Multifunctional safety helmet management method and system based on work type classification

Abstract

The invention provides a multifunctional safety helmet management method and system based on work type classification, and the method comprises the following steps: carrying out the correlation of a plurality of sensing modules and work type information based on a time window according to the work type classification, and obtaining a work type map; acquiring a face image of a constructor, and performing work type matching based on the face matching square values of the partitions; activating a corresponding sensing module to enter a standby state according to the optimized activation strategy, and automatically granting authority; dividing the sampling reference points into a plurality of construction sub-regions, and binding corresponding sensing modules to obtain a construction region map based on a network map; and when a designated construction area is identified, a construction area map is called, a sensing module is started to perform real-time data acquisition, data optimization is performed through a smooth threshold function, corresponding standard data is called based on the map, comparison is performed based on a time sequence, and abnormal personnel and the area are rapidly positioned. According to the invention, real-time monitoring, operation behavior analysis, environmental safety monitoring and authority management of constructors are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of construction safety management, and in particular to a multifunctional safety helmet management method and system based on work type classification. Background Art

[0002] With the rapid development of industries like construction, mining, and power generation, construction site safety is receiving increasing attention. In these industries, the safety of construction workers is crucial to ensuring the smooth progress of projects. However, traditional safety management methods are no longer sufficient to meet the demands of increasingly complex construction environments and the ever-changing nature of work. Traditional safety management relies heavily on manual inspections, safety officer supervision, and the use of basic personal protective equipment (PPE). These methods are not only inefficient but also present significant safety risks and management loopholes, making them unable to cope with the rapidly changing construction site environment and workforce dynamics.

[0003] During the construction process, workers in different trades face varying working environments and potential safety risks. For example, electricians face the dangers of high voltage electricity, welders need to guard against high-temperature burns, and those working at heights must remain vigilant to the risk of falling. Currently, while hard hats on construction sites provide basic protection, their functionality is limited and cannot meet the needs of more targeted safety protection. Furthermore, the environmental conditions on construction sites are complex, with multiple potential risk factors such as air pollution, noise, and temperature. These factors require hard hats to be able to sense and provide feedback on relevant data in a timely manner so that workers can take timely countermeasures.

[0004] At present, with the advancement of science and technology, intelligent sensing technology is gradually being applied to the field of safety management, especially in the field of personnel safety protection. As an innovative product in this field, the multifunctional safety helmet has the capabilities of embedded sensors, data collection and real-time monitoring. However, the existing multifunctional safety helmet system mainly focuses on individual aspects of monitoring and lacks the ability to accurately identify and manage workers of different types of work. In addition, most systems are unable to dynamically adjust the working status of sensors in real time according to the specific needs of the work tasks and the workers' environment. Furthermore, the current intelligent safety helmet management system can usually only provide a single safety protection function and lacks the ability to effectively integrate and collaborate with environmental information, personnel behavior and other data on the construction site, making it difficult for construction workers to obtain comprehensive safety protection in complex construction environments.

[0005] Therefore, there is an urgent need for an intelligent, systematic and highly flexible safety helmet management method that can be customized according to the needs of different types of work on the construction site, and monitor the health status, behavioral performance and environmental safety conditions of construction workers in real time, so as to effectively prevent various potential dangers and ensure the safety of construction workers. Summary of the Invention

[0006] In order to overcome the shortcomings of the existing technology, the present invention provides a multifunctional safety helmet management method and system based on work type classification. By combining the work requirements of different work types with multiple sensing technologies, real-time monitoring of construction personnel, work behavior analysis, environmental safety monitoring and authority management can be achieved.

[0007] In order to solve the above problems, the technical solutions adopted by the present invention are as follows: A multifunctional safety helmet management method based on work type classification includes the following steps: Provide multifunctional helmets with multiple built-in sensor modules based on job classification, and associate them with job information based on time windows to obtain a job type map; When wearing a hard hat, the construction worker's facial image is collected and the job type is matched based on the partitioned face matching square value; Based on the successfully matched job types and job type maps, the corresponding sensor modules are activated into standby mode with an optimized activation strategy, and permissions are automatically granted; The construction area is divided into several sub-areas based on the sampling reference points and bound to the corresponding sensor modules to obtain a construction area map based on the network graph; When entering the designated construction area, the construction area map is retrieved to start the sensor module for real-time data collection, and data optimization is performed through the smoothing threshold function. Based on the map, the corresponding standard data is retrieved and compared based on the time series to quickly locate abnormal personnel and areas.

