Interactive biological virtual experiment teaching system based on deep learning

Abstract

The invention discloses an interactive biological virtual experiment teaching system based on deep learning, and the system comprises a dynamic scene intelligent construction module, a multi-modal operation perception module, a real-time feedback guidance module, a comprehensive report generation module, and a self-adaptive learning optimization module. Component interaction, state simulation and parameter adaptive adjustment are supported, learning progress and difficulty data are received to realize linkage, a knowledge graph contains principle, specification, equipment characteristics and safety knowledge, staged loading, dangerous operation early warning and real-time knowledge node pushing are supported, and the knowledge graph is used for learning. The multi-mode operation sensing module integrates computer vision and sensor technologies to collect action data, identifies an operation sequence and a pre-judgment intention based on a space-time sequence model, adopts an environment adaptability algorithm to adjust collected parameters, and achieves autonomous optimization of the parameters through historical data analysis.

Description

Technical Field

[0001] This invention relates to the field of virtual experimental teaching, and in particular to an interactive biological virtual experimental teaching system based on deep learning. Background Technology

[0002] In high school education, biology, as an experimental natural science, plays a crucial role in cultivating students' practical abilities and scientific literacy. Biology experimental teaching is an important part of cultivating high school students' practical abilities and scientific literacy. With the deep integration of information technology and education, virtual experimental teaching systems are gradually being applied to biology teaching. These systems aim to overcome the limitations of traditional experimental teaching in terms of venue, equipment, and safety by simulating experimental scenarios, capturing operational behaviors, and providing feedback and guidance. The related technologies involve multiple fields such as computer vision, sensor technology, deep learning, and knowledge graph construction, and are committed to improving the convenience, safety, and effectiveness of experimental teaching.

[0003] Currently, existing virtual experimental teaching systems for biology often struggle to create realistic 3D dynamic scenes. User interaction with scene components is insufficient, experimental parameter adjustments lack adaptability, the linkage between the scene and the learning process is weak, the knowledge graph lacks comprehensive coverage, and it's difficult to achieve phased loading and real-time warnings for dangerous operations. Furthermore, relevant knowledge nodes cannot be synchronously pushed during operation. In terms of operational perception, the collection of user action data is not precise enough, making it difficult to accurately identify operation sequences and predict operational intentions. Environmental adaptability is poor, collected parameters cannot be autonomously optimized, and recognition accuracy is low in complex environments. Feedback and guidance mechanisms lack specificity, and the adjustment of experimental difficulty is inadequate. While flexible enough, the limited interaction methods make it difficult to select appropriate guidance formats based on user response efficiency. The fixed text display method affects the timeliness and flexibility of guidance. The comprehensive report generation is not comprehensive enough, lacking effective integration of operational data, experimental results, and learning assessment information. It does not support online revision and record keeping, making it difficult to provide revision suggestions and update knowledge mastery analysis based on knowledge graphs. The learning assessment system is not perfect, lacking analysis of indicators such as operational coherence, unable to generate personalized learning paths, and the scenarios are difficult to adapt to individual learning needs, making it difficult to achieve a comprehensive and personalized learning process and effect assessment. Therefore, an interactive biological virtual experiment teaching system based on deep learning is proposed. Summary of the Invention

[0004] This invention provides the following technical solution: an interactive biological virtual experiment teaching system based on deep learning, comprising:

[0005] Dynamic Scene Intelligent Construction Module:

[0006] The system constructs a 3D scene based on experimental standard procedures and knowledge graphs, supporting component interaction, state simulation, and adaptive parameter adjustment; it receives learning progress and difficulty data to achieve linkage and sends interaction rules to the perception module; the knowledge graph contains principles, specifications, equipment characteristics, and safety knowledge, supports phased loading and dangerous operation warnings, and pushes knowledge nodes in real time related to experimental stages.

[0007] Multimodal operation sensing module:

[0008] It integrates computer vision and sensor technologies to collect motion data, identifies operation sequences and predicts intentions based on a spatiotemporal sequence model, and uses an environmentally adaptive algorithm to dynamically adjust the acquisition parameters, accurately capture operation details and standardization, and achieves autonomous parameter optimization through historical data analysis to improve recognition accuracy in complex environments.

[0009] Real-time feedback and guidance module:

[0010] It receives operation and scene data to generate feedback suggestions and dynamically adjusts the difficulty level; it supports multimodal interaction of voice, annotation, animation and text to achieve real-time guidance response that blends the virtual and real worlds; it intelligently selects the guidance form based on user response efficiency; and the text description adopts a layered display to support switching between operation prompts and principle explanations.

