Intelligent device rack intelligent workshop safety system

Abstract

The application discloses an intelligent workshop safety system of an intelligent equipment rack, comprising a face recognition module, a voice interaction module and an intelligent monitoring module which are sequentially linked, the intelligent monitoring module is connected with an environment monitoring module and a device module to establish a bidirectional signal connection, and is connected with a wireless server through a wireless communication link to realize the whole-process safety management and control of workshop personnel access, function matching, device operation and environment state, solve the problems of inaccurate personnel access, single interaction mode, disconnection of device and environment management and control, fuzzy early warning level, poor data traceability and insufficient system compatibility in the prior art, realize the whole-process intelligent management and control of workshop personnel access, function matching, device operation and environment state through the linkage design of face recognition, voice interaction, intelligent monitoring and environment monitoring, and improve the accuracy, efficiency and reliability of workshop safety management.

Description

Technical Field

[0001] This invention relates to the field of intelligent equipment rack processing technology, and more specifically to an intelligent workshop safety system for intelligent equipment racks. Background Technology

[0002] Intelligent equipment rack production workshops involve high-precision equipment operation and multi-workstation collaborative work, placing extremely high demands on personnel access control, equipment safety, and environmental management. Existing workshop safety systems suffer from significant deficiencies: First, personnel access management is rudimentary, often relying on traditional access cards or password verification, which easily leads to issues like fake attendance and misuse of permissions, and fails to accurately match personnel identities with workstation functions. Second, interaction methods are limited, lacking intelligent voice guidance and permission query functions; operators must manually consult operation manuals or contact management personnel, impacting work efficiency. Third, equipment monitoring and environmental management are disconnected; equipment operation status monitoring focuses only on single parameters, failing to link with environmental factors such as workshop temperature, humidity, and harmful gases, making equipment malfunctions or safety accidents prone to occur due to environmental anomalies. Fourth, warning levels are vague; only a single warning is issued for environmental or equipment anomalies, failing to implement differentiated control measures based on risk levels. Fifth, data traceability is poor, lacking long-term data storage and log recording functions, making it difficult to trace the cause after a safety incident. Sixth, system compatibility is insufficient; each module operates independently, data interaction is inefficient, and a closed-loop management system cannot be formed throughout the entire process. Existing solutions are insufficient to meet the core requirements of "precise access, intelligent interaction, coordinated control, hierarchical early warning, and data traceability," and there is an urgent need for an integrated intelligent workshop safety system. Summary of the Invention

[0003] The technical problem this invention aims to solve is to provide an intelligent workshop safety system for intelligent equipment racks, addressing issues in existing technologies such as inaccurate personnel access control, limited interaction methods, disconnect between equipment and environmental management, ambiguous early warning levels, poor data traceability, and insufficient system compatibility. Through the integrated design of facial recognition, voice interaction, intelligent monitoring, and environmental monitoring, the system achieves intelligent control over the entire process of personnel access, job matching, equipment operation, and environmental status in the workshop, thereby improving the accuracy, efficiency, and reliability of workshop safety management.

