Textile workshop safety management method and system based on Internet of Things

Abstract

The invention is suitable for the field of safety production, and provides a textile workshop safety management method and system based on the Internet of Things, and the method comprises the following steps: obtaining abnormal activity data in a textile workshop through a monitoring device; detecting the state of the employee by using a multi-modal fusion technology based on the abnormal activity data to obtain employee state information; when it is judged that the employee is injured, position information of the employee when the employee is injured is obtained based on the monitoring device; and converting the position information and the employee state information into voice information through an intelligent voice system, and establishing connection with communication terminals of related personnel and hospitals. According to the invention, the injured employees in the textile workshop can be automatically identified, and communication interaction about rescue can be carried out in time, so that the safety accident processing efficiency of the textile workshop is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of safe production, and in particular to a textile workshop safety management method and system based on the Internet of Things. Background Art

[0002] A textile workshop refers to a place used for textile production and processing, where a series of processes from raw material preparation to finished product manufacturing can be completed.

[0003] Modern textile workshops are usually equipped with advanced mechanical equipment, such as automatic spinning machines, looms, dyeing machines, etc. With the advancement of technology, the degree of intelligence and automation is constantly improving, and many workshops have also introduced computer control systems to optimize production processes. Of course, even with the current high degree of automation, human participation is still required, and the mechanical equipment in the textile workshop is mostly based on mechanical actions such as winding, shearing, and cutting, which are highly dangerous, so the safety production regulations of the workshop are very important.

[0004] However, in actual situations, it is inevitable that personnel in the workshop will be injured due to misoperation. In the prior art, when this happens, the follow-up treatment of the injured still needs to be done manually, such as contacting a hospital for rescue, and the processing efficiency cannot be guaranteed. Therefore, a textile workshop safety management method and system based on the Internet of Things is proposed to solve the above problems. Summary of the invention

[0005] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a textile workshop safety management method and system based on the Internet of Things to solve the problems existing in the above-mentioned background technology.

[0006] The present invention is implemented as follows: a textile workshop safety management method based on the Internet of Things, the method comprising the following steps:

[0007] Acquiring abnormal activity data in a textile workshop through a monitoring device, wherein the monitoring device is a device installed in the textile workshop and used to monitor production activities in the workshop;

[0008] Based on abnormal activity data, the multimodal fusion technology is used to detect the employee's status and obtain the employee status information;

[0009] When it is determined that an employee is injured, obtaining the location information of the employee at the time of injury based on the monitoring device;

[0010] The location information and employee status information are converted into voice information through the intelligent voice system and connected with the communication terminals of relevant personnel and the hospital.

[0011] As a further solution of the present invention: the step of obtaining abnormal activity data in the textile workshop by means of a monitoring device specifically comprises:

[0012] Receive video data of the workshop uploaded in real time by a camera device, the resolution and frame rate of which must be greater than a set value;

[0013] Receiving personal data of employees uploaded in real time by sensor equipment, the personal data being individual physiological indicators of employees, the sensor equipment being worn by employees, and the monitoring device including a camera device and a sensor device;

[0014] Build a deep learning model and analyze the employee's motion characteristics based on video data, and judge the employee's physical condition in combination with the acquired personal data to identify abnormal behavior of employees;

[0015] When an employee exhibits abnormal behavior, the monitoring device continuously records the employee's work content to obtain abnormal activity data.

[0016] As a further solution of the present invention: the step of constructing a deep learning model and analyzing the action characteristics of employees based on video data specifically includes:

[0017] The acquired video data is divided into multiple segments, and key points of key actions of employees in the video content are annotated to obtain training data;

[0018] Perform model training on the selected deep learning model based on the training data;

[0019] The real-time uploaded video data is received by a server, and online analysis is performed using a trained deep learning model;

[0020] When an abnormal action is detected, the segment integration time and employee number corresponding to the video data are archived and stored.