[0011] Comprehensive report generation module:

[0012] It integrates operational data, experimental results, and learning assessments to generate standardized reports, including time-series records, anomaly tracing, and knowledge analysis, to meet the needs of teaching assessment and experimental review; it supports online revision and record keeping, and provides revision suggestions and updates mastery analysis based on knowledge graphs.

[0013] Adaptive learning optimization module:

[0014] A dynamic evaluation system is built to collect accuracy, duration, and report quality indicators, increase operational consistency analysis, optimize training focus to improve learning effectiveness, generate personalized learning paths, push scenario adjustment instructions to the construction module, and provide evaluation data for report generation.

[0015] The dynamic scene intelligent construction module constructs a 3D scene based on the experimental standard process and knowledge graph. This 3D scene supports component interaction, state simulation, and adaptive parameter adjustment. At the same time, the dynamic scene intelligent construction module receives learning progress and difficulty data to achieve linkage and sends the interaction rules to the multimodal operation perception module. The knowledge graph contains principles, specifications, equipment characteristics, and safety knowledge, supports staged loading and dangerous operation warnings, and can push knowledge nodes in real time related to the experimental stage.

[0016] The multimodal operation perception module receives interaction rules from the dynamic scene intelligent construction module, integrates computer vision and sensor technology to collect action data, and identifies operation sequences and predicts intentions based on a spatiotemporal sequence model. At the same time, it adopts an environmentally adaptive algorithm to dynamically adjust the acquisition parameters to accurately capture operation details and standardization, and achieves autonomous parameter optimization through historical data analysis to improve the recognition accuracy in complex environments.

[0017] The real-time feedback guidance module receives operation data from the multimodal operation perception module and scene data from the dynamic scene intelligent construction module, generates feedback suggestions, and dynamically adjusts the difficulty level. This module supports multimodal interaction of voice, annotation, animation, and text, realizing real-time guidance response that blends the virtual and real worlds. It also intelligently selects the guidance format based on user response efficiency, and the text description adopts a layered display that supports switching between operation prompts and principle explanations.

[0018] The comprehensive report generation module integrates operational data from the multimodal operation perception module, experimental results related to the dynamic scene intelligent construction module, and learning evaluation from the adaptive learning optimization module to generate a standardized report. This report includes time-series records, anomaly tracing, and knowledge analysis, meeting the needs of teaching evaluation and experiment review. At the same time, this module supports online revision and trace recording, providing revision suggestions based on knowledge graphs and updating mastery analysis.

[0019] The adaptive learning optimization module constructs a dynamic evaluation system, collects accuracy, duration, and report quality indicators, adds operational coherence analysis, optimizes training focus to improve learning effectiveness, generates personalized learning paths, pushes scene adjustment instructions to the dynamic scene intelligent construction module, and provides evaluation data for the comprehensive report generation module.

[0020] Preferably, the adaptive learning optimization module analyzes the user's weaknesses in experimental operation coherence based on the standardized report generated by the comprehensive report generation module, generates targeted operation coherence training tasks, and pushes them to the dynamic scene intelligent construction module to construct a special training scene.

[0021] The adaptive learning optimization module receives the standardized report generated by the comprehensive report generation module, analyzes the user's weaknesses in the continuity of experimental operations in the report, generates targeted operation continuity training tasks based on the analysis results, and pushes the training tasks to the dynamic scene intelligent construction module, which then constructs specialized training scenarios.

[0022] Preferably, the experimental equipment component with physical collision detection function generated by the dynamic scene intelligent construction module calculates the physical feedback effect based on the collision force and angle when a collision occurs, and updates the component status and scene display in real time.

[0023] The dynamic scene intelligent construction module generates experimental equipment components with physical collision detection function. When the experimental equipment components collide, the module calculates the physical feedback effect based on the collision force and angle, and updates the status of the experimental equipment components and the scene display in real time.

[0024] Preferably, the safety knowledge association data in the knowledge graph of the dynamic scene intelligent construction module is automatically triggered based on the experimental equipment and operation steps in the scene, and safety warning information is pushed to the real-time feedback guidance module in advance when the user approaches a dangerous operation.

[0025] The knowledge graph of the dynamic scene intelligent construction module contains safety knowledge-related data. In the constructed scene, when experimental equipment and operation steps are identified, the safety knowledge-related data is automatically triggered. When the user approaches a dangerous operation, safety warning information is pushed to the real-time feedback and guidance module in advance.

[0026] Preferably, the environmental adaptability algorithm of the multimodal operation perception module dynamically adjusts the acquisition parameters of computer vision and sensors by identifying the lighting and background interference conditions of the operation environment, thereby improving the accuracy of operation action data acquisition in complex environments.