[0004] To address the aforementioned technical problems, this invention provides an intelligent equipment rack smart workshop safety system, comprising a face recognition module, a voice interaction module, and an intelligent monitoring module linked in sequence. The intelligent monitoring module establishes bidirectional signal connections with the environmental monitoring module and the equipment module, and interacts with a wireless server via a wireless communication link, realizing full-process safety control of workshop personnel access, job matching, equipment operation, and environmental status. The specific structure and process are as follows: The face recognition module consists of a data acquisition unit, an image preprocessing unit, an image processing unit, a matching unit, and an analysis unit connected in series. The data acquisition unit is a high-definition binocular camera used to acquire real-time images of the user's face and facial depth information. The image preprocessing unit performs noise reduction, grayscale correction, size normalization, and face region cropping on the acquired images. The image processing unit extracts image information, including facial organ size parameters, facial organ skin color features, and coordinates of key local facial feature points. The matching unit compares the extracted image information with pre-stored authorized personnel image information in the system database for similarity. The analysis unit outputs an identity recognition result based on the comparison result; if the similarity is ≥ a preset threshold, authorization is granted; otherwise, an access control interception signal is triggered. Voice interaction module: Automatically starts after facial recognition authorization is successful. Its execution process includes: (1) Multimedia information is acquired synchronously through audio acquisition device and image acquisition device, wherein the multimedia information includes first voice information and real-time scene image; (2) Locate the sound source of the first voice information and calculate the spatial coordinates of the current user using the microphone array; (3) Based on the spatial orientation coordinates, select candidate users in that orientation from the real-time scene image, and combine the identity recognition results output by the face recognition module to lock in the unique current user; (4) Based on the voice command content and identity recognition results in the multimedia information, match the pre-stored work position function database of the system to determine the current user's work position function and the group identifier to which they belong; (5) Invoke the corresponding preset voice interaction strategy according to the group identifier. The voice interaction strategy includes operation guidance script, permission query feedback, risk warning prompt, and execute the voice interaction function. The intelligent monitoring module includes a core control microprocessor. The input of the microprocessor is connected to a detection unit via a signal conditioning circuit. The detection unit comprises a current sensor, a temperature sensor, and a vibration sensor, used to collect real-time operating parameters of the device module. The output of the microprocessor is connected to a switching unit via a drive circuit. The switching unit is a relay or an AC contactor. The switching unit, the device module, and the power supply are connected in series to form a closed loop. The microprocessor controls the switching unit's on / off state based on the data collected by the detection unit. The microprocessor also connects to a wireless server via a detection wireless communication unit. This detection wireless communication unit uses LoRa or 5G communication protocols to upload device operating data and receive remote control commands. The intelligent monitoring module is connected to the environmental monitoring module and controls it. The environmental monitoring module is used to monitor the workshop environment in real time and feed it back to the microprocessor, forming a linkage control between equipment operation and environmental status.

[0005] Furthermore, the matching unit of the face recognition module adopts a feature point weighted comparison algorithm, and the preset threshold is 95%; if the similarity is <95%, the analysis unit will trigger an audible and visual alarm signal and send an interception command to the access control module.

[0006] Furthermore, the convolutional layers of the CNN model use 3×3 kernels with a stride of 1 and SamePadding; the pooling layers use 2×2 max pooling kernels with a stride of 2; the fully connected layers contain two hidden layers with 512 and 256 neurons respectively, the activation function is ReLU, and the output layer uses Softmax to normalize the feature vector.

[0007] Furthermore, the convolutional layers of the CNN model use 3×3 kernels with a stride of 1 and SamePadding; the pooling layers use 2×2 max pooling kernels with a stride of 2; the fully connected layers contain two hidden layers with 512 and 256 neurons respectively, the activation function is ReLU, and the output layer uses Softmax to normalize the feature vector.

[0008] Furthermore, the audio acquisition device of the voice interaction module is a high-fidelity noise-canceling microphone, and the image acquisition device shares a high-definition binocular camera with the acquisition unit of the face recognition module; the workstation function database includes operation positions, technical positions, management positions, and maintenance positions, and different workstation functions correspond to different voice interaction strategy permissions, among which the maintenance position can trigger equipment shutdown operation guidance scripts, while the operation position can only obtain equipment startup and parameter adjustment guidance scripts.

[0009] Furthermore, the environmental monitoring module includes an environmental monitoring unit, a control unit, a display device, a video monitoring unit, and an alarm unit connected in sequence, as well as an environmental data analyzer that establishes data connections with each unit. The environmental monitoring unit includes a temperature and humidity sensor, a gas sensor, and a dust sensor, which are used to collect environmental information such as workshop temperature, humidity, concentration of harmful gases, and concentration of dust, respectively. The control unit adopts a master-slave control architecture, including one master control board and at least two slave control boards. The master control board communicates with the microprocessor of the intelligent monitoring module, and the slave control boards control the environmental monitoring units in different areas respectively. The display device includes a main display screen and at least two slave display screens. The main display screen is located in the workshop control room, and the slave display screens are located in each workstation area to display environmental information and equipment operating status in real time. The video monitoring unit is a high-definition network camera that is linked in real time with the display device to capture video footage of areas with abnormal environments and store it synchronously. The warning unit includes three sub-warning units, corresponding to three levels: general warning, important warning, and emergency warning. Each sub-warning unit is an audible and visual alarm. Different warning levels correspond to different frequencies of audible and visual signals. In the event of an emergency warning, the intelligent monitoring module will be triggered to control the switching unit to cut off the power supply to the relevant equipment.

