Construction site pre-class education supervision system based on AI group image face recognition technology

Abstract

The invention discloses a construction site pre-class education supervision system based on an AI group image face recognition technology, which can intelligently recognize a worker gathering scene and automatically start capturing and recognition functions through a crowd gathering trigger algorithm, remarkably improves the working efficiency, is high in face recognition accuracy, ensures that the identity of a pre-class education worker is correct, effectively avoids substituted signing and missed signing, and improves the working efficiency. Meanwhile, voice and environment data in the education process are collected in real time, comprehensive and accurate information is provided for follow-up supervision and analysis, education content is converted into characters, storage and consulting are facilitated, standardized safety production management archive formation is facilitated, activity records are automatically generated, manual operation complexity and errors are reduced, facial images, environment data and archives are deeply counted and analyzed, and the education efficiency is improved. Potential safety hazards and education missing can be found in time, managers are automatically pre-warned, rapid response is assisted, workers are reminded of supplementary learning or reeducation, safety education is ensured to be complete and standard, and safety awareness and operation skills of the workers are improved.

Description

Technical Field

[0001] The present invention relates to the field of education supervision technology, and in particular to a pre-shift education supervision system for construction sites based on AI group portrait face recognition technology. Background Art

[0002] In the construction industry, safe production is always the top priority, especially on busy construction sites. Ensuring that every worker receives adequate pre-shift safety education is crucial. However, traditional pre-shift education supervision methods face many challenges, such as manual record-keeping prone to errors, inefficient supervision, and difficulty tracking worker participation. These problems not only affect the effectiveness of education, but also increase safety hazards on construction sites. Summary of the Invention

[0003] In view of this, the present invention proposes a pre-shift education and supervision system for construction sites based on AI group portrait face recognition technology, which can solve the shortcomings of traditional construction site safety management methods and traditional pre-shift education and supervision methods, improve the construction company's production safety management level, and reduce the risk of safety accidents.

[0004] The technical solution of the present invention is achieved as follows:

[0005] The construction site pre-shift education and supervision system based on AI group portrait facial recognition technology includes:

[0006] A camera module is used to automatically capture and recognize facial images of workers participating in pre-shift training based on crowd gathering triggering algorithms;

[0007] Data collection module, used to collect voice content and on-site environment data during pre-class education;

[0008] The speech recognition and conversion module is used to convert the educational content of the team leader in the pre-shift education into text form and save it as a safety production management file;

[0009] The safety activity record generation module is used to automatically generate pre-shift safety activity records based on the previous day's project safety inspection records and the team leader's daily education content, and push them to workers for signature confirmation through the mobile mini-program;

[0010] The behavior monitoring and early warning module is used to perform statistical analysis on workers' facial image data, on-site environmental data, and safety production management files. When it detects that workers have not attended pre-shift training on time, the briefing content is missing, or does not comply with safety regulations, it automatically sends an early warning message to the manager and prompts the workers to make up for the lessons or receive re-education.

[0011] As a further optional solution to the construction site pre-shift education supervision system based on AI group portrait facial recognition technology, the crowd gathering-based triggering algorithm automatically captures and recognizes the facial images of workers participating in pre-shift education, specifically including:

[0012] Perform real-time analysis of video streams to detect the location and number of people;

[0013] Calculate the crowd concentration index in the current video frame based on the location and number of the crowd;

[0014] When the crowd concentration index exceeds a preset threshold, a face recognition algorithm is used to capture the worker’s facial image from the video stream;

[0015] The captured facial image is compared with the pre-stored worker facial database to identify the worker's identity information.

[0016] As a further optional solution to the construction site pre-shift education and supervision system based on AI group portrait face recognition technology, the crowd concentration index in the current video frame is calculated based on the position and number of the crowd. The specific calculation formula is as follows:

[0017] PDI = (D / A) × N;

[0018] Among them, PDI represents the crowd concentration index, D represents the area occupied by the crowd, A represents the area of the preset safe area, and N represents the number of people in the detection area.

[0019] As a further optional solution to the construction site pre-shift education supervision system based on AI group portrait face recognition technology, the collection of voice content and on-site environment data during the pre-shift education process specifically includes:

[0020] Using audio acquisition equipment and environmental monitoring sensors, the system collects the voice content of pre-class education and on-site temperature and humidity data.