[0021] As a further solution of the present invention: the step of detecting the employee's status using multimodal fusion technology based on abnormal activity data to obtain employee status information specifically includes:

[0022] Parsing the abnormal activity data to obtain video sub-data, audio sub-data and individual feature data;

[0023] Perform time series alignment processing on the video sub-data, audio sub-data and individual feature data;

[0024] Detect employees' body movements and facial expressions based on video sub-data, detect abnormal sounds using audio analysis technology based on audio sub-data, and detect employees' heart rates based on individual feature data;

[0025] Multimodal fusion technology is used to integrate the above detection results to evaluate the current status of employees and obtain employee status information.

[0026] As a further solution of the present invention: the step of obtaining the position information of the employee when he is injured based on the monitoring device specifically includes:

[0027] When it is learned through a monitoring device that an employee is injured, multiple monitoring devices adjacent to the monitoring device are called to observe the scene when the employee was injured to obtain multiple observation videos;

[0028] Calculate the actual coordinates of the employee based on multiple observation videos and using triangulation;

[0029] And according to the retrieved workshop plan, the actual coordinates are converted into location information with workshop markings.

[0030] As a further solution of the present invention: the step of converting the location information and employee status information into voice information through the intelligent voice system and establishing a connection with the communication terminal of the relevant personnel and the hospital specifically includes:

[0031] Presetting a template for voice calls, and filling the location information and employee status information into the template;

[0032] Converting the filled template into voice information using text-to-speech technology;

[0033] Selecting the contact numbers of relevant personnel and hospitals from the contact list, wherein the contact list has a plurality of contact numbers preset;

[0034] When the communication is established, the natural language generation technology is used in combination with the voice information to complete the information communication interaction.

[0035] Another object of the present invention is to provide a textile workshop safety management system based on the Internet of Things, the system comprising:

[0036] A data acquisition module, used to acquire abnormal activity data in a textile workshop through a monitoring device, wherein the monitoring device is a device arranged in the textile workshop and used to monitor production activities in the workshop;

[0037] The status detection module detects the status of employees based on abnormal activity data using multimodal fusion technology to obtain employee status information;

[0038] An information acquisition module, for acquiring location information of an employee at the time of injury based on the monitoring device when it is determined that an employee is injured;

[0039] The emergency communication module is used to convert location information and employee status information into voice information through the intelligent voice system and establish connections with relevant personnel and the hospital's communication terminals.

[0040] As a further solution of the present invention: the data acquisition module includes:

[0041] A video transmission unit, used to receive video data of the workshop uploaded in real time by a camera device, the resolution and frame rate of which must be greater than a set value;

[0042] An individual monitoring unit, used to receive personal data of employees uploaded in real time by a sensor device, wherein the personal data is a physiological index of an individual employee. The sensor device is worn by the employee, and the monitoring device includes a camera device and a sensor device;

[0043] The behavior analysis unit builds a deep learning model and analyzes the employee's action characteristics based on video data, and judges the employee's physical condition in combination with the acquired personal data to identify abnormal behavior of employees;

[0044] The record upload unit is used to continuously record the employee's work content through the monitoring device to obtain abnormal activity data when the employee exhibits abnormal behavior.

[0045] As a further solution of the present invention: the state detection module includes:

[0046] A data analysis unit, used for analyzing the abnormal activity data to obtain video sub-data, audio sub-data and individual feature data;

[0047] A data alignment unit, used for performing time series alignment processing on the video sub-data, the audio sub-data and the individual feature data;

[0048] The detection and analysis unit detects the employee's body movements and facial expressions based on the video sub-data, detects abnormal sounds using audio analysis technology based on the audio sub-data, and detects the employee's heart rate based on the individual characteristic data;

[0049] The comprehensive evaluation unit uses multimodal fusion technology to integrate the above detection results to evaluate the current status of the employee and obtain the employee status information.