[0010] Furthermore, the environmental data analyzer has a built-in environmental threshold database that stores the safety threshold ranges for different environmental parameters. The environmental data analyzer compares the environmental information collected by the environmental monitoring unit with the safety threshold ranges and outputs the environmental level determination results. The environmental level determination results include normal, general abnormal, important abnormal, and emergency abnormal, which correspond to different warning levels of the warning unit.

[0011] Furthermore, the intelligent monitoring module also includes a data storage unit, which is a solid-state drive, used to store facial recognition records, voice interaction logs, equipment operating parameters, environmental monitoring data, and video surveillance footage, with a storage duration of not less than 90 days.

[0012] The beneficial effects of this invention are as follows: 1. Precise access control and job function matching, and improved security: The face recognition module adopts a binocular acquisition and feature point weighted comparison algorithm, and a 95% similarity threshold ensures accurate identity recognition and avoids misuse of permissions; the voice interaction module realizes accurate matching of identity and job function, and differentiated permission control for different positions, ensuring operational security from the source.

[0013] 2. Intelligent interaction and operation guidance, with outstanding convenience: The voice interaction strategy adapts to the needs of different workstations. Maintenance personnel can obtain shutdown guidance, while operators can only obtain startup and parameter adjustment guidance, which not only meets the needs of the operation but also avoids operational risks; voice guidance replaces manual query, improving work efficiency.

[0014] 3. High reliability with integrated equipment and environment control: The intelligent monitoring module and the environmental monitoring module work together in two directions, and the equipment operating status and environmental parameters interact in real time. The hierarchical early warning mechanism enables differentiated control, and the system automatically shuts down in case of emergency or abnormality, minimizing the risk of safety accidents.

[0015] 4. End-to-end data traceability and efficient management: The data storage unit stores various records for a long period of no less than 90 days, supporting the traceability and responsibility identification of security incidents; the wireless server enables remote data query and control, facilitating centralized management in the workshop.

[0016] 5. Modular design and compatibility, with wide adaptability: Each module adopts a standardized interface, supporting functional expansion and device compatibility; the CNN model optimizes feature extraction accuracy, and the environmental monitoring unit covers multiple environmental parameters, making it suitable for intelligent equipment rack production workshops of different sizes and types. Detailed Implementation

[0017] The present invention will be further described below with reference to specific embodiments, so that those skilled in the art can better understand and implement the present invention, but the embodiments are not intended to limit the present invention.

[0018] The intelligent equipment rack intelligent workshop safety system provided by the present invention includes a face recognition module, a voice interaction module, and an intelligent monitoring module that are linked in sequence. The intelligent monitoring module establishes bidirectional signal connections with the environmental monitoring module and the equipment module, and interacts with the wireless server through a wireless communication link.

[0019] The face recognition module consists of a data acquisition unit, an image preprocessing unit, an image processing unit, a matching unit, and an analysis unit connected in series. The data acquisition unit is a high-definition binocular camera. The image preprocessing unit performs noise reduction, grayscale correction, size normalization, and face region cropping on the acquired images. The image processing unit extracts image information, including facial organ size parameters, facial organ skin color features, and coordinates of key local facial feature points. The matching unit uses a feature point weighted comparison algorithm to compare the extracted image information with the pre-stored authorized personnel image information in the system database, with a preset threshold of 95%. The analysis unit analyzes the results based on the comparison... If the output identity recognition result has a similarity of ≥95%, it is determined that the authorization is passed; otherwise, the access control interception signal and the sound and light alarm signal are triggered, and an interception command is sent to the access control module. The image processing unit uses a CNN model for feature extraction. The convolutional layer of the CNN model uses a 3×3 kernel with a stride of 1 and SamePadding. The pooling layer uses a 2×2 max pooling kernel with a stride of 2. The fully connected layer contains two hidden layers with 512 and 256 neurons respectively. The activation function is ReLU. The output layer uses Softmax to normalize the feature vector.