[0027] The environmental adaptability algorithm of the multimodal operation perception module identifies the lighting and background interference conditions of the operation environment, and dynamically adjusts the acquisition parameters of computer vision and sensors based on the recognition results to improve the accuracy of operation action data acquisition in complex environments.

[0028] Preferably, the environmental adaptability algorithm of the multimodal operation sensing module establishes an environmental parameter optimization model by analyzing the correlation between historical environmental data and the acquisition effect, thereby realizing the autonomous iterative adjustment of the acquisition parameters.

[0029] The environmental adaptability algorithm of the multimodal operation sensing module analyzes the correlation between historical environmental data and acquisition results, establishes an environmental parameter optimization model through analysis, and uses this model to realize the autonomous iterative adjustment of acquisition parameters.

[0030] Preferably, when the dynamic scene intelligent construction module constructs a 3D scene based on the experimental standard process and knowledge graph, it generates experimental equipment components with operation guidance labels for different experimental stages. The operation guidance labels dynamically adjust their display status and position as the experiment progresses to guide the user to complete the experimental operation.

[0031] When constructing a 3D scene based on experimental standard procedures and knowledge graphs, the dynamic scene intelligent construction module generates experimental equipment components with operation guidance icons for different experimental stages. The operation guidance icons dynamically adjust their display status and position as the experiment progresses to guide users to complete the experimental operations.

[0032] Preferably, when the multimodal operation perception module identifies the operation sequence and predicts the intention, it performs time-series analysis on the force, angle and speed characteristics of the operation action, generates an operation fluency score based on a deep learning model, and feeds back the score result to the real-time feedback guidance module in real time.

[0033] In the process of recognizing operation sequences and predicting intentions, the multimodal operation perception module performs temporal analysis on the force, angle and speed characteristics of the operation actions, generates an operation fluency score based on a deep learning model, and feeds the score result back to the real-time feedback guidance module in real time.

[0034] Preferably, the real-time feedback guidance module provides step-by-step correction guidance for users' non-standard operations based on the analysis results of the multimodal operation perception module, and demonstrates the standard operation process through animation, while also pushing relevant experimental principle explanations in conjunction with the knowledge graph.

[0035] The real-time feedback guidance module receives the analysis results from the multimodal operation perception module, provides step-by-step correction guidance for users' non-standard operations, demonstrates standard operating procedures through animation, and pushes relevant experimental principle explanations in conjunction with knowledge graphs.

[0036] Preferably, the online revision function in the comprehensive report generation module includes revision suggestion push, which judges the rationality of the user's revisions based on the experimental knowledge graph and standard operating procedures and provides reference opinions.

[0037] The online revision function in the comprehensive report generation module includes revision suggestion push. Based on the experimental knowledge graph and standard operating procedures, the module judges the rationality of the user's revisions and provides the user with reference opinions.

[0038] In summary, compared with existing technologies, this invention provides an interactive virtual experimental teaching system for biology based on deep learning, which has the following beneficial effects:

[0039] The dynamic scene intelligent construction module in this invention can construct realistic 3D scenes based on experimental standard procedures and knowledge graphs, supporting user interaction with components in the scene, simulating various states during the experimental process, and the relevant parameters can be adaptively adjusted according to the actual situation; by receiving learning progress and difficulty data, it realizes the linkage between the scene and the learning process, so that the scene can match the user's learning status; the knowledge graph covers comprehensive content such as principles, specifications, equipment characteristics and safety knowledge, supports phased loading of knowledge, can provide early warning of dangerous operations, and push relevant knowledge nodes in real time at different stages of the experiment, helping users to simultaneously master experimental knowledge, specifications and safety points during operation;

[0040] The multimodal operation perception module integrates computer vision and sensor technology, which can effectively collect user action data, accurately identify operation sequences and predict user operation intentions with the help of spatiotemporal sequence models.

[0041] The environmentally adaptive algorithm can dynamically adjust the acquisition parameters, accurately capture the details and standardization of the operation, and achieve autonomous optimization of the acquisition parameters by analyzing historical data. This ensures that the recognition accuracy remains high even in complex environments, thus ensuring accurate perception and judgment of user operations.

[0042] After receiving operation data and scenario data, the real-time feedback guidance module can generate targeted feedback suggestions and dynamically adjust the difficulty level of the experiment. It supports multimodal interaction methods such as voice, annotation, animation and text, realizing real-time guidance response that blends the virtual and real worlds. It can also intelligently select the appropriate guidance form based on the user's response efficiency. The text instructions are displayed in a layered manner, allowing users to flexibly switch between operation prompts and explanations of principles, improving the timeliness, relevance and flexibility of the guidance.