[0008] As a preferred embodiment of the present invention, when performing association based on a time window, it includes: Obtain the set of built-in sensor modules, the type of work to be associated, and the current time window; By obtaining the sensor output vector of each sensor module in the set and the work type output vector of each work type information to be associated, a correlation coefficient is generated, and the window correlation coefficient and window correlation threshold of each work type information to be associated with each sensor module in the set are generated in combination with the current time window; The sensor modules whose window association threshold is greater than a preset value are associated with the corresponding work type information.

[0009] As a preferred embodiment of the present invention, when performing job matching, it includes: Performing histogram equalization preprocessing on the facial image to obtain an enhanced facial image and dividing it into several image partitions; use Operator, by adjusting the size of the pixel block to obtain the histogram features of several image partitions at different scales, and combining them to obtain the multi-scale histogram features of each image partition as a variable; The histogram features of each image partition in the construction worker's face image to be matched in the database are taken as the mean to obtain the face matching square value; When the face matching value is less than the threshold, the match is considered successful, and the pre-stored information is retrieved to complete the job matching.

[0010] As a preferred embodiment of the present invention, when activating the corresponding sensor module with the optimized activation strategy, it includes: Get the moment when the sensor module changes from high power consumption to low power consumption , the moment of transition from low power consumption to high power consumption and the moments of high and low power consumption 、 ; Provide an initial activation frequency, based on the time to , the static power consumption of the sensor module and the power consumption when the sensor module is activated, the total power consumption of the sensor module is obtained, and the activation frequency of the sensor module is adjusted.

[0011] As a preferred embodiment of the present invention, segmentation based on sampling reference points includes: Select an initial sub-area in the designated construction area as the positioning area, divide the positioning area into several grids, and use the center point and each vertex of each grid as sampling reference points; Obtain the Euclidean distance between the sampling reference point and the point to be measured, obtain the point with the smallest Euclidean distance, and determine whether it is the center point of the grid; If yes, shrink the four vertices of the positioning area halfway toward the center point respectively; if no, take the grid where the point with the smallest Euclidean distance is located as the positioning area; Repeat the above steps until the area of ​​the grid where the point with the smallest Euclidean distance is located is smaller than the preset value; All sampling reference points in the final generated grid are used The algorithm selects the point with the smallest Euclidean distance and uses the coordinates of the point to mark the final generated grid as the construction sub-area.

[0012] As a preferred embodiment of the present invention, when obtaining a construction area map based on a network diagram, the method includes: Generate a network diagram for representing each construction sub-area, as shown in Formula 6: (6); Where, For the network diagram, is the set of edges in the network graph, is a set of sampling reference points in the network diagram, each sampling reference point is associated with a corresponding sensor module; The sampling data type and sampling value range of each sampling reference point are obtained, added to the network diagram, and the construction area map is generated.

[0013] As a preferred embodiment of the present invention, when identifying entry into a designated construction area, the method includes: The BDS / GPS dual positioning module embedded in the helmet can be used to obtain the location of construction workers and determine whether they have entered the designated construction area. If so, the construction worker's position is compared with the sub-area marker coordinates to obtain the construction sub-area where the construction worker is located.

[0014] As a preferred embodiment of the present invention, when performing data optimization through a smoothing threshold function, it includes: Decompose the real-time multi-source data through wavelet basis functions to obtain wavelet decomposition coefficients at each scale; Estimate the noise variance in each scale data and obtain the threshold; Will The function is dynamically combined with the threshold function adjustment parameters to generate a smooth threshold function; Filtering is performed through the threshold value and smoothing threshold function to obtain the wavelet decomposition coefficients of each scale; The wavelet decomposition coefficients of each scale are filtered and reconstructed, and the approximate coefficients and detail coefficients are merged to obtain the optimized real-time multi-source data.

[0015] As a preferred embodiment of the present invention, when performing comparison based on time series, it includes: The optimized real-time multi-source data is used to fill missing values ​​using the cubic spline interpolation method; Obtain the time series of the filled real-time multi-source data and the interval time series of the standard data, and generate the sequence distance between the time series and the interval time series; When the sequence distance is greater than a preset value, it is considered that abnormal data exists, and the abnormal data is aligned through the time series.