[0020] The voice interaction module automatically starts after face recognition authorization is passed. Its audio acquisition device is a high-fidelity noise-canceling microphone, and the image acquisition device and the acquisition unit of the face recognition module share a high-definition binocular camera. The execution process of the voice interaction module includes: (1) synchronously acquiring multimedia information through the audio acquisition device and the image acquisition device. The multimedia information includes first voice information and real-time scene image; (2) performing sound source localization on the first voice information and calculating the spatial orientation coordinates of the current user through the microphone array; (3) based on the spatial orientation coordinates, selecting candidate users in that orientation from the real-time scene image and combining the identity information output by the face recognition module. (3) Based on the results, lock the unique current user; (4) According to the voice command content and identity recognition results in the multimedia information, match the pre-stored work position function database of the system to determine the current user's work position function and the group identifier; (5) According to the group identifier, call the corresponding preset voice interaction strategy, the voice interaction strategy includes operation guidance script, permission query feedback, risk warning prompt, and execute the voice interaction function; the work position function database includes operation position, technical position, management position, and maintenance position. Different work position functions correspond to different voice interaction strategy permissions. Among them, the maintenance position can trigger the equipment shutdown operation guidance script, while the operation position can only obtain the equipment start-up and parameter adjustment guidance script.

[0021] The intelligent monitoring module includes a core control microprocessor. The input of the microprocessor is connected to a detection unit via a signal conditioning circuit. The detection unit includes a current sensor, a temperature sensor, and a vibration sensor. The output of the microprocessor is connected to a switching unit via a drive circuit. The switching unit is a relay or an AC contactor. The switching unit, the device module, and the power supply are connected in series to form a closed loop. The microprocessor is connected to a wireless server via a detection wireless communication unit, which uses LoRa or 5G communication protocols. The intelligent monitoring module also includes a data storage unit, which is a solid-state drive, used to store facial recognition records, voice interaction logs, device operating parameters, environmental monitoring data, and video monitoring footage for a storage time of no less than 90 days. The intelligent monitoring module is signal-connected to the environmental monitoring module and controls it.

[0022] The environmental monitoring module includes an environmental monitoring unit, a control unit, a display device, a video monitoring unit, and an alarm unit connected in sequence, as well as an environmental data analyzer that establishes data connections with each unit. The environmental monitoring unit includes a temperature and humidity sensor, a gas sensor, and a dust sensor. The control unit adopts a master-slave control architecture, including one master control board and at least two slave control boards. The master control board communicates with the microprocessor of the intelligent monitoring module, and the slave control boards control the environmental monitoring units in different areas. The display device includes one master display screen and at least two slave display screens. The master display screen is located in the workshop control room, and the slave display screens are located in each workstation area. The video monitoring unit is... A high-definition network camera is linked in real time with the display device. The warning unit includes three sub-warning units, corresponding to three levels: general warning, important warning, and emergency warning. Each sub-warning unit is an audible and visual alarm. In the event of an emergency warning, the intelligent monitoring module will trigger the control switch unit to cut off the power to the relevant equipment. The environmental data analyzer has a built-in environmental threshold database that stores the safety threshold ranges for different environmental parameters. The environmental data analyzer compares the environmental information collected by the environmental monitoring unit with the safety threshold ranges and outputs the environmental level judgment result. The environmental level judgment result includes normal, general abnormal, important abnormal, and emergency abnormal, which correspond to different warning levels of the warning unit.

[0023] The working principle and specific operation process of this invention are as follows: 1. System Deployment and Initialization Complete the installation and debugging of each module: Install the high-definition binocular camera of the face recognition module at the workshop entrance access control point; deploy the high-fidelity noise-canceling microphone of the voice interaction module in conjunction with the camera; integrate the microprocessor, detection unit, and switch unit of the intelligent monitoring module into the equipment control cabinet, install the detection unit in key parts of the equipment, and connect the switch unit in series to the equipment power supply circuit; deploy the environmental monitoring units of the environmental monitoring module according to the workshop area, set the main display screen in the central control room, deploy the secondary display screens at each workstation, and install the warning unit and video monitoring unit in corresponding areas; deploy the data storage unit and wireless server in the central control room, and complete the signal connection and communication debugging of each module.