[0043] The comprehensive report generation module integrates operational data, experimental results, and learning assessment information to generate standardized reports that include time-series records, anomaly tracing, and knowledge analysis, meeting the needs of teaching assessment and experimental review. It supports online revision of reports and records revision history, provides suggestions for revision based on knowledge graphs, and updates the user's knowledge mastery analysis to help fully understand the experimental situation and learning outcomes.

[0044] The adaptive learning optimization module constructs a dynamic evaluation system, collects indicators such as accuracy, duration, and report quality, and adds operational consistency analysis to optimize training focus and improve learning effectiveness. It generates personalized learning paths, pushes scene adjustment instructions to the dynamic scene intelligent construction module to adapt the scene to individual learning needs, and provides evaluation data to the comprehensive report generation module to support comprehensive and personalized evaluation of the learning process and effect. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Please see Figure 1This invention provides a technical solution: an interactive biological virtual experiment teaching system based on deep learning, comprising:

[0048] Dynamic Scene Intelligent Construction Module:

[0049] The system constructs a 3D scene based on experimental standard procedures and knowledge graphs, supporting component interaction, state simulation, and adaptive parameter adjustment; it receives learning progress and difficulty data to achieve linkage and sends interaction rules to the perception module; the knowledge graph contains principles, specifications, equipment characteristics, and safety knowledge, supports phased loading and dangerous operation warnings, and pushes knowledge nodes in real time related to experimental stages.

[0050] Multimodal operation sensing module:

[0051] It integrates computer vision and sensor technologies to collect motion data, identifies operation sequences and predicts intentions based on a spatiotemporal sequence model, and uses an environmentally adaptive algorithm to dynamically adjust the acquisition parameters, accurately capture operation details and standardization, and achieves autonomous parameter optimization through historical data analysis to improve recognition accuracy in complex environments.

[0052] Real-time feedback and guidance module:

[0053] It receives operation and scene data to generate feedback suggestions and dynamically adjusts the difficulty level; it supports multimodal interaction of voice, annotation, animation and text to achieve real-time guidance response that blends the virtual and real worlds; it intelligently selects the guidance form based on user response efficiency; and the text description adopts a layered display to support switching between operation prompts and principle explanations.

[0054] Comprehensive report generation module:

[0055] It integrates operational data, experimental results, and learning assessments to generate standardized reports, including time-series records, anomaly tracing, and knowledge analysis, to meet the needs of teaching assessment and experimental review; it supports online revision and record keeping, and provides revision suggestions and updates mastery analysis based on knowledge graphs.

[0056] Adaptive learning optimization module:

[0057] A dynamic evaluation system is built to collect accuracy, duration, and report quality indicators, increase operational consistency analysis, optimize training focus to improve learning effectiveness, generate personalized learning paths, push scenario adjustment instructions to the construction module, and provide evaluation data for report generation.

[0058] The dynamic scene intelligent construction module constructs a 3D scene based on the experimental standard process and knowledge graph. This 3D scene supports component interaction, state simulation, and adaptive parameter adjustment. At the same time, the dynamic scene intelligent construction module receives learning progress and difficulty data to achieve linkage and sends the interaction rules to the multimodal operation perception module. The knowledge graph contains principles, specifications, equipment characteristics, and safety knowledge, supports staged loading and dangerous operation warnings, and can push knowledge nodes in real time related to the experimental stage.

[0059] The multimodal operation perception module receives interaction rules from the dynamic scene intelligent construction module, integrates computer vision and sensor technology to collect action data, and identifies operation sequences and predicts intentions based on a spatiotemporal sequence model. At the same time, it adopts an environmentally adaptive algorithm to dynamically adjust the acquisition parameters to accurately capture operation details and standardization, and achieves autonomous parameter optimization through historical data analysis to improve the recognition accuracy in complex environments.

[0060] The real-time feedback guidance module receives operation data from the multimodal operation perception module and scene data from the dynamic scene intelligent construction module, generates feedback suggestions, and dynamically adjusts the difficulty level. This module supports multimodal interaction of voice, annotation, animation, and text, realizing real-time guidance response that blends the virtual and real worlds. It also intelligently selects the guidance format based on user response efficiency, and the text description adopts a layered display that supports switching between operation prompts and principle explanations.

[0061] The comprehensive report generation module integrates operational data from the multimodal operation perception module, experimental results related to the dynamic scene intelligent construction module, and learning evaluation from the adaptive learning optimization module to generate a standardized report. This report includes time-series records, anomaly tracing, and knowledge analysis, meeting the needs of teaching evaluation and experiment review. At the same time, this module supports online revision and trace recording, providing revision suggestions based on knowledge graphs and updating mastery analysis.