[0016] A multifunctional safety helmet management system based on work type classification, including: The first map generation unit is used to provide a multifunctional helmet with multiple built-in sensor modules according to the type of work, and associate it with the type of work information based on a time window to obtain a type of work map; Matching unit: When wearing a helmet, the facial image of the construction worker is collected and the job type is matched based on the partitioned face matching square value; Authorization unit: Based on the successfully matched job types and job type maps, it activates the corresponding sensor module into standby state with an optimized activation strategy and automatically grants permissions; The second map generation unit: divides the area into several construction sub-areas based on the sampling reference points and binds the corresponding sensor modules to obtain a construction area map based on the network graph; Data acquisition unit: When entering the designated construction area, it retrieves the construction area map and activates the sensor module for real-time data collection; Data identification unit: used to optimize data through smooth threshold function, retrieve corresponding standard data based on the map, compare based on time series, and quickly locate abnormal people and areas.

[0017] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention combines the operational requirements of different types of work with a variety of sensing technologies to achieve real-time monitoring of construction personnel, operational behavior analysis, environmental safety monitoring, and authority management. This method and system can optimize safety management at the construction site, accurately identify the types of work of construction personnel and match them with their corresponding working environments, ensure that construction personnel perform their work under conditions that meet safety requirements, and improve the real-time and accuracy of construction safety. The implementation of the present invention can effectively reduce safety risks during the construction process, provide personalized safety protection measures, and promptly issue warnings and handle abnormalities when they are discovered.

[0018] (2) The present invention provides an innovative solution for comprehensively improving the efficiency and accuracy of construction site safety management through work type classification, real-time data monitoring and analysis, and dynamic authority management. This method not only customizes the activation strategy of the sensor module according to the needs of different work types, but also enables real-time safety monitoring at the construction site and rapid location and warning of potential safety hazards, thereby ensuring the safety of construction workers.

[0019] (3) The modular design adopted by the present invention makes the system highly adaptable in different construction environments. The sensor modules can be flexibly selected and configured according to actual needs, meeting the requirements of various types of work and construction scenarios, and have good scalability.

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a step diagram of the multifunctional safety helmet management method based on work type classification provided by the present invention; Figure 2 This is a flow chart of the segmentation based on sampling reference points provided by the present invention. DETAILED DESCRIPTION

[0022] The multifunctional helmet management method based on work type classification provided by the present invention is as follows: Figure 1 As shown, the following steps are included: Step S1: Based on the job classification at the construction site, a multifunctional helmet with multiple built-in sensor modules is provided, and each sensor module is associated with the job information based on a time window to obtain a job type map; Step S2: After the construction worker puts on the helmet, the built-in image acquisition module of the helmet captures the construction worker's facial image, and performs job matching based on the partitioned face matching square value; Step S3: Based on the successfully matched work types and work type maps, the corresponding sensor modules are activated to enter the standby state using an optimized activation strategy, and permission to enter the designated construction area is automatically granted; Step S4: Divide different designated construction areas into several construction sub-areas based on sampling reference points, and bind corresponding sensor modules to obtain a construction area map based on the network diagram; Step S5: When a construction worker is identified as entering a designated construction area, a construction area map is retrieved and corresponding sensor modules are activated according to the construction sub-area where the construction worker is located to collect real-time data; Step S6: Optimize the collected real-time multi-source data using a smoothing threshold function, retrieve corresponding standard data based on the work type map and construction area map, and compare them based on the time series; Step S7: When data anomalies occur, the abnormal construction personnel and abnormal construction areas are quickly located based on the work type map and construction area map, and the corresponding abnormal information is pushed; Among them, the work type information includes: work type, work task, accessible construction area and authority to use equipment.

[0023] Specifically, the sensor modules may include: a worker health monitoring sensor module, a construction environment monitoring sensor module, a worker behavior monitoring sensor module, a sensor module dedicated to a specific type of work, and a safety helmet wearing monitoring module.

[0024] Among them, the worker health monitoring sensor module is used to monitor the physiological status of workers and ensure that workers work under safe physical conditions.

[0025] The construction environment monitoring sensor module is used to monitor the safety of the construction environment in which workers are located, including parameters such as air quality and physical environment.

[0026] Worker behavior monitoring sensor modules These sensor modules are mainly used to analyze workers' activity status and behavior and identify potential dangerous actions or abnormal behaviors.

