resources, at the same time, based on the user group information, the video content meeting the user's needs is pushed, and the personalized service and utilization efficiency of the teaching resources are enhanced.

[0070] It should be noted that: the above formula, through the principle of dimensional consistency and mathematical standardization means (for example, normalization processing, dimensionless parameter conversion or unit system unification), different properties of physical quantities can be translated into unitless standard values or the same dimension superimposable parameters, so as to eliminate the interference of different dimensions on the operation logic, so that the formula retains the original data distribution characteristics while having mathematical operation rationality and objective law adaptability. The above merely illustrates the exemplary embodiments of the present application, and cannot limit the scope of the present application.

Claims

1. A method for intelligent recording and broadcasting management of educational courses, characterized in that: Includes the following: S1. Collect multimodal data in real time using a multi-sensor group deployed in the classroom. The multimodal data includes characteristics of the teacher's teaching movements and students' body movements, audio characteristics of the teacher's and students' speeches, and characteristics of the operating status of teaching equipment. The sensor group includes visual sensors, audio sensors, and IoT device status monitors. S2. Automatically identify the current teaching stage as theoretical explanation, classroom interaction, practical operation, or group discussion based on a preset teaching template library and real-time multimodal data collection; S3. Automatically switch the corresponding recording mode according to the identified teaching stage, and synthesize the videos recorded by different cameras in real time according to the time sequence to generate a composite video with synchronized main and auxiliary images; S4. Automatically identify key teaching nodes during the teaching recording process and add their corresponding time points to the composite video timeline to generate a structured course index; S5. Based on the structured course index and the preset classification rules, the processed composite video is automatically stored in the corresponding classification directory and pushed to the preset relevant user groups.

2. The method for intelligent recording and broadcasting management of educational courses according to claim 1, characterized in that: The multi-modal data is collected in real time by a multi-sensor group arranged in the classroom, wherein: The visual sensors include wide-angle cameras deployed on the front and back walls of the classroom, as well as a main camera and auxiliary cameras that track the teacher's teaching movements and students' body movements. The wide-angle camera is used to capture a panoramic view of the classroom, while the main camera and auxiliary cameras are used to capture the teacher's teaching movement characteristics or students' body movement characteristics within the display range of the main screen and auxiliary screens respectively. The audio sensor includes a microphone array evenly distributed on the top of the classroom to collect audio features of teachers and students' speeches; The IoT device status monitor is connected to the electronic whiteboard and experimental instruments to collect the operating status characteristics of the teaching equipment.

3. The method for intelligent recording and broadcasting management of educational courses according to claim 1, characterized in that: The S2 content includes: Retrieving the corresponding template feature combination of each teaching stage pre-stored in the teaching template library, wherein the template feature combination is generated by multimodal data with different feature combinations; The multimodal data collected by the multi-sensor group in real time are compared with the pre-stored template feature combinations corresponding to each teaching stage. When the template feature combination corresponding to a certain teaching stage matches the data feature combination collected by the sensor in real time, the teaching stage is triggered.

4. The method for intelligent recording and broadcasting management of educational courses according to claim 3, characterized in that: The corresponding template feature combinations for each teaching stage include: The theoretical explanation stage is generated by combining the corresponding template features of the teacher's concentrated position, high frequency of teaching equipment operation, and student concentration; In the classroom interaction stage, the teacher's movement trajectory is expanded, and the student's speech event triggers the generation of corresponding template feature combinations; The practical operation stage is generated by the combination of the operating status of the experimental equipment, the high frequency of the students' experimental actions, and the corresponding template features of multiple sound sources; The group discussion stage is generated by dispersed clustering of student groups and concurrent combination of corresponding template features of multiple sound sources.

5. The method for intelligent recording and broadcasting management of educational courses according to claim 2, characterized in that: The automatic switching of the corresponding recording mode according to the identified teaching stage includes: When the theoretical explanation phase is identified, the main camera is controlled to adjust to the medium focal length shooting mode to capture a panoramic view of the electronic whiteboard display screen, and the auxiliary cameras are controlled to capture the teacher's teaching screen or the student area panoramic view; When the classroom interaction phase is identified, the main camera is controlled to switch to a close-up of the teacher or the speaking student, and the auxiliary cameras are controlled to switch to a wide-angle lens to cover the panoramic view of the teacher and students or the close-up of the student who raised his hand; When it is identified as the practical operation stage, the main camera is controlled to switch to the close-focus shooting mode to shoot the experimental instrument parameter display screen, and the auxiliary cameras are controlled to shoot close-up pictures of the teacher's demonstration movements or the students' experimental movements at corresponding angles; When a group discussion phase is identified, the main camera and auxiliary cameras are controlled to patrol each group in turn, and the main and auxiliary screen patrol switching rules are set. The main and auxiliary screen patrol switching rules include the following: the teacher's participation group has a higher priority than the student's spontaneous discussion group, and the group with the loudest audio volume has a higher priority than the group with the highest visual activity; When the recording mode is automatically switched, the corresponding speech audio of the teacher or student speaking on the main screen is used as the accompanying audio of the recorded video, and the microphone beam direction and width of the sound source recording position of the teacher or student speaking on the main screen in the microphone array are synchronously adjusted to enhance the audio input of the directional microphone.