[0062] The adaptive learning optimization module constructs a dynamic evaluation system, collects accuracy, duration, and report quality indicators, adds operational coherence analysis, optimizes training focus to improve learning effectiveness, generates personalized learning paths, pushes scene adjustment instructions to the dynamic scene intelligent construction module, and provides evaluation data for the comprehensive report generation module.

[0063] The dynamic scene intelligent construction module builds 3D scenes based on experimental standard procedures and knowledge graphs. It supports component interaction, state simulation and parameter adaptive adjustment, and can receive learning progress and difficulty data to achieve linkage. It can also send interaction rules to the perception module. At the same time, the knowledge graph can push knowledge nodes in real time related to the experimental stage. This is conducive to building virtual experimental scenes that fit the experimental needs of learners and can be dynamically adjusted, so that learners can be better immersed in the experimental environment, acquire knowledge related to the experimental stage in a timely manner, and lay a good foundation for the smooth conduct of the experiment.

[0064] The multimodal operation perception module integrates computer vision and sensor technology to collect action data, identifies operation sequences and predicts intentions based on a spatiotemporal sequence model, dynamically adjusts the acquisition parameters using an environmentally adaptive algorithm, accurately captures operation details and standardization, and achieves autonomous parameter optimization through historical data analysis, thereby improving the recognition accuracy in complex environments. It can accurately and comprehensively perceive the learner's operation behavior, providing reliable operation data for subsequent feedback guidance, report generation, and learning optimization.

[0065] The real-time feedback guidance module receives operation and scenario data to generate feedback suggestions, dynamically adjusts the difficulty level, supports multimodal interaction to achieve instant guidance response, intelligently selects guidance form based on user response efficiency, and uses layered display of text descriptions to support switching between operation prompts and principle explanations. It can guide learners' operations in a timely and effective manner, help learners correct errors in a timely manner, understand experimental principles, and improve the pertinence and efficiency of learning.

[0066] The comprehensive report generation module integrates operational data, experimental results, and learning assessments to generate standardized reports containing various analytical contents, supporting online revision and record keeping, providing revision suggestions based on knowledge graphs and updating mastery analysis, meeting the needs of teaching assessment and experimental review, facilitating teachers to comprehensively assess learners' learning, and also facilitating learners to review and summarize the experimental process, thus promoting the improvement of learning outcomes;

[0067] The adaptive learning optimization module constructs a dynamic evaluation system, collects multiple indicators and adds operational coherence analysis, generates personalized learning paths, pushes scenario adjustment instructions to the construction module, provides evaluation data for report generation, and can perform personalized learning optimization according to the learner's specific situation, making the learner's training more targeted, effectively improving learning outcomes, and achieving personalized learning goals.

[0068] Through the collaborative work of its various modules, the overall system forms a complete, closed-loop interactive virtual experiment teaching system for biology. From scene construction, operation perception, feedback guidance, report generation to learning optimization, each link is closely connected and cooperates with each other, comprehensively improving the quality and efficiency of virtual experiment teaching in biology and providing learners with a high-quality virtual experiment learning experience.

[0069] The adaptive learning optimization module generates standardized reports based on the comprehensive report generation module, analyzes the user's weaknesses in the continuity of experimental operations, generates targeted operation continuity training tasks, and pushes them to the dynamic scene intelligent construction module to build specialized training scenarios.

[0070] The adaptive learning optimization module receives the standardized report generated by the comprehensive report generation module, analyzes the weak links in the user's experimental operation coherence in the report, generates targeted operation coherence training tasks based on the analysis results, and pushes the training tasks to the dynamic scene intelligent construction module, which then constructs a special training scene.

[0071] It can accurately pinpoint users' shortcomings in the continuity of experimental operations, and by generating specialized training scenarios, it can provide users with targeted training, which helps to improve the continuity of users' experimental operations.

[0072] The dynamic scene intelligent construction module generates experimental equipment components with physical collision detection function. When a collision occurs, it calculates the physical feedback effect based on the collision force and angle, and updates the component status and scene display in real time.

[0073] The dynamic scene intelligent construction module generates experimental equipment components with physical collision detection function. When the experimental equipment components collide, the module calculates the physical feedback effect based on the collision force and angle, and updates the status of the experimental equipment components and the scene display in real time.

[0074] This makes the collision feedback in the experimental scenario more realistic, improving the realism of the scenario. At the same time, the real-time updates of status and display allow users to understand the impact of their operations in a timely manner, enhancing the user experience.