[0027] Specialized sensor modules for specific types of work are functional modules designed to meet the special needs of different types of work, such as: Electromagnetic field sensor: used in electrical work to monitor high-voltage electric fields or current intensity to prevent the risk of electric shock.

[0028] Temperature gradient sensor: For welders and foundry workers, it monitors temperature changes in high-temperature environments to prevent burns or heatstroke.

[0029] Tilt sensor: Suitable for high-altitude work or crane operations, to detect whether workers are at dangerous tilt angles.

[0030] Vibration sensor: used by heavy machinery operators to monitor the vibration frequency and amplitude of the machinery and prevent operational failures.

[0031] The helmet wearing monitoring module is used to ensure the correct wearing and use of helmets: Pressure sensor: Detects whether the helmet is worn correctly (by pressure points on the head).

[0032] Infrared sensor: Identifies whether there is human body temperature inside the helmet to determine whether it is being worn.

[0033] Specifically, this invention integrates multiple sensor modules (such as those for worker health monitoring, construction environment monitoring, and worker behavior monitoring). This not only monitors construction workers' physiological status and environmental safety in real time, but also identifies potentially dangerous behaviors, provides real-time warnings, and prevents accidents. This multi-sensor collaboration significantly enhances construction site safety.

[0034] In the above step S1, when performing association based on a time window, the following steps are included: Obtain the set of built-in sensor modules, the type of work to be associated, and the current time window; By obtaining the sensor output vector of each sensor module in the set and the work type output vector of each work type information to be associated, the correlation coefficient is generated, as shown in Formula 1: (1); Where, is the correlation coefficient, is the number of sensor modules, For the A sensor module, is the number of job types to be associated. For the Types of work information to be associated, For the The sensor output vector of each sensor module, For the The output vector of the type of work to be associated; Based on the correlation coefficient, the window correlation coefficient between each type of work information to be associated and each sensor module in the set in the current time window is generated, as shown in Formula 2: (2); Where, is the size of the current time window, The current time window Types of work information to be associated, is the window correlation coefficient; Based on the window correlation coefficient, the window correlation threshold between each type of work information to be associated and each sensor module in the set in the current time window is generated, as shown in Formula 3: (3); Where, is the window association threshold; The sensor modules whose window association threshold is greater than a preset value are associated with the corresponding work type information.

[0035] In the above step S2, when performing job matching based on the partitioned face matching square value, it includes: Performing histogram equalization preprocessing on the facial image to obtain an enhanced facial image; Divide the enhanced facial image into several image partitions, and use operator, and obtain histogram features of different scales of several image partitions by adjusting the size of the pixel block; The histogram features of different scales of each image partition are combined to obtain the multi-scale histogram features of each image partition; The multi-scale histogram features of each image partition are used as variables, and the histogram features of each image partition in the construction worker face images to be matched in the database are used as the mean to obtain the face matching square value, as shown in Formula 4: (4); Where, For the The face matching value of each image partition, For the The weights of the image partitions, For the Image partition scale histogram features, is the face image of the construction worker to be matched Histogram features of image partitions; When the face matching square value is less than the threshold, the face matching is considered successful, and the corresponding construction personnel's pre-stored information is retrieved to complete the job type matching.

[0036] Specifically, this invention uses job type atlases and facial image recognition technology to identify construction workers in real time and automatically grant them access to designated construction areas based on their job type. This precise matching of identity and job type effectively prevents unauthorized access to dangerous areas, improving construction site safety and management efficiency.

[0037] In the above step S3, when the corresponding sensor module is activated to enter the standby state by using the optimized activation strategy, the following steps are included: Get the moment when the sensor module changes from high power consumption to low power consumption , the moment when the sensor module reaches low power consumption ; Get the moment when the sensor module changes from low power consumption to high power consumption , the moment when the sensor module reaches high power consumption ; An initial activation frequency is provided, and the total power consumption of the sensor module is obtained based on the static power consumption of the sensor module and the power consumption when the sensor module is activated, as shown in Formula 5: (5); Where, is the total power consumption of the sensor module, The standby time. is the initial activation frequency, is the power consumption when the sensor module is activated at the initial activation frequency, is the static power consumption of the sensor module; The activation frequency of the sensor module is dynamically adjusted according to the total power consumption to optimize the duration of the sensor module in the low power state.