[0021] The data quality index is calculated based on the collected voice content and the temperature and humidity environmental data on site, and the voice content and environmental data with a quality index greater than a preset threshold are saved in the database.

[0022] As a further optional solution to the construction site pre-shift education and supervision system based on AI group portrait face recognition technology, the data quality index is calculated based on the collected voice content and the on-site temperature and humidity environmental data. The specific calculation formula is as follows:

[0023] DQI=(V / V_max)×W_audio+(E / E_max)×W_env;

[0024] Among them, DQI represents the data quality index, V represents the actual amount of valid voice data collected, V_max represents the preset maximum valid voice data amount threshold, E represents the actual amount of valid environmental data collected, E_max represents the preset maximum valid environmental data amount threshold, W_audio represents the voice data weight coefficient, W_env represents the environmental data weight coefficient, and W_audio+W_env=1.

[0025] As a further option for the construction site pre-shift education and supervision system based on AI group portrait facial recognition technology, the system performs statistical analysis on workers' facial image data, on-site environmental data, and safety production management files. When it is detected that workers have not attended pre-shift education on time, the briefing content is missing, or they do not meet safety regulations, an early warning message is automatically sent to the manager. Specifically, it includes:

[0026] Based on the workers' facial image data, on-site environmental data and production safety management files, the consistency index of facial image data, environmental data and production safety management files is calculated;

[0027] Based on the consistency index of facial image data, environmental data and production safety management files, a comprehensive early warning index is calculated. When the comprehensive early warning index exceeds the preset threshold, the early warning mechanism is triggered and an early warning message is automatically sent to the manager.

[0028] As a further optional solution to the construction site pre-shift education and supervision system based on AI group portrait face recognition technology, the camera module also includes a crowd density detection function, which automatically starts the video recording function when it detects that the crowd density exceeds a preset threshold.

[0029] A construction site pre-shift education and supervision method based on AI group portrait facial recognition technology, specifically including:

[0030] Using cameras, based on a preset crowd gathering trigger algorithm, the system automatically captures and recognizes the facial images of workers participating in pre-shift training.

[0031] Real-time collection of voice content and on-site environmental data during pre-class education;

[0032] Accurately convert the oral education content of the team leader in pre-shift education into text form, and automatically organize and save it as a safety production management file;

[0033] Based on the previous day's project safety inspection records and the team leader's daily education content, a pre-shift safety activity record is intelligently generated and pushed to workers for signature confirmation via a mobile mini-program.

[0034] Statistical analysis is performed on workers' facial image data, on-site environmental data, and safety production management files. When it is detected that workers have not attended pre-shift training on time, the briefing content is missing, or does not comply with safety regulations, an early warning message is automatically sent to the manager, prompting the workers to make up for the lessons or receive re-education.

[0035] A computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method for pre-shift education and supervision at a construction site based on AI group portrait face recognition technology are implemented.

[0036] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned method for pre-shift education and supervision at a construction site based on AI group portrait face recognition technology.

[0037] The beneficial effects of the present invention are as follows: through the crowd gathering trigger algorithm, it can intelligently identify the scene of workers gathering, automatically start the capture and recognition function, without manual intervention, greatly improving work efficiency, and the accuracy of facial image recognition is high, ensuring the correct identity of workers participating in pre-shift education, effectively preventing the occurrence of situations such as proxy signing and missed signing, and collecting voice content and on-site environmental data in the pre-shift education process in real time, providing a comprehensive and accurate information basis for subsequent supervision and analysis, and converting the education content into text form, which is not only convenient for storage and reference, but also helps to form standardized safety production management files, automatically generate pre-shift safety activity records, and reduce It reduces the tediousness and errors of manual operations, improves the accuracy and timeliness of records, and pushes them to workers for signature confirmation through mobile applets, which enhances workers' sense of participation and responsibility, while also ensuring the legality of the records. It conducts in-depth statistical analysis of workers' facial image data, on-site environmental data, and production safety management files, which can timely identify potential safety hazards and education gaps, and automatically send early warning information to managers, helping managers to take quick measures to prevent accidents, prompting workers to make up for the lessons or re-education, ensuring that every worker can receive complete and standardized safety education, and improving workers' safety awareness and operational skills. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0039] Figure 1This is a schematic diagram of the composition of a construction site pre-shift education and supervision system based on AI group portrait face recognition technology in the present invention;

[0040] Figure 2 This is a flow chart of a method for pre-shift education and supervision at a construction site based on AI group portrait face recognition technology according to the present invention;