[0050] As a further solution of the present invention: the emergency communication module includes:

[0051] An information filling unit, used to pre-set a template for a voice call and fill the location information and employee status information into the template;

[0052] A speech conversion unit, which converts the filled template into speech information using text-to-speech technology;

[0053] A communication selection unit, used to select the communication numbers of relevant personnel and hospitals from a communication list, wherein the communication list is preset with multiple communication numbers;

[0054] The intelligent voice interaction unit is used to complete information communication and interaction by using natural language generation technology and combining the voice information after the communication is established.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention will first use the existing video recognition technology to judge the behavior of employees. When the employee behavior is abnormal, the employee will be paid special attention to obtain the abnormal activity data and use multimodal fusion technology to judge the employee's status, because it can more accurately judge whether the employee is injured. More importantly, when an employee is injured, the intelligent voice system can be used to establish communication with relevant management, security personnel and hospitals, so that the situation can be fed back in time, which not only enables the company's management and security personnel to arrive at the scene in time, but also can greatly save the rescue time for injured employees. In summary, the present invention can automatically identify injured employees in a textile workshop, and can communicate and interact about rescue in time, thereby effectively improving the efficiency of the textile workshop in handling safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 The figure is a flow chart of a textile workshop safety management method based on the Internet of Things.

[0058] Figure 2 The present invention is a flow chart of obtaining abnormal activity data in a textile workshop through a monitoring device in a textile workshop safety management method based on the Internet of Things.

[0059] Figure 3 A flowchart for building a deep learning model in a textile workshop safety management method based on the Internet of Things and analyzing employees' motion characteristics based on video data.

[0060] Figure 4 The present invention is a flowchart of a textile workshop safety management method based on the Internet of Things, which detects the status of employees based on abnormal activity data using multimodal fusion technology to obtain employee status information.

[0061] Figure 5 The present invention is a flow chart of obtaining the position information of an employee when he / she is injured based on the monitoring device in a textile workshop safety management method based on the Internet of Things.

[0062] Figure 6The present invention is a flowchart of a textile workshop safety management method based on the Internet of Things, in which location information and employee status information are converted into voice information through an intelligent voice system and a connection is established with communication terminals of relevant personnel and hospitals.

[0063] Figure 7 This is a structural diagram of a textile workshop safety management system based on the Internet of Things.

[0064] Figure 8 This is a structural diagram of a data acquisition module in a textile workshop safety management system based on the Internet of Things.

[0065] Fig. 9 This is a structural diagram of a status detection module in a textile workshop safety management system based on the Internet of Things.

[0066] Fig.10 This is a structural diagram of an emergency communication module in a textile workshop safety management system based on the Internet of Things. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0068] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0069] like Figure 1 As shown, an embodiment of the present invention provides a textile workshop safety management method based on the Internet of Things, and the method includes the following steps:

[0070] S100, obtaining abnormal activity data in a textile workshop through a monitoring device, wherein the monitoring device is a device disposed in the textile workshop and used to monitor production activities in the workshop;

[0071] S200, detecting the employee's status based on the abnormal activity data using multimodal fusion technology to obtain employee status information;

[0072] S300, when it is determined that an employee is injured, obtaining location information of the employee at the time of injury based on the monitoring device;

[0073] S400, through the intelligent voice system, converts location information and employee status information into voice information and establishes a connection with the communication terminals of relevant personnel and the hospital.

[0074] It should be noted that the multimodal fusion technology refers to the combination of various data such as video, audio and sensors to judge the status of employees. For example, when an employee is injured, he or she will have rapidly changing facial expressions, screams, specific body movements, changes in heart rate and other performance characteristics. These characteristics can all be recorded by the monitoring device. Therefore, the monitoring device does not represent a single device, but represents the cameras, sensors and other equipment in the monitoring workshop. Only when the data is collected comprehensively can the accuracy of the judgment of the employee's status be guaranteed.

[0075] In the embodiment of the present invention, the present invention will first use the existing video recognition technology to judge the behavior of employees. When the employee behavior is abnormal, the employee will be paid special attention to obtain the abnormal activity data, and the multimodal fusion technology will be used to judge the status of the employee, because it can more accurately judge whether the employee is injured. More importantly, when an employee is injured, the intelligent voice system can be used to establish communication with relevant management, security personnel and hospitals, so that the situation can be fed back in time, which not only enables the company's management and security personnel to arrive at the scene in time, but also can greatly save the rescue time for injured employees. In summary, the present invention can automatically identify injured employees in a textile workshop, and then can timely communicate and interact about rescue, thereby effectively improving the efficiency of the textile workshop in handling safety accidents.