[0075] The safety knowledge-related data in the knowledge graph of the dynamic scene intelligent construction module is automatically triggered based on the experimental equipment and operation steps in the scene, and safety warning information is pushed to the real-time feedback guidance module in advance when the user approaches a dangerous operation.

[0076] The knowledge graph of the dynamic scene intelligent construction module contains safety knowledge-related data. In the constructed scene, when the experimental equipment and operation steps in the scene are identified, the safety knowledge-related data is automatically triggered. When the user approaches a dangerous operation, the safety warning information is pushed to the real-time feedback and guidance module in advance.

[0077] It can issue timely safety warnings before users approach dangerous operations, avoid danger in advance, improve the safety of experimental operations, and reduce the occurrence of dangerous operations.

[0078] The environment adaptability algorithm of the multimodal operation perception module dynamically adjusts the acquisition parameters of computer vision and sensors by identifying the lighting and background interference of the operation environment, thereby improving the accuracy of operation action data acquisition in complex environments.

[0079] The environmental adaptability algorithm of the multimodal operation perception module identifies the lighting and background interference of the operation environment, and dynamically adjusts the acquisition parameters of computer vision and sensors based on the recognition results to improve the accuracy of operation action data acquisition in complex environments.

[0080] It can adapt to different complex operating environments and effectively improve the accuracy of data collection in complex environments by dynamically adjusting the acquisition parameters, thus ensuring data reliability.

[0081] The environmental adaptability algorithm of the multimodal operation sensing module establishes an environmental parameter optimization model by analyzing the correlation between historical environmental data and acquisition results, thereby enabling autonomous iterative adjustment of acquisition parameters.

[0082] The environmental adaptability algorithm of the multimodal operation sensing module analyzes the correlation between historical environmental data and acquisition results, establishes an environmental parameter optimization model through analysis, and uses this model to achieve autonomous iterative adjustment of acquisition parameters;

[0083] This allows the collected parameters to be continuously and autonomously optimized based on historical data without manual intervention, thereby continuously improving the rationality of the collected parameters and enhancing the effectiveness of data collection in different environments.

[0084] When constructing a 3D scene based on experimental standard procedures and knowledge graphs, the dynamic scene intelligent construction module generates experimental equipment components with operation guidance labels for different experimental stages. The operation guidance labels dynamically adjust their display status and position as the experiment progresses, guiding users to complete the experimental operations.

[0085] When constructing a 3D scene based on experimental standard procedures and knowledge graphs, the dynamic scene intelligent construction module generates experimental equipment components with operation guidance icons for different experimental stages. The operation guidance icons dynamically adjust their display status and position as the experiment progresses to guide users to complete the experimental operations.

[0086] By dynamically adjusting the operation guidance labels, clear operation instructions can be provided to users at different stages of the experiment, helping users to accurately grasp the order and key points of the experimental operation and successfully complete the experiment.

[0087] When recognizing operation sequences and predicting intentions, the multimodal operation perception module performs temporal analysis on the force, angle, and speed characteristics of the operation actions, generates an operation fluency score based on a deep learning model, and feeds the score results back to the real-time feedback guidance module in real time.

[0088] In the process of recognizing operation sequences and predicting intentions, the multimodal operation perception module performs temporal analysis on the force, angle and speed characteristics of the operation actions, generates an operation fluency score based on a deep learning model, and feeds the score result back to the real-time feedback guidance module in real time.

[0089] It can quantitatively evaluate the smoothness of user operations and provide real-time feedback to users, helping them understand the smoothness of their operations and making targeted improvements to enhance the smoothness of their operations.

[0090] The real-time feedback guidance module provides step-by-step correction guidance for users' non-standard operations based on the analysis results of the multimodal operation perception module. It also demonstrates the standard operation process through animation and pushes relevant experimental principle explanations in conjunction with the knowledge graph.

[0091] The real-time feedback guidance module receives the analysis results from the multimodal operation perception module, provides step-by-step correction guidance for users' non-standard operations, demonstrates standard operating procedures through animation, and pushes relevant experimental principle explanations in conjunction with knowledge graphs.

[0092] It provides specific and intuitive corrective guidance for users' non-standard operations. The combination of animated demonstrations and explanations of principles helps users quickly understand and master standard operations, improving the standardization of operations and understanding of experimental principles.

[0093] The online revision function in the comprehensive report generation module includes revision suggestion push. Based on the experimental knowledge graph and standard operating procedures, it judges the rationality of the user's revisions and provides reference opinions.

[0094] The online revision function in the comprehensive report generation module includes revision suggestion push. The module judges the rationality of the user's revision content based on the experimental knowledge graph and standard operating procedures, and provides the user with reference opinions.