[0038] Specifically, the present invention utilizes an intelligent optimization strategy to dynamically adjust the sensor module's operating state based on its power consumption, extending the device's lifespan and reducing energy consumption. In particular, in low-power mode, precise control of the sensor module ensures optimal operation, effectively controlling the device's energy consumption while ensuring safety.

[0039] In the above step S3, when automatically granting permission to enter the designated construction area, it includes: Obtain the corresponding private key based on the successfully matched work type, and perform permission tracking and decryption based on the work type map to obtain permission to enter the designated construction area and use the designated equipment; Among them, the job type map uses The algorithm encrypts the access to designated construction areas and the use of designated equipment based on the type of work.

[0040] In the above step S4, if Figure 2 As shown, segmentation based on sampling reference points includes: Select an initial sub-area in the designated construction area as the positioning area, divide the positioning area into several grids, and use the center point and each vertex of each grid as sampling reference points; Obtain the Euclidean distance between the sampling reference point and the point to be measured, and obtain the point with the minimum Euclidean distance; Determine whether the point with the smallest Euclidean distance is the center point of the grid. If so, shrink the four vertices of the positioning area halfway toward the center point. If not, use the grid where the point with the smallest Euclidean distance is located as the positioning area. Repeat the above steps until the area of ​​the grid where the point with the smallest Euclidean distance is located is smaller than the preset value; All sampling reference points in the final generated grid are used The algorithm selects the point with the smallest Euclidean distance and uses the coordinates of this point to mark the final generated grid as the construction sub-area; The sub-area marker coordinates are used to identify the construction sub-area where the construction personnel are located after entering the designated construction area.

[0041] Specifically, the present invention divides the construction area into multiple sub-areas based on sampling reference points. Combined with the positioning and data collection of sensor modules, it can monitor various safety indicators within the construction area in real time. The real-time location of construction workers can be accurately associated with their sub-area, ensuring the effectiveness of data collection and monitoring.

[0042] In the above step S4, when obtaining the construction area map based on the network diagram, it includes: Generate a network diagram for representing each construction sub-area, as shown in Formula 6: (6); Where, For the network diagram, is the set of edges in the network graph, is a set of sampling reference points in the network diagram, each sampling reference point is associated with a corresponding sensor module; Obtain the sampling data type and sampling value range of each sampling reference point, add it to the network diagram, and generate a construction area map, as shown in Formula 7: (7); Where, For the construction area map, is the number of edges in the network graph, Sampling reference point With sampling reference point The weight of the edge between them, Sampling reference point The out-degree, Sampling reference point The out-degree, Sampling reference point The number of the construction sub-area to which it belongs, Sampling reference point The number of the construction sub-area to which it belongs, Is the judgment coefficient, used to judge the sampling reference point With sampling reference point Whether they belong to the same construction sub-area.

[0043] In the above step S5, when it is detected that the construction personnel have entered the designated construction area, the following steps are performed: The BDS / GPS dual positioning module embedded in the helmet can be used to obtain the location of construction workers and determine whether they have entered the designated construction area. If so, the construction worker's position is compared with the sub-area marker coordinates to obtain the construction sub-area where the construction worker is located.

[0044] In the above step S6, when the optimization process is performed by the smoothing threshold function, it includes: The real-time multi-source data is decomposed by selecting the wavelet basis function according to the data characteristics to obtain the wavelet decomposition coefficients of each scale; Estimate the noise variance in each scale data and obtain the threshold value through the threshold selection method; Will The function is dynamically combined with the threshold function adjustment parameters to generate a smooth threshold function; The wavelet decomposition coefficients of each scale are filtered through the threshold value and the smoothing threshold function to obtain the filtered wavelet decomposition coefficients of each scale; The wavelet decomposition coefficients of each scale are filtered and reconstructed, and the approximate coefficients and detail coefficients are merged to complete the data recovery and obtain the optimized real-time multi-source data; Among them, when the wavelet decomposition coefficients of each scale of filtering are obtained, as shown in Formula 8: (8); (9); Where, After filtering Scale, wavelet coefficients, for Scale, wavelet coefficients, for The output of the function, is the global threshold, Adjust the parameters for the threshold function, is the smoothing parameter, is a natural constant.