[0076] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of obtaining abnormal activity data in a textile workshop by a monitoring device specifically includes:

[0077] S101, receiving video data of the workshop uploaded in real time by a camera device, wherein the resolution and frame rate of the camera device must be greater than a set value;

[0078] S102, receiving personal data of an employee uploaded in real time by a sensor device, wherein the personal data is an individual physiological index of the employee, the sensor device is worn by the employee, and the monitoring device includes a camera device and a sensor device;

[0079] S103, building a deep learning model and analyzing the employee's motion characteristics based on the video data, and judging the employee's physical condition in combination with the acquired personal data to identify abnormal behavior of the employee;

[0080] S104, when an employee exhibits abnormal behavior, the monitoring device continuously records the employee's work content to obtain abnormal activity data.

[0081] In the embodiment of the present invention, there are certain requirements for the specifications of the camera equipment, so as to ensure the quality of the video, which is more conducive to the subsequent recognition and analysis. After obtaining the video data and personal data about the employee, it is first necessary to identify the employee's action characteristics to determine whether it meets the safety production standards, that is, to identify whether there are some dangerous actions. At the same time, the personal data can be used to determine whether the employee's body is in a state of fatigue. The above aspects are the main causes of safety accidents, and therefore can be used as a basis for judging whether to focus on a certain employee.

[0082] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of building a deep learning model and analyzing the action characteristics of employees based on video data specifically includes:

[0083] S1031, dividing the acquired video data into multiple segments, and annotating key points of key actions of employees in the video content to obtain training data;

[0084] S1032, performing model training on the selected deep learning model based on the training data;

[0085] S1033, receiving the real-time uploaded video data through a server, and performing online analysis using a trained deep learning model;

[0086] S1034, when an abnormal action is detected, the segment integration time and employee number corresponding to the video data are archived and stored.

[0087] In the embodiment of the present invention, when the key points of the employee's actions are marked, the OpenPose real-time multi-person 2D posture estimation model is generally selected, which can capture the key points of the human body, such as the neck, wrist, knee, etc., and then generate a confidence map of the key points through a convolutional neural network (CNN), and then connect the key points through a two-part matching algorithm in graph theory to form a human skeleton. The training data includes not only the coordinates of each key point, but also the coordinate change trend of the key point, as well as the speed and acceleration of the key point. These data will be used for subsequent model training. The real-time uploaded video data only records the employee status, and only after the employee is judged to be injured by the online analysis method, will it be focused on and recorded, so as to facilitate the subsequent further identification and analysis of the specific circumstances of the employee's injury. Of course, in this embodiment, it is not only used to detect whether the employee is injured, but also can judge some dangerous actions or irregular behaviors of the employee, and also has the function of reminding the employee, which plays a preventive role.

[0088] like Figure 4As shown, as a preferred embodiment of the present invention, the step of detecting the employee's status based on abnormal activity data using multimodal fusion technology to obtain employee status information specifically includes:

[0089] S201, parsing the abnormal activity data to obtain video sub-data, audio sub-data and individual feature data;

[0090] S202, performing time series alignment processing on the video sub-data, the audio sub-data and the individual feature data;

[0091] S203, detecting the employee's body movements and facial expressions based on the video sub-data, detecting abnormal sounds using audio analysis technology based on the audio sub-data, and detecting the employee's heart rate based on the individual characteristic data;

[0092] S204, using multimodal fusion technology to combine the above detection results to evaluate the current status of the employee to obtain employee status information.

[0093] In the embodiment of the present invention, although the employee's body movements can be identified through the deep learning model, this cannot fully determine the specific situation of the employee. Therefore, the multimodal fusion technology can be used, covering the three aspects of audio, video and personal signs. The first thing to do is to align the three in time series. For example, when the employee's face is distorted, the wound is covered, and screams at the same time, it can be determined that the employee is injured. If the employee only has specific movements, but is not accompanied by changes in sound and personal signs, it is often just some random movements of the employee, and there is no injury. Of course, the employee status information is generated according to the injured part of the employee, and the injured part can be identified through the employee's movements in the video.