6. The method for intelligent recording and broadcasting management of educational courses according to claim 5, characterized in that: When the recording mode is automatically switched, the content also includes: Define the multi-screen conflict condition as triggering multiple devices to run in the same teaching stage or the existence of multiple teaching stages in the same period; If the multi-screen conflict condition is triggered in the same teaching stage, all teaching screens captured by multiple cameras are integrated, and each teaching screen is prioritized by analyzing its relevance to the current teaching stage and its image quality. The main and auxiliary screens are switched in descending order of priority. Through the built-in device status monitoring module, a combination of periodic detection and event-triggered detection is used to continuously monitor various devices involved in teaching screen processing; Obtain monitoring results and determine whether various types of equipment have faults based on preset equipment fault judgment standards; When a faulty device is detected, the teaching screen directly associated with the faulty device is obtained and recorded as the faulty screen. The main and auxiliary screens are then dynamically swapped according to the following rules: When the fault screen is the main screen, obtain the corresponding priority ranking of the currently collected auxiliary screens, and then select the auxiliary screen with the highest priority as the new main screen for full-screen display. At the same time, the original main screen is reduced and placed in a specific area of ​​the screen as an auxiliary screen for display; When the faulty screen is a secondary screen, the next priority screen of the secondary screen in the priority sorting is obtained as a replacement screen for the secondary screen.

7. The method for intelligent recording and broadcasting management of educational courses according to claim 1, characterized in that: The method of synthesizing videos recorded by different cameras in real time in a time sequence to generate a composite video with synchronized main and auxiliary images includes: Set the corresponding recording end action for each teaching stage. When the visual sensor marks the recording end action, the corresponding recording screen end instruction for each teaching stage is triggered; Videos recorded by different cameras are time-aligned based on a unified clock source, and the main and auxiliary screen display layouts and audio broadcast layouts are dynamically adjusted according to the teaching stage.

8. The method for intelligent recording and broadcasting management of educational courses according to claim 1, characterized in that: The S4 content includes: Based on multimodal data collected in real time by a multi-sensor group, key teaching nodes and their timings, including electronic whiteboard page turning actions, teacher voice keywords, and student hand-raising actions, are identified; Add index tags to the corresponding time points of each identified key teaching node, and then add the index tags to the video timeline according to the time points, thereby generating a structured course index. The index tags record the teaching content type of the corresponding key teaching node.

9. The method for intelligent recording and broadcasting management of educational courses according to claim 8, characterized in that: The identification includes the key teaching nodes and their time points of the electronic whiteboard page turning action, the teacher's voice keywords, and the student's hand raising action, including: Obtain the course outline and continuously capture the electronic whiteboard page-turning action through the IoT device status monitor. Then, compare the courseware content on the display screen after the electronic whiteboard page is turned with the course outline. If it matches the main knowledge points or chapter titles in the course outline, mark the time point and mark it as the key teaching node corresponding to the electronic whiteboard page-turning action; The teacher's speech audio collected in real time by the audio sensor is converted into text, which is then compared with the preset terminology library. When a word in the terminology library is recognized in the text, the time point at which the word appears is marked, and the time point is marked as the key teaching node corresponding to the teacher's voice keyword; The visual sensor is used to capture the student's body movements in real time, and hand raising is set as the recognition action. When the recognition action is triggered, if the corresponding student's body movement in the recording screen includes a standing action, the corresponding time point of the hand raising action is marked, and the time point is marked as the corresponding key teaching node of the student's hand raising action.

10. The method for intelligent recording and broadcasting management of educational courses according to claim 1, characterized in that: The S5 content includes: Based on the structured course index, the processed recorded video is divided into video clips corresponding to each key teaching node, and then the course content characteristics of each key teaching node are determined based on the picture content of each video clip; Based on the knowledge system of the teaching field, a course classification rule containing different course content characteristics is established, and then the course content characteristics of each key teaching node are matched with the established course classification rule to determine the corresponding course classification category of the processed composite video; By analyzing the preset relevant user group information, a personalized push video category is established for the user, and then the composite video stored in the corresponding category directory is pushed to the corresponding user.

Images