[0041] Figure 3 A schematic diagram of the composition of a computing device according to the present invention. DETAILED DESCRIPTION

[0042] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0043] refer to Figures 1 to 3 The construction site pre-shift education and supervision system based on AI group portrait face recognition technology includes:

[0044] A camera module is used to automatically capture and recognize facial images of workers participating in pre-shift training based on crowd gathering triggering algorithms;

[0045] Data collection module, used to collect voice content and on-site environment data during pre-class education;

[0046] The speech recognition and conversion module is used to convert the educational content of the team leader in the pre-shift education into text form and save it as a safety production management file;

[0047] The safety activity record generation module is used to automatically generate pre-shift safety activity records based on the previous day's project safety inspection records and the team leader's daily education content, and push them to workers for signature confirmation through the mobile mini-program;

[0048] The behavior monitoring and early warning module is used to perform statistical analysis on workers' facial image data, on-site environmental data, and safety production management files. When it detects that workers have not attended pre-shift training on time, the briefing content is missing, or does not comply with safety regulations, it automatically sends an early warning message to the manager and prompts the workers to make up for the lessons or receive re-education.

[0049] In this embodiment, through the crowd gathering trigger algorithm, it can intelligently identify the scene where workers gather, automatically start the capture and recognition function, and no human intervention is required, which greatly improves work efficiency. The accuracy of facial image recognition is high, ensuring the correct identity of workers participating in pre-shift education, effectively preventing the occurrence of situations such as proxy signing and missed signing, and collecting voice content and on-site environmental data in the pre-shift education process in real time, providing a comprehensive and accurate information basis for subsequent supervision and analysis. Converting the education content into text form not only facilitates storage and review, but also helps to form standardized safety production management files, automatically generate pre-shift safety activity records, and reduce It eliminates the tediousness and errors of manual operations, improves the accuracy and timeliness of records, and pushes them to workers for signature confirmation through mobile applets, which enhances workers' sense of participation and responsibility, while also ensuring the legality of the records. It conducts in-depth statistical analysis of workers' facial image data, on-site environmental data, and production safety management files, which can timely discover potential safety hazards and education gaps, and automatically send early warning information to managers, helping managers to take quick measures to prevent accidents, and prompt workers to make up for or re-educate, ensuring that every worker can receive complete and standardized safety education, and improving workers' safety awareness and operational skills.

[0050] Preferably, the method of automatically capturing and identifying facial images of workers participating in pre-shift education based on a crowd gathering triggering algorithm specifically includes:

[0051] Perform real-time analysis of video streams to detect the location and number of people;

[0052] Calculate the crowd concentration index in the current video frame based on the location and number of the crowd;

[0053] When the crowd concentration index exceeds a preset threshold, a face recognition algorithm is used to capture the worker’s facial image from the video stream;

[0054] The captured facial image is compared with the pre-stored worker facial database to identify the worker's identity information.

[0055] In this embodiment, by real-time analysis of the video stream, the dynamic changes of the crowd can be quickly captured, and the position and number of the crowd can be accurately detected, which provides basic data for the subsequent calculation of the crowd concentration, ensuring the response speed and accuracy of the system; the calculation of the crowd concentration index can quantify the degree of crowd concentration, providing a clear basis for triggering the face recognition algorithm, and by setting a reasonable threshold, the face recognition function can be automatically started when the crowd gathers to a certain extent, avoiding unnecessary resource consumption and false triggering; by combining the judgment of the crowd concentration index, the face recognition algorithm can be started at the most appropriate time to ensure that the facial images of workers participating in pre-shift education are captured, thereby improving the accuracy and practicality of recognition; through facial image comparison, the identity information of the workers can be quickly identified, providing basic data for subsequent attendance, management and other work. This step ensures the accuracy and traceability of the workers' identities, and helps to improve the efficiency and safety of construction site management.

[0056] Preferably, the crowd concentration index in the current video frame is calculated based on the position and number of the crowd. The specific calculation formula is as follows:

[0057] PDI = (D / A) × N;

[0058] Among them, PDI represents the crowd concentration index, D represents the area occupied by the crowd, A represents the area of the preset safe area, and N represents the number of people in the detection area.

[0059] In this embodiment, in the pre-shift education scenario at a construction site, the system can automatically capture and identify the facial images of workers participating in the education, and at the same time evaluate the degree of crowd gathering based on the PDI index. When the PDI value is too high, the system can automatically activate the facial image acquisition mechanism, further ensuring the response speed and accuracy of the acquisition.