[0094] like Figure 5 As shown, as a preferred embodiment of the present invention, the step of obtaining the position information of the employee when he is injured based on the monitoring device specifically includes:

[0095] S301, when it is learned through a monitoring device that an employee is injured, multiple monitoring devices adjacent to the monitoring device are called to observe the scene when the employee is injured to obtain multiple observation videos;

[0096] S302, calculating the actual coordinates of the employee based on multiple observation videos and using triangulation method;

[0097] S303, and convert the actual coordinates into location information with workshop identification according to the retrieved workshop plan.

[0098] In the embodiment of the present invention, cameras are generally installed at various positions in the textile workshop. These cameras have different viewing angles, and the viewing angles can also be adjusted. The positions and viewing angles of these cameras are known, that is, when a certain employee is identified as injured by a camera, other cameras will be called for synchronous identification. Therefore, the relative positions of the employees and the cameras are known. Of course, the positions are all expressed in the form of coordinates. After calculation using the triangulation formula, the specific and accurate position of the employee can be obtained, and then converted into a position relative to the textile workshop, that is, the position information, so that others can know the position of the injured employee, which is convenient for subsequent rescue operations.

[0099] like Figure 6 As shown, as a preferred embodiment of the present invention, the step of converting the location information and employee status information into voice information through the intelligent voice system and establishing a connection with the communication terminal of the relevant personnel and the hospital specifically includes:

[0100] S401, presetting a template for voice calls, and filling the location information and employee status information into the template;

[0101] S402, converting the filled template into voice information using text-to-speech technology;

[0102] S403, selecting the contact numbers of the relevant personnel and the hospital from the contact list, wherein the contact list has a plurality of contact numbers preset;

[0103] S404, when the communication is established, the natural language generation technology is used in combination with the voice information to complete the information communication interaction.

[0104] In the embodiment of the present invention, it is explained that the situation feedback in the textile workshop is in the form of a template. When the communication is connected, the relevant personnel and the hospital will be able to know the location in the workshop and the approximate situation of the employee's injury, so that the management personnel and security personnel can arrive at the scene in time, and the hospital's rescue operation can be deployed according to the employee's injury situation. In summary, this embodiment uses an intelligent voice assistant to feedback relevant information of employee injuries to other personnel, eliminating the tedious problem of frequent communication and determination between people in the existing situation.

[0105] like Figure 7 As shown, the embodiment of the present invention also provides a textile workshop safety management system based on the Internet of Things, the system comprising:

[0106] The data acquisition module 100 is used to acquire abnormal activity data in the textile workshop through a monitoring device, wherein the monitoring device is a device arranged in the textile workshop and used to monitor production activities in the workshop;

[0107] A status detection module 200 detects the status of employees using multimodal fusion technology based on abnormal activity data to obtain employee status information;

[0108] The information acquisition module 300 is used to acquire the position information of the employee when the employee is injured based on the monitoring device when it is determined that the employee is injured;

[0109] The emergency communication module 400 is used to convert the location information and employee status information into voice information through the intelligent voice system and establish a connection with the communication terminals of relevant personnel and the hospital.

[0110] In the embodiment of the present invention, the present invention will first use the existing video recognition technology to judge the behavior of employees. When the employee behavior is abnormal, the employee will be paid special attention to obtain the abnormal activity data, and the multimodal fusion technology will be used to judge the status of the employee, because it can more accurately judge whether the employee is injured. More importantly, when an employee is injured, the intelligent voice system can be used to establish communication with relevant management, security personnel and hospitals, so that the situation can be fed back in time, which not only enables the company's management and security personnel to arrive at the scene in time, but also can greatly save the rescue time for injured employees. In summary, the present invention can automatically identify injured employees in a textile workshop, and then can timely communicate and interact about rescue, thereby effectively improving the efficiency of the textile workshop in handling safety accidents.

[0111] like Figure 8 As shown, as a preferred embodiment of the present invention, the data acquisition module 100 includes:

[0112] The video transmission unit 101 is used to receive the video data of the workshop uploaded by the camera in real time. The resolution and frame rate of the camera must be greater than the set value.