[0045] In the above step S6, when the comparison is performed based on the time series, it includes: The optimized real-time multi-source data is used to fill missing values ​​using the cubic spline interpolation method; Get the populated real-time multi-source data time series and the interval time series of standard data , and generate the sequence distance between the time series and the interval time series, as shown in Formula 10 and Formula 11: (10); (11); Where, is the amount of data in the time series, For the Sequence data, For the Group interval time series data, is the sequence distance; When the sequence distance is greater than the preset value, it is considered that there is abnormal data, and the abnormal data is aligned through the time series.

[0046] Specifically, this invention utilizes advanced data optimization techniques such as smoothing threshold functions and wavelet transforms to efficiently filter noise and recover data from real-time multi-source data, ensuring data accuracy. Furthermore, a time series comparison method enables rapid problem location when anomalies are detected, providing timely notification of anomalies and ensuring the safety of construction workers.

[0047] The multifunctional safety helmet management system based on work type classification provided by the present invention includes: a first map generation unit, a matching unit, an authorization unit, a second map generation unit, a data acquisition unit and a data recognition unit.

[0048] The first map generation unit is used to provide a multifunctional safety helmet with multiple built-in sensor modules according to the type of work at the construction site, and associate each sensor module with the type of work information based on a time window to obtain a type of work map.

[0049] Matching unit: When construction workers put on safety helmets, the built-in image acquisition module is used to collect facial images of construction workers, and match them to their jobs based on the partitioned face matching method.

[0050] Authorization unit: Based on the successfully matched types of work and types of work maps, the corresponding sensor modules are activated into standby mode with an optimized activation strategy, and permission to enter the designated construction area is automatically granted.

[0051] The second map generation unit is used to divide different designated construction areas into several construction sub-areas based on sampling reference points, and bind corresponding sensor modules to obtain a construction area map based on the network diagram.

[0052] Data collection unit: When a construction worker is identified as entering a designated construction area, the construction area map is retrieved and the corresponding sensor module is activated according to the construction sub-area where the construction worker is located to collect real-time data.

[0053] Data identification unit: used to optimize the collected real-time multi-source data through a smooth threshold function, retrieve the corresponding standard data based on the work type map and construction area map, and compare it based on the time series; when data anomalies occur, the abnormal construction personnel and abnormal construction areas are quickly located based on the work type map and construction area map, and the corresponding abnormal information is pushed.

[0054] Among them, the work type information includes: work type, work task, accessible construction area and authority to use equipment.

[0055] The above embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and replacements made by technicians in this field on the basis of the present invention fall within the scope of protection required by the present invention.

Claims

1. A multifunctional helmet management method based on work type classification, characterized in that: The following steps are involved: Provide multifunctional helmets with multiple built-in sensor modules based on job classification, and associate them with job information based on time windows to obtain a job type map; When wearing a hard hat, the construction worker's facial image is collected and the job type is matched based on the partitioned face matching square value; Based on the successfully matched job types and job type maps, the corresponding sensor modules are activated into standby mode with an optimized activation strategy, and permissions are automatically granted; The construction area is divided into several sub-areas based on the sampling reference points and bound to the corresponding sensor modules to obtain a construction area map based on the network graph; When entering the designated construction area, the construction area map is retrieved to start the sensor module for real-time data collection, and data optimization is performed through the smoothing threshold function. Based on the map, the corresponding standard data is retrieved and compared based on the time series to quickly locate abnormal personnel and areas.

2. The multifunctional helmet management method based on work type classification according to claim 1 is characterized in that: When performing time window-based association, the following are included: Obtain the set of built-in sensor modules, the type of work to be associated, and the current time window; By obtaining the sensor output vector of each sensor module in the set and the work type output vector of each work type information to be associated, a correlation coefficient is generated, and the window correlation coefficient and window correlation threshold of each work type information to be associated with each sensor module in the set are generated in combination with the current time window; The sensor modules whose window association threshold is greater than a preset value are associated with the corresponding work type information.

3. The multifunctional helmet management method based on work type classification according to claim 1 is characterized in that: When matching jobs, include: Performing histogram equalization preprocessing on the facial image to obtain an enhanced facial image and dividing it into several image partitions; use Operator, by adjusting the size of the pixel block to obtain the histogram features of several image partitions at different scales, and combining them to obtain the multi-scale histogram features of each image partition as a variable; The histogram features of each image partition in the construction worker's face image to be matched in the database are taken as the mean to obtain the face matching square value; When the face matching value is less than the threshold, the match is considered successful, and the pre-stored information is retrieved to complete the job matching.