[0095] It helps users judge the rationality of revisions and provides professional reference opinions, which helps improve the accuracy and rationality of user revisions and enhance the quality of comprehensive reports.

[0096] This solution: The dynamic scene intelligent construction module constructs a 3D scene based on experimental standard procedures and a knowledge graph. This 3D scene supports component interaction, state simulation, and adaptive parameter adjustment. Simultaneously, the dynamic scene intelligent construction module receives learning progress and difficulty data to achieve linkage, and sends interaction rules to the multimodal operation perception module. The knowledge graph contains principles, specifications, equipment characteristics, and safety knowledge, supports phased loading and dangerous operation warnings, and can push knowledge nodes in real time related to experimental stages. The dynamic scene intelligent construction module generates experimental equipment components with physical collision detection capabilities. When an experimental equipment component collides, the module calculates the impact force and angle. The system provides feedback on the experimental equipment and updates the status and scene display of the experimental equipment components in real time. When constructing a 3D scene based on the experimental standard process and knowledge graph, the dynamic scene intelligent construction module generates experimental equipment components with operation guidance icons for different experimental stages. The operation guidance icons dynamically adjust their display status and position as the experiment progresses to guide the user to complete the experimental operation. The knowledge graph of the dynamic scene intelligent construction module contains safety knowledge-related data. In the constructed scene, when experimental equipment and operation steps are identified, the safety knowledge-related data is automatically triggered. When the user approaches a dangerous operation, safety warning information is pushed to the real-time feedback guidance module in advance.

[0097] The multimodal operation perception module receives interaction rules from the dynamic scene intelligent construction module, integrates computer vision and sensor technologies to collect action data, and identifies operation sequences and predicts intentions based on a spatiotemporal sequence model. Simultaneously, it employs an environmentally adaptive algorithm to dynamically adjust acquisition parameters to accurately capture operation details and standardization, and achieves autonomous parameter optimization through historical data analysis, improving recognition accuracy in complex environments. The multimodal operation perception module's environmentally adaptive algorithm identifies lighting and background interference in the operation environment, dynamically adjusting the acquisition parameters of computer vision and sensors based on the identification results to improve the accuracy of action data acquisition in complex environments. The algorithm also analyzes the correlation between historical environmental data and acquisition results, establishing an environmental parameter optimization model to achieve autonomous iterative adjustment of acquisition parameters. During the process of identifying operation sequences and predicting intentions, the multimodal operation perception module performs temporal analysis on the force, angle, and speed characteristics of the operation actions, generates an operation fluency score based on a deep learning model, and feeds this score back to the real-time feedback guidance module in real time.

[0098] The real-time feedback guidance module receives operation data from the multimodal operation perception module and scene data from the dynamic scene intelligent construction module, generates feedback suggestions, and dynamically adjusts the difficulty level. This module supports multimodal interaction via voice, annotation, animation, and text, achieving real-time guidance response that blends virtual and real elements. It intelligently selects the guidance format based on user response efficiency, and the text descriptions are displayed in a layered manner, supporting switching between operation prompts and principle explanations. The real-time feedback guidance module receives analysis results from the multimodal operation perception module, provides step-by-step correction guidance for users' non-standard operations, demonstrates standard operating procedures through animation, and pushes relevant experimental principle explanations using a knowledge graph. The real-time feedback guidance module also receives advance safety warning information pushed by the dynamic scene intelligent construction module.

[0099] The comprehensive report generation module integrates operational data from the multimodal operation perception module, experimental results related to the dynamic scene intelligent construction module, and learning evaluation from the adaptive learning optimization module to generate a standardized report. This report includes time-series records, anomaly tracing, and knowledge analysis, meeting the needs of teaching evaluation and experimental review. Simultaneously, this module supports online revision and record keeping, providing revision suggestions based on a knowledge graph and updating mastery analysis. The online revision function in the comprehensive report generation module includes revision suggestion push notifications. Based on the experimental knowledge graph and standard operating procedures, the module judges the rationality of the user's revisions and provides reference opinions to the user.

[0100] The adaptive learning optimization module constructs a dynamic evaluation system, collects accuracy, duration, and report quality indicators, adds operational coherence analysis, optimizes training focus to improve learning effectiveness, generates personalized learning paths, pushes scene adjustment instructions to the dynamic scene intelligent construction module, and provides evaluation data for the comprehensive report generation module. The adaptive learning optimization module receives standardized reports generated by the comprehensive report generation module, analyzes the user's weaknesses in experimental operational coherence in the reports, generates targeted operational coherence training tasks based on the analysis results, and pushes the training tasks to the dynamic scene intelligent construction module, which then constructs specialized training scenarios.