[0060] Preferably, the collection of voice content and on-site environment data during the pre-class education process specifically includes:

[0061] Using audio acquisition equipment and environmental monitoring sensors, the system collects the voice content of pre-class education and on-site temperature and humidity data.

[0062] The data quality index is calculated based on the collected voice content and the temperature and humidity environmental data on site, and the voice content and environmental data with a quality index greater than a preset threshold are saved in the database.

[0063] In this embodiment, the audio acquisition equipment can fully record the voice content of the pre-shift education process, including the team leader's explanation, workers' questions and feedback, etc. The environmental monitoring sensor is responsible for collecting on-site environmental data such as temperature and humidity. This data helps to understand the actual environmental conditions of the educational activities and is of great significance for evaluating the educational effect, worker comfort and potential safety risks. The introduction of the data quality index as a screening criterion ensures that the data saved in the database has high availability and reliability. The quality index is calculated based on multiple factors, such as speech clarity, background noise level, and accuracy of environmental data. By comparing the quality index with a preset threshold, low-quality data can be automatically eliminated, avoiding misjudgment or misunderstanding caused by inaccurate or ambiguous data, and improving the accuracy and efficiency of data analysis. Only data with a qualified quality index is saved, which reduces the storage burden of the database, improves the speed of data retrieval and processing, and is conducive to establishing a long-term and systematic data archive, making it easier for managers to trace historical education records, analyze trends and changes in education effects, and provide data support for continuous improvement.

[0064] Preferably, the data quality index is calculated based on the collected voice content and the temperature and humidity environment data on site. The specific calculation formula is as follows:

[0065] DQI=(V / V_max)×W_audio+(E / E_max)×W_env;

[0066] Among them, DQI represents the data quality index, V represents the actual amount of valid voice data collected, V_max represents the preset maximum valid voice data amount threshold, E represents the actual amount of valid environmental data collected, E_max represents the preset maximum valid environmental data amount threshold, W_audio represents the voice data weight coefficient, W_env represents the environmental data weight coefficient, and W_audio+W_env=1.

[0067] In this embodiment, by comparing the actual amount of valid data collected (V and E) with the preset maximum valid data amount thresholds (V_max and E_max), and normalizing the results to the range of 0-1, data collected in different batches and at different times are made comparable, which helps to establish a unified data quality assessment standard and improve the standardization of data management; the calculation result of the data quality index (DQI) can be directly used to screen high-quality data. When the DQI value exceeds the preset threshold, the corresponding voice content and environmental data can be saved in the database. This method can ensure that the data saved in the database has high availability and reliability.

[0068] Preferably, the worker's facial image data, on-site environmental data and safety production management files are statistically analyzed, and when it is detected that a worker has not attended pre-shift training on time, the briefing content is missing or does not comply with safety regulations, an early warning message is automatically sent to the manager, specifically including:

[0069] Based on the workers' facial image data, on-site environmental data and production safety management files, the consistency index of facial image data, environmental data and production safety management files is calculated;

[0070] Based on the consistency index of facial image data, environmental data and production safety management files, a comprehensive early warning index is calculated. When the comprehensive early warning index exceeds the preset threshold, the early warning mechanism is triggered and an early warning message is automatically sent to the manager.

[0071] In this embodiment, through the recognition and analysis of facial image data, the identity of the worker can be accurately identified, ensuring that the monitored object is consistent with the worker information in the safety production management file, avoiding monitoring failure due to identity recognition errors, the collection and analysis of on-site environmental data can reflect the safety status of the working environment in real time, and provide an important reference for safe production. The establishment and improvement of the safety production management file provides detailed data support for monitoring, ensuring the comprehensiveness and accuracy of the monitoring content; this technical solution can monitor the workers' facial image data, on-site environmental data and safety production management files in real time. Once an abnormal situation is found (such as the worker fails to attend pre-shift education on time, the briefing content is missing or does not meet safety regulations), the early warning mechanism is immediately triggered and an early warning message is automatically sent to the manager. This real-time early warning method can quickly attract the attention of the manager, take timely measures to solve the problem, and effectively avoid the accumulation of safety hazards and the occurrence of accidents; through statistical analysis of facial image data, on-site environmental data and safety production management files, the manager can fully understand the workers' attendance, work This comprehensive data analysis of the safety status of the working environment and the implementation of safety production management helps managers quickly identify problems, formulate targeted management measures, and improve management efficiency. At the same time, the technical solution can also automatically record early warning information and the implementation of management measures, providing data support for subsequent safety production management. This technical solution promotes the standardization and regularization of safety production management through the monitoring and early warning of facial image data, on-site environmental data, and safety production management files. Workers must attend pre-shift training on time, and the content of the briefing must comply with safety regulations. These requirements are clearly written into the safety production management files and enforced through technical means. This standardized management method helps to enhance workers' safety awareness and reduce the occurrence of safety accidents. Through this technical solution, managers can understand workers' attendance and the safety status of the work environment in real time, so as to rationally arrange work tasks and resource allocation. When safety hazards are found in certain workers or the work environment, managers can promptly adjust work plans to ensure the smooth completion of production tasks and the safety and health of workers.