[0024] System initialization: Enter the facial image information of authorized personnel into the system database of the face recognition module to establish a complete authorized personnel database; configure the workstation functions and group identifiers of each personnel in the workstation function database of the voice interaction module, and preset the corresponding voice interaction strategies; set the safety threshold range of each environmental parameter in the environmental threshold database of the environmental monitoring module, and associate the environmental level with the warning level; set the storage strategy of the data storage unit to ensure that the data storage time of various types of data is not less than 90 days; complete the CNN model parameter configuration to ensure the accuracy of feature extraction.

[0025] 2. Personnel Admission and Job Matching Process Facial recognition access control: When personnel enter the workshop, a high-definition binocular camera captures facial images and facial depth information in real time; the image preprocessing unit performs noise reduction, grayscale correction, size normalization, and facial region cropping on the captured images; the image processing unit extracts facial organ size parameters, skin color features, and key local feature point coordinates through a CNN model; the matching unit uses a feature point weighted comparison algorithm to compare the extracted features with the authorization information in the system database; if the similarity is ≥95%, the analysis unit outputs an authorization pass signal, and the access control automatically opens; if the similarity is <95%, an audible and visual alarm is triggered, the access control is blocked, and the information of unauthorized personnel is recorded at the same time.

[0026] Voice Interaction and Job Function Matching: After authorization, the voice interaction module automatically starts; a high-fidelity noise-canceling microphone collects the user's initial voice information, and a high-definition binocular camera simultaneously captures real-time scene images; the system performs sound source localization on the voice information, calculates the user's spatial coordinates, and locates the current user from the scene images; based on the voice command content and identity recognition results, it matches the job function database to determine the user's job function and group identifier; and invokes the corresponding voice interaction strategy. For example, when a maintenance worker inquires about equipment shutdown procedures, the system provides shutdown procedure guidance; when an operator queries equipment startup parameters, the system provides startup guidance and parameter range prompts.

[0027] 3. Equipment and Environment Linkage Control Process Real-time monitoring data acquisition: The detection unit of the intelligent monitoring module collects equipment operating parameters (current, temperature, vibration data) in real time and transmits them to the microprocessor through the signal conditioning circuit; the environmental monitoring unit of the environmental monitoring module collects information such as workshop temperature and humidity, harmful gas concentration, and dust concentration, which is then aggregated from the control board to the main control board and transmitted to the microprocessor; the microprocessor synchronously uploads the equipment data and environmental data to the wireless server and stores them in the data storage unit.

[0028] Data processing and level determination: The environmental data analyzer compares the collected environmental information with the safety threshold range, outputs the environmental level determination result (normal, general abnormal, important abnormal, emergency abnormal), and links the corresponding warning level; the microprocessor analyzes the equipment operating parameters to determine whether the equipment is in normal operating condition.

[0029] Linked control execution: If the environmental level is normal and the equipment is operating normally, the system maintains real-time monitoring; if the environmental level is a general or major anomaly, the corresponding audible and visual alarm is activated, the video monitoring unit captures images of the abnormal area, and the display screen simultaneously displays the abnormal information and warning prompts; if the environmental level is an emergency anomaly, the warning unit activates an emergency warning, and the microprocessor controls the switching unit to cut off the power to the relevant equipment to prevent the accident from escalating; if the equipment operating parameters are abnormal, the microprocessor first sends a risk warning prompt to the personnel at the corresponding workstation through the voice interaction module. If the anomaly is not handled in time, the corresponding warning level is triggered according to the severity, and the equipment is shut down if necessary.

[0030] 4. Data Interaction and Traceability Management The wireless server receives and stores equipment operation data, environmental monitoring data, facial recognition records, voice interaction logs, and other information in real time, supporting remote query and monitoring by central control room administrators. Administrators can send remote control commands to the microprocessor through the wireless server to remotely start and stop the equipment and adjust parameters.