[0101] The overall system forms a complete, closed-loop interactive virtual experiment teaching system for biology through the collaborative work of the dynamic scene intelligent construction module, multimodal operation perception module, real-time feedback guidance module, comprehensive report generation module, and adaptive learning optimization module. From scene construction, operation perception, feedback guidance, report generation to learning optimization, each link is closely connected and cooperates with each other to comprehensively improve the quality and efficiency of virtual experiment teaching in biology and provide learners with a high-quality virtual experiment learning experience.

[0102] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0103] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An interactive virtual experimental teaching system for biology based on deep learning, characterized in that, include: Dynamic Scene Intelligent Construction Module: A 3D scene is constructed based on experimental standard procedures and knowledge graphs, supporting component interaction, state simulation and adaptive parameter adjustment; It receives learning progress and difficulty data to achieve linkage and sends interaction rules to the perception module; the knowledge graph contains principles, specifications, equipment characteristics and safety knowledge, supports staged loading and dangerous operation warnings, and pushes knowledge nodes in real time related to the experimental stage; Multimodal operation sensing module: It integrates computer vision and sensor technologies to collect motion data, identifies operation sequences and predicts intentions based on a spatiotemporal sequence model, and uses an environmentally adaptive algorithm to dynamically adjust the acquisition parameters, accurately capture operation details and standardization, and achieves autonomous parameter optimization through historical data analysis to improve recognition accuracy in complex environments. Real-time feedback and guidance module: Receive operation and scenario data to generate feedback suggestions and dynamically adjust the difficulty level; It supports multimodal interaction including voice, annotation, animation, and text, enabling real-time guidance and response that blends the virtual and real worlds. It intelligently selects the guidance format based on user response efficiency, and the text descriptions are displayed in layers, supporting switching between operation prompts and principle explanations. Comprehensive report generation module: Integrate operational data, experimental results, and learning assessments to generate standardized reports, including time-series records, anomaly tracing, and knowledge analysis, to meet the needs of teaching assessment and experimental review. Supports online revision and record keeping, provides revision suggestions based on knowledge graphs and updates mastery analysis; Adaptive learning optimization module: A dynamic evaluation system is built to collect accuracy, duration, and report quality indicators, increase operational consistency analysis, optimize training focus to improve learning effectiveness, generate personalized learning paths, push scenario adjustment instructions to the construction module, and provide evaluation data for report generation.

2. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The adaptive learning optimization module analyzes the user's weaknesses in experimental operation consistency based on the standardized report generated by the comprehensive report generation module, generates targeted operation consistency training tasks, and pushes them to the dynamic scene intelligent construction module to construct special training scenarios.

3. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The experimental equipment components with physical collision detection function generated by the dynamic scene intelligent construction module calculate the physical feedback effect based on the collision force and angle when a collision occurs, and update the component status and scene display in real time.

4. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The safety knowledge association data in the knowledge graph of the dynamic scene intelligent construction module is automatically triggered based on the experimental equipment and operation steps in the scene, and safety warning information is pushed to the real-time feedback guidance module in advance when the user approaches a dangerous operation.

5. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The environmental adaptability algorithm of the multimodal operation perception module dynamically adjusts the acquisition parameters of computer vision and sensors by identifying the lighting and background interference of the operation environment, thereby improving the accuracy of operation action data acquisition in complex environments.

6. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The environmental adaptability algorithm of the multimodal operation sensing module establishes an environmental parameter optimization model by analyzing the correlation between historical environmental data and acquisition results, thereby enabling autonomous iterative adjustment of acquisition parameters.

7. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: When the dynamic scene intelligent construction module constructs a 3D scene based on the experimental standard process and knowledge graph, it generates experimental equipment components with operation guidance icons for different experimental stages. The operation guidance icons dynamically adjust their display status and position as the experiment progresses to guide the user to complete the experimental operation.

8. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: When recognizing operation sequences and predicting intentions, the multimodal operation perception module performs temporal analysis on the force, angle, and speed characteristics of the operation actions, generates an operation fluency score based on a deep learning model, and feeds the score results back to the real-time feedback guidance module in real time.

9. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The real-time feedback guidance module provides step-by-step correction guidance for users' non-standard operations based on the analysis results of the multimodal operation perception module, and demonstrates the standard operation process through animation, while also pushing relevant experimental principle explanations in conjunction with the knowledge graph.

10. The interactive biological virtual experiment teaching system based on deep learning according to claim 1, characterized in that: The online revision function in the comprehensive report generation module includes revision suggestion push, which judges the rationality of the user's revisions and provides reference opinions based on the experimental knowledge graph and standard operating procedures.

Images