[0113] The individual monitoring unit 102 is used to receive the personal data of the employee uploaded in real time by the sensor device, wherein the personal data is the physiological index of the individual employee. The sensor device is worn by the employee, and the monitoring device includes a camera device and a sensor device;

[0114] The behavior analysis unit 103 constructs a deep learning model and analyzes the employee's action characteristics based on the video data, and determines the employee's physical state in combination with the acquired personal data to identify abnormal behavior of the employee;

[0115] The record uploading unit 104 is used to continuously record the work content of the employee through the monitoring device to obtain abnormal activity data when the employee exhibits abnormal behavior.

[0116] like Fig. 9As shown, as a preferred embodiment of the present invention, the state detection module 200 includes:

[0117] A data parsing unit 201 is used to parse the abnormal activity data to obtain video sub-data, audio sub-data and individual feature data;

[0118] A data alignment unit 202, used for performing time series alignment processing on the video sub-data, the audio sub-data and the individual feature data;

[0119] The detection and analysis unit 203 detects the employee's body movements and facial expressions based on the video sub-data, detects abnormal sounds using audio analysis technology based on the audio sub-data, and detects the employee's heart rate based on the individual characteristic data;

[0120] The comprehensive evaluation unit 204 uses multimodal fusion technology to comprehensively evaluate the above detection results to obtain employee status information.

[0121] like Fig.10 As shown, as a preferred embodiment of the present invention, the emergency communication module 400 includes:

[0122] An information filling unit 401 is used to pre-set a template for a voice call and fill the location information and employee status information into the template;

[0123] A speech conversion unit 402 converts the filled template into speech information using text-to-speech technology;

[0124] A communication selection unit 403 is used to select the communication numbers of relevant personnel and hospitals in the communication list, wherein the communication list is preset with multiple communication numbers;

[0125] The intelligent voice interaction unit 404 is used to complete information communication and interaction by using natural language generation technology and combining the voice information after the communication is established.

[0126] The above only describes in detail the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0127] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0128] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0129] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A textile workshop safety management method based on the Internet of Things, characterized in that: The method comprises the following steps: Acquiring abnormal activity data in a textile workshop through a monitoring device, wherein the monitoring device is a device installed in the textile workshop and used to monitor production activities in the workshop; Based on abnormal activity data, the multimodal fusion technology is used to detect the employee's status and obtain the employee status information; When it is determined that an employee is injured, obtaining the location information of the employee at the time of injury based on the monitoring device; The location information and employee status information are converted into voice information through the intelligent voice system and connected with the communication terminals of relevant personnel and the hospital.

2. The textile workshop safety management method based on the Internet of Things according to claim 1 is characterized in that: The step of obtaining abnormal activity data in the textile workshop through the monitoring device specifically includes: Receive video data of the workshop uploaded in real time by a camera device, the resolution and frame rate of which must be greater than a set value; Receiving personal data of employees uploaded in real time by sensor equipment, the personal data being individual physiological indicators of employees, the sensor equipment being worn by employees, and the monitoring device including a camera device and a sensor device; Build a deep learning model and analyze the employee's motion characteristics based on video data, and judge the employee's physical condition in combination with the acquired personal data to identify abnormal behavior of employees; When an employee exhibits abnormal behavior, the monitoring device continuously records the employee's work content to obtain abnormal activity data.

3. The textile workshop safety management method based on the Internet of Things according to claim 2 is characterized in that: The steps of building a deep learning model and analyzing the employee's action characteristics based on video data specifically include: The acquired video data is divided into multiple segments, and key points of key actions of employees in the video content are annotated to obtain training data; Perform model training on the selected deep learning model based on the training data; The real-time uploaded video data is received by a server, and online analysis is performed using a trained deep learning model; When an abnormal action is detected, the segment integration time and employee number corresponding to the video data are archived and stored.