[0072] It should be noted that the specific formula for calculating the consistency index of facial image data, environmental data and production safety management files is:

[0073] DCI_face=(M_face / T_face)×100%;

[0074] DCI_env=(M_env / T_env)×100%;

[0075] DCI_doc=(M_doc / T_doc)×100%;

[0076] Among them, M_face represents the actual number of worker facial images collected, T_face represents the number of worker facial images that should be collected (based on sign-in records), M_env represents the actual number of valid environmental data collected, T_env represents the number of environmental data that should be collected (based on the preset sampling frequency), M_doc represents the actual number of safety production management files generated, T_doc represents the number of safety production management files that should be generated (based on the pre-shift education plan), DCI_face represents the consistency index of facial image data, DCI_env represents the consistency index of environmental data, and DCI_doc represents the consistency index of safety production management files.

[0077] Based on the consistency index of facial image data, environmental data and production safety management files, the comprehensive early warning index is calculated. The specific formula is:

[0078] CWI=α×(1-DCI_face)+β×(1-DCI_env)+γ×(S_doc);

[0079] Among them, α, β, γ: weight coefficients, and α+β+γ=1;

[0080] S_doc: Severity score of missing or non-compliance with safety regulations in production safety management files (standardized value);

[0081] When DCI_face and DCI_env are lower than the preset threshold, or S_doc is higher than the preset threshold, CWI will increase, triggering the early warning mechanism.

[0082] Preferably, the camera module also includes a crowd density detection function, and automatically starts the video recording function when it is detected that the crowd density exceeds a preset threshold.

[0083] In this embodiment, once the crowd density is detected to be excessive, the camera module will automatically start the video recording function to capture and save the on-site footage. This not only provides valuable visual evidence for subsequent safety analysis and accident investigation, but also reduces the need for manual intervention, ensuring that video data at critical moments can be quickly and accurately obtained.

[0084] A construction site pre-shift education and supervision method based on AI group portrait facial recognition technology, specifically including:

[0085] Using cameras, based on a preset crowd gathering trigger algorithm, the system automatically captures and recognizes the facial images of workers participating in pre-shift training.

[0086] Real-time collection of voice content and on-site environmental data during pre-class education;

[0087] Accurately convert the oral education content of the team leader in pre-shift education into text form, and automatically organize and save it as a safety production management file;

[0088] Based on the previous day's project safety inspection records and the team leader's daily education content, a pre-shift safety activity record is intelligently generated and pushed to workers for signature confirmation via a mobile mini-program.

[0089] Statistical analysis is performed on workers' facial image data, on-site environmental data, and safety production management files. When it is detected that workers have not attended pre-shift training on time, the briefing content is missing, or does not comply with safety regulations, an early warning message is automatically sent to the manager, prompting the workers to make up for the lessons or receive re-education.

[0090] A computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method for pre-shift education and supervision at a construction site based on AI group portrait face recognition technology are implemented.

[0091] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned method for pre-shift education and supervision at a construction site based on AI group portrait face recognition technology.

[0092] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology is characterized by: include: A camera module is used to automatically capture and recognize facial images of workers participating in pre-shift training based on crowd gathering triggering algorithms; Data collection module, used to collect voice content and on-site environment data during pre-class education; The speech recognition and conversion module is used to convert the educational content of the team leader in the pre-shift education into text form and save it as a safety production management file; The safety activity record generation module is used to automatically generate pre-shift safety activity records based on the previous day's project safety inspection records and the team leader's daily education content, and push them to workers for signature confirmation through the mobile mini-program; The behavior monitoring and early warning module is used to perform statistical analysis on workers' facial image data, on-site environmental data, and safety production management files. When it detects that workers have not attended pre-shift training on time, the briefing content is missing, or does not comply with safety regulations, it automatically sends an early warning message to the manager and prompts the workers to make up for the lessons or receive re-education.

2. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology according to claim 1 is characterized in that: The crowd gathering-based triggering algorithm automatically captures and recognizes facial images of workers participating in pre-shift education, specifically including: Perform real-time analysis of video streams to detect the location and number of people; Calculate the crowd concentration index in the current video frame based on the location and number of the crowd; When the crowd concentration index exceeds a preset threshold, a face recognition algorithm is used to capture the worker’s facial image from the video stream; The captured facial image is compared with the pre-stored worker facial database to identify the worker's identity information.

3. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology according to claim 2 is characterized in that: The crowd concentration index in the current video frame is calculated based on the position and number of the crowd. The specific calculation formula is as follows: PDI = (D / A) × N; Among them, PDI represents the crowd concentration index, D represents the area occupied by the crowd, A represents the area of the preset safe area, and N represents the number of people in the detection area.

4. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology according to claim 3 is characterized in that: The collection of voice content and on-site environment data during the pre-class education process specifically includes: Using audio acquisition equipment and environmental monitoring sensors, the system collects the voice content of pre-class education and on-site temperature and humidity data. The data quality index is calculated based on the collected voice content and the temperature and humidity environmental data on site, and the voice content and environmental data with a quality index greater than a preset threshold are saved in the database.

5. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology according to claim 4 is characterized in that: The data quality index is calculated based on the collected voice content and the temperature and humidity environment data on site. The specific calculation formula is as follows: DQI=(V / V_max)×W_audio+(E / E_max)×W_env; Among them, DQI represents the data quality index, V represents the actual amount of valid voice data collected, V_max represents the preset maximum valid voice data amount threshold, E represents the actual amount of valid environmental data collected, E_max represents the preset maximum valid environmental data amount threshold, W_audio represents the voice data weight coefficient, W_env represents the environmental data weight coefficient, and W_audio+W_env=1.

6. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology according to claim 5 is characterized in that: The system performs statistical analysis on workers' facial image data, on-site environmental data, and safety production management files. When it detects that workers have not attended pre-shift training on time, that the briefing content is missing, or that they do not comply with safety regulations, it automatically sends an early warning message to the manager, specifically including: Based on the workers' facial image data, on-site environmental data and production safety management files, the consistency index of facial image data, environmental data and production safety management files is calculated; Based on the consistency index of facial image data, environmental data and production safety management files, a comprehensive early warning index is calculated. When the comprehensive early warning index exceeds the preset threshold, the early warning mechanism is triggered and an early warning message is automatically sent to the manager.

7. The construction site pre-shift education and supervision system based on AI group portrait face recognition technology according to claim 6 is characterized in that: The camera module also includes a crowd density detection function, which automatically starts the video recording function when it detects that the crowd density exceeds a preset threshold.

8. A construction site pre-shift education and supervision method based on AI group portrait face recognition technology, characterized in that: Specifically include: Using cameras, based on a preset crowd gathering trigger algorithm, the system automatically captures and recognizes the facial images of workers participating in pre-shift training. Real-time collection of voice content and on-site environmental data during pre-class education; Accurately convert the oral education content of the team leader in pre-shift education into text form, and automatically organize and save it as a safety production management file; Based on the previous day's project safety inspection records and the team leader's daily education content, a pre-shift safety activity record is intelligently generated and pushed to workers for signature confirmation via a mobile mini-program. Statistical analysis is performed on workers' facial image data, on-site environmental data, and safety production management files. When it is detected that workers have not attended pre-shift training on time, the briefing content is missing, or does not comply with safety regulations, an early warning message is automatically sent to the manager, prompting the workers to make up for the lessons or receive re-education.

9. A computing device, characterized in that It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for pre-shift education and supervision of construction sites based on AI group portrait face recognition technology as described in claim 8 are implemented.

10. A computer-readable storage medium, characterized in that The storage medium stores a computer program, which, when executed by the processor, implements the steps of the construction site pre-shift education and supervision method based on AI group portrait face recognition technology as described in claim 8.

Images