[0031] When a safety incident occurs, managers can retrieve historical data through the data storage unit, including personnel access records before and after the incident, equipment operating parameters, environmental monitoring data, video surveillance footage, etc., to achieve incident tracing and responsibility determination.

[0032] 5. System maintenance and upkeep Routine maintenance: Check the operating status of each module daily to ensure that the acquisition devices such as cameras, microphones, and sensors are clean and unobstructed; check whether the display screen is normal and whether the audible and visual signals of the warning unit are clear; check the storage status of the data storage unit to ensure sufficient storage space; clean the dust on the surface of the equipment and maintain good heat dissipation.

[0033] Regular maintenance: Conduct a comprehensive system inspection weekly, calibrate the acquisition accuracy of each sensor to ensure data accuracy; check the signal connection and communication status of each module to avoid poor contact or communication interruption; verify the switching function of the switching unit to ensure that the power supply can be cut off normally in an emergency; back up the critical data of the data storage unit to prevent data loss; check the running status of the CNN model and optimize parameters if necessary.

[0034] Long-term maintenance: Conduct a comprehensive overhaul of the system hardware every quarter, replacing aging sensors, alarms, and other components; update the environmental threshold database and workstation function database to adapt to workshop production adjustments; upgrade system software and communication protocols to improve system compatibility and security; and conduct comprehensive testing of data storage units to ensure long-term stable storage.

[0035] The above-described embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A smart equipment rack intelligent workshop safety system, characterized in that, The system includes a face recognition module, a voice interaction module, and an intelligent monitoring module that are linked sequentially. The intelligent monitoring module establishes bidirectional signal connections with the environmental monitoring module and the equipment module, and interacts with the wireless server via a wireless communication link. This enables full-process safety management and control of personnel access, job matching, equipment operation, and environmental status in the workshop. The specific structure and process are as follows: The face recognition module consists of a data acquisition unit, an image preprocessing unit, an image processing unit, a matching unit, and an analysis unit connected in series. The data acquisition unit is a high-definition binocular camera used to acquire real-time images of the user's face and facial depth information. The image preprocessing unit performs noise reduction, grayscale correction, size normalization, and face region cropping on the acquired images. The image processing unit extracts image information, including facial organ size parameters, facial organ skin color features, and coordinates of key local facial feature points. The matching unit compares the extracted image information with pre-stored authorized personnel image information in the system database for similarity. The analysis unit outputs an identity recognition result based on the comparison result; if the similarity is ≥ a preset threshold, authorization is granted; otherwise, an access control interception signal is triggered. Voice interaction module: Automatically starts after facial recognition authorization is successful. Its execution process includes: (1) Multimedia information is acquired synchronously through audio acquisition device and image acquisition device, wherein the multimedia information includes first voice information and real-time scene image; (2) Locate the sound source of the first voice information and calculate the spatial coordinates of the current user using the microphone array; (3) Based on the spatial orientation coordinates, select candidate users in that orientation from the real-time scene image, and combine the identity recognition results output by the face recognition module to lock in the unique current user; (4) Based on the voice command content and identity recognition results in the multimedia information, match the pre-stored work position function database of the system to determine the current user's work position function and the group identifier to which they belong; (5) Invoke the corresponding preset voice interaction strategy according to the group identifier. The voice interaction strategy includes operation guidance script, permission query feedback, risk warning prompt, and execute the voice interaction function. The intelligent monitoring module includes a core control microprocessor. The input of the microprocessor is connected to a detection unit via a signal conditioning circuit. The detection unit comprises a current sensor, a temperature sensor, and a vibration sensor, used to collect real-time operating parameters of the device module. The output of the microprocessor is connected to a switching unit via a drive circuit. The switching unit is a relay or an AC contactor. The switching unit, the device module, and the power supply are connected in series to form a closed loop. The microprocessor controls the switching unit's on / off state based on the data collected by the detection unit. The microprocessor also connects to a wireless server via a detection wireless communication unit. This detection wireless communication unit uses LoRa or 5G communication protocols to upload device operating data and receive remote control commands. The intelligent monitoring module is connected to the environmental monitoring module and controls it. The environmental monitoring module is used to monitor the workshop environment in real time and feed it back to the microprocessor, forming a linkage control between equipment operation and environmental status.