4. The textile workshop safety management method based on the Internet of Things according to claim 1 is characterized in that: The step of detecting the employee's status using multimodal fusion technology based on abnormal activity data to obtain employee status information specifically includes: Parsing the abnormal activity data to obtain video sub-data, audio sub-data and individual feature data; Perform time series alignment processing on the video sub-data, audio sub-data and individual feature data; Detect employees' body movements and facial expressions based on video sub-data, detect abnormal sounds using audio analysis technology based on audio sub-data, and detect employees' heart rates based on individual feature data; Multimodal fusion technology is used to integrate the above detection results to evaluate the current status of employees and obtain employee status information.

5. The textile workshop safety management method based on the Internet of Things according to claim 1 is characterized in that: The step of obtaining the position information of the employee when he / she is injured based on the monitoring device specifically includes: When it is learned through a monitoring device that an employee is injured, multiple monitoring devices adjacent to the monitoring device are called to observe the scene when the employee was injured to obtain multiple observation videos; Calculate the actual coordinates of the employee based on multiple observation videos and using triangulation; And according to the retrieved workshop plan, the actual coordinates are converted into location information with workshop markings.

6. The textile workshop safety management method based on the Internet of Things according to claim 1 is characterized in that: The step of converting the location information and employee status information into voice information through the intelligent voice system and establishing a connection with the communication terminal of the relevant personnel and the hospital specifically includes: Presetting a template for voice calls, and filling the location information and employee status information into the template; Converting the filled template into voice information using text-to-speech technology; Selecting the contact numbers of relevant personnel and hospitals from the contact list, wherein the contact list has a plurality of contact numbers preset; When the communication is established, the natural language generation technology is used in combination with the voice information to complete the information communication interaction.

7. A textile workshop safety management system based on the Internet of Things, characterized in that: The system comprises: A data acquisition module, used to acquire abnormal activity data in a textile workshop through a monitoring device, wherein the monitoring device is a device arranged in the textile workshop and used to monitor production activities in the workshop; The status detection module detects the status of employees based on abnormal activity data using multimodal fusion technology to obtain employee status information; An information acquisition module, for acquiring location information of an employee at the time of injury based on the monitoring device when it is determined that an employee is injured; The emergency communication module is used to convert location information and employee status information into voice information through the intelligent voice system and establish connections with relevant personnel and the hospital's communication terminals.

8. The textile workshop safety management system based on the Internet of Things according to claim 7 is characterized in that: The data acquisition module comprises: A video transmission unit, used to receive video data of the workshop uploaded in real time by a camera device, the resolution and frame rate of which must be greater than a set value; An individual monitoring unit, used to receive personal data of employees uploaded in real time by a sensor device, wherein the personal data is a physiological index of an individual employee. The sensor device is worn by the employee, and the monitoring device includes a camera device and a sensor device; The behavior analysis unit builds a deep learning model and analyzes the employee's action characteristics based on video data, and judges the employee's physical condition in combination with the acquired personal data to identify abnormal behavior of employees; The record upload unit is used to continuously record the employee's work content through the monitoring device to obtain abnormal activity data when the employee exhibits abnormal behavior.

9. The textile workshop safety management system based on the Internet of Things according to claim 7 is characterized in that: The state detection module comprises: A data analysis unit, used for analyzing the abnormal activity data to obtain video sub-data, audio sub-data and individual feature data; A data alignment unit, used for performing time series alignment processing on the video sub-data, the audio sub-data and the individual feature data; The detection and analysis unit detects the employee's body movements and facial expressions based on the video sub-data, detects abnormal sounds using audio analysis technology based on the audio sub-data, and detects the employee's heart rate based on the individual characteristic data; The comprehensive evaluation unit uses multimodal fusion technology to integrate the above detection results to evaluate the current status of the employee and obtain the employee status information.

10. The textile workshop safety management system based on the Internet of Things according to claim 7 is characterized in that: The emergency communication module comprises: An information filling unit, used to pre-set a template for a voice call and fill the location information and employee status information into the template; A speech conversion unit, which converts the filled template into speech information using text-to-speech technology; A communication selection unit, used to select the communication numbers of relevant personnel and hospitals from a communication list, wherein the communication list is preset with multiple communication numbers; The intelligent voice interaction unit is used to complete information communication and interaction by using natural language generation technology and combining the voice information after the communication is established.

Images