4. The multifunctional helmet management method based on work type classification according to claim 1 is characterized in that: When activating the corresponding sensor module with the optimized activation strategy, it includes: Get the moment when the sensor module changes from high power consumption to low power consumption , the moment of transition from low power consumption to high power consumption and the moments of high and low power consumption 、 ; Provide an initial activation frequency, based on the time to , the static power consumption of the sensor module and the power consumption when the sensor module is activated, the total power consumption of the sensor module is obtained, and the activation frequency of the sensor module is adjusted.

5. The multifunctional helmet management method based on work type classification according to any one of claims 1 to 4, characterized in that: Segmentation based on sampling reference points, including: Select an initial sub-area in the designated construction area as the positioning area, divide the positioning area into several grids, and use the center point and each vertex of each grid as sampling reference points; Obtain the Euclidean distance between the sampling reference point and the point to be measured, obtain the point with the smallest Euclidean distance, and determine whether it is the center point of the grid; If yes, shrink the four vertices of the positioning area halfway toward the center point respectively; if no, take the grid where the point with the smallest Euclidean distance is located as the positioning area; Repeat the above steps until the area of ​​the grid where the point with the smallest Euclidean distance is located is smaller than the preset value; All sampling reference points in the final generated grid are used The algorithm selects the point with the smallest Euclidean distance and uses the coordinates of the point to mark the final generated grid as the construction sub-area.

6. The multifunctional helmet management method based on work type classification according to claim 5 is characterized in that: When obtaining a construction area map based on a network diagram, it includes: Generate a network diagram for representing each construction sub-area, as shown in Formula 6: （6）； Where, For the network diagram, is the set of edges in the network graph, is a set of sampling reference points in the network diagram, each sampling reference point is associated with a corresponding sensor module; The sampling data type and sampling value range of each sampling reference point are obtained, added to the network diagram, and the construction area map is generated.

7. The multifunctional helmet management method based on work type classification according to claim 6 is characterized in that: When identifying entry into a designated construction area, include: The BDS / GPS dual positioning module embedded in the helmet can be used to obtain the location of construction workers and determine whether they have entered the designated construction area. If so, the construction worker's position is compared with the sub-area marker coordinates to obtain the construction sub-area where the construction worker is located.

8. The multifunctional helmet management method based on work type classification according to any one of claims 1 to 4, characterized in that: When performing data optimization via a smoothing threshold function, this includes: Decompose the real-time multi-source data through wavelet basis functions to obtain wavelet decomposition coefficients at each scale; Estimate the noise variance in each scale data and obtain the threshold; Will The function is dynamically combined with the threshold function adjustment parameters to generate a smooth threshold function; Filtering is performed through the threshold value and smoothing threshold function to obtain the wavelet decomposition coefficients of each scale; The wavelet decomposition coefficients of each scale are filtered and reconstructed, and the approximate coefficients and detail coefficients are merged to obtain the optimized real-time multi-source data.

9. The multifunctional helmet management method based on work type classification according to claim 8 is characterized in that: When comparing based on time series, this includes: The optimized real-time multi-source data is used to fill missing values ​​using the cubic spline interpolation method; Obtain the time series of the filled real-time multi-source data and the interval time series of the standard data, and generate the sequence distance between the time series and the interval time series; When the sequence distance is greater than a preset value, it is considered that abnormal data exists, and the abnormal data is aligned through time series.

10. A multifunctional safety helmet management system based on job classification, characterized in that: include: The first map generation unit is used to provide a multifunctional helmet with multiple built-in sensor modules according to the type of work, and associate it with the type of work information based on a time window to obtain a type of work map; Matching unit: When wearing a helmet, the facial image of the construction worker is collected and the job type is matched based on the partitioned face matching square value; Authorization unit: Based on the successfully matched job types and job type maps, it activates the corresponding sensor modules into standby state with an optimized activation strategy and automatically grants permissions; The second map generation unit: divides the area into several construction sub-areas based on the sampling reference points, binds the corresponding sensor modules, and obtains a construction area map based on the network graph; Data acquisition unit: When entering the designated construction area, it retrieves the construction area map and activates the sensor module for real-time data collection; Data identification unit: used to optimize data through smooth threshold function, retrieve corresponding standard data based on the map, compare based on time series, and quickly locate abnormal people and areas.

Images