2. The intelligent equipment rack intelligent workshop safety system as described in claim 1, characterized in that, The matching unit of the face recognition module uses a feature point weighted comparison algorithm, and the preset threshold is 95%. If the similarity is less than 95%, the analysis unit will trigger an audible and visual alarm signal and send an interception command to the access control module.

3. The intelligent equipment rack intelligent workshop safety system as described in claim 1, characterized in that, The CNN model uses 3×3 kernels for its convolutional layers with a stride of 1 and SamePadding for padding; 2×2 max pooling kernels for its pooling layers with a stride of 2; and two fully connected layers with 512 and 256 neurons respectively, using ReLU as the activation function. The output layer uses Softmax to normalize the feature vectors.

4. The intelligent equipment rack intelligent workshop safety system as described in claim 1, characterized in that, The CNN model uses 3×3 kernels for its convolutional layers with a stride of 1 and SamePadding for padding; 2×2 max pooling kernels for its pooling layers with a stride of 2; and two fully connected layers with 512 and 256 neurons respectively, using ReLU as the activation function. The output layer uses Softmax to normalize the feature vectors.

5. The intelligent equipment rack intelligent workshop safety system as described in claim 1, characterized in that, The audio acquisition device of the voice interaction module is a high-fidelity noise-canceling microphone, and the image acquisition device shares a high-definition binocular camera with the acquisition unit of the face recognition module. The workstation function database includes operation, technical, management and maintenance positions. Different workstation functions correspond to different voice interaction strategy permissions. Among them, the maintenance position can trigger equipment shutdown operation guidance scripts, while the operation position can only obtain equipment startup and parameter adjustment guidance scripts.

6. The intelligent equipment rack intelligent workshop safety system as described in claim 1, characterized in that, The environmental monitoring module includes an environmental monitoring unit, a control unit, a display device, a video monitoring unit, and an alarm unit, which are connected in sequence, as well as an environmental data analyzer that establishes data connections with each unit. The environmental monitoring unit includes a temperature and humidity sensor, a gas sensor, and a dust sensor, which are used to collect environmental information such as workshop temperature, humidity, concentration of harmful gases, and concentration of dust, respectively. The control unit adopts a master-slave control architecture, including one master control board and at least two slave control boards. The master control board communicates with the microprocessor of the intelligent monitoring module, and the slave control boards control the environmental monitoring units in different areas respectively. The display device includes a main display screen and at least two slave display screens. The main display screen is located in the workshop control room, and the slave display screens are located in each workstation area to display environmental information and equipment operating status in real time. The video monitoring unit is a high-definition network camera that is linked in real time with the display device to capture video footage of areas with abnormal environments and store it synchronously. The warning unit includes three sub-warning units, corresponding to three levels: general warning, important warning, and emergency warning. Each sub-warning unit is an audible and visual alarm. Different warning levels correspond to different frequencies of audible and visual signals. In the event of an emergency warning, the intelligent monitoring module will be triggered to control the switching unit to cut off the power supply to the relevant equipment.

7. The intelligent equipment rack intelligent workshop safety system as described in claim 6, characterized in that, The environmental data analyzer has a built-in environmental threshold database that stores the safety threshold ranges for different environmental parameters. The environmental data analyzer compares the environmental information collected by the environmental monitoring unit with the safety threshold ranges and outputs the environmental level determination results. The environmental level determination results include normal, general abnormal, important abnormal, and emergency abnormal, which correspond to different warning levels of the warning unit.

8. The intelligent equipment rack intelligent workshop safety system as described in claim 1, characterized in that, The intelligent monitoring module also includes a data storage unit, which is a solid-state drive, used to store facial recognition records, voice interaction logs, equipment operating parameters, environmental monitoring data and video surveillance footage, with a storage duration of no less than 90 days.