Visitor identity management system with accurate and efficient face recognition

Abstract

The invention discloses a visitor identity management system with accurate and efficient face recognition, relates to the technical field of visitor management, and aims to deploy high-definition cameras with an active near-infrared technology in each key area of a cruise ship, extract facial features by using a deep learning convolutional neural network model, accurately verify the identity and effectively guarantee the safety of the cruise ship. Multiple standard network protocols are applied, different systems are connected to obtain multi-source data such as cruise ship tracks, equipment operation, weather and personnel health, and a basis is provided for operation decision making. And passenger flow monitoring equipment is installed in each playing area, real-time passenger flow monitoring, risk early warning and flow limiting are realized in combination with a risk assessment model, and the tourist experience is optimized. The APP integrates positioning and path planning functions, evacuation is guided by means of multi-mode interaction and AR technology in emergency, and the emergency response efficiency is improved by cross-system collaborative linkage. Meanwhile, data security and authority management are guaranteed, early warning optimization and intelligent adjustment are achieved, and system performance and service quality are comprehensively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of visitor management, in particular to a visitor identity management system with precise and efficient face recognition. BACKGROUND

[0002] In the operation and management of cruise ships, traditional visitor identity management methods have many drawbacks. For example, manual registration is inefficient and prone to errors, making it difficult to quickly and accurately verify visitor identities, and can cause congestion during peak passenger flow, affecting the boarding experience of tourists. At the same time, there is a lack of real-time monitoring and risk warning mechanism for visitor activities, which cannot timely discover and respond to safety hazards such as excessive gathering of personnel, failure of facilities and equipment, and adverse weather. In terms of emergency response, data is isolated between systems, with poor collaboration, making it difficult to quickly and effectively integrate resources for emergency handling. Moreover, as cruise ships continue to expand in size and the number of tourists increases, the demand for data security and permission management is also increasing, and traditional methods are difficult to meet these needs. The limitations of existing technologies have led to the birth of the present application to address the low efficiency, poor security, and lack of collaboration in visitor identity management on cruise ships.

[0003] In view of the above, the present application is proposed. SUMMARY

[0004] The purpose of the present application is to provide a visitor identity management system with precise and efficient face recognition to solve the problems in the background art.

[0005] To solve the above technical problems, the present application provides a visitor identity management system with precise and efficient face recognition, comprising: A face recognition identity verification module: high-definition cameras equipped with active near-infrared technology are deployed at each entrance and key area of the cruise ship to collect facial images, edge computing nodes run a convolutional neural network model based on deep learning for feature extraction, and the identity is verified by comparing the Euclidean distance with the visitor face database in the cloud server; Active near-infrared technology in conjunction with high-definition cameras allows the system to stably collect clear facial images under various lighting conditions, ensuring that identity verification is not affected by light. The convolutional neural network model based on deep learning can accurately extract facial features, improve recognition accuracy, and reduce false rejection rate and false acceptance rate. Distributing the computing task to edge computing nodes can reduce the pressure on the cloud and speed up the processing. By comparing with the cloud database, the visitor's identity can be quickly verified, effectively preventing illegal personnel from entering the key areas of the cruise ship, building the first line of defense for safe operation of the cruise ship, and protecting the personal and property safety of passengers and crew; Data access module: Use standard network communication interface and TCP / IP protocol to access the cruise ship trajectory tracking system to obtain cruise ship position information, use Modbus TCP protocol to access the cruise ship equipment operation monitoring system to obtain equipment data, use HTTP / HTTPS protocol to access the weather data provider to obtain weather information, and use the secure data interaction protocol to access the personnel health management system to obtain personnel health information. Through various standard protocols, different systems are connected to realize efficient collection of multi-source data, providing comprehensive information support for cruise ship operation. The obtained cruise ship position information helps to reasonably plan the route and arrange the stopover points. The equipment operation data can monitor the equipment status in real time, timely detect hidden troubles, arrange maintenance in advance, and reduce the impact of equipment failure on cruise ship operation. Weather information can help the cruise ship to respond to severe weather in advance and ensure the safety of navigation. Personnel health information provides a basis for medical care and epidemic prevention, ensuring the health and safety of passengers and crew. At the same time, the use of standard protocols enhances system compatibility, making it easy to extend and access new data sources in the future; Passenger flow monitoring and risk warning module: Install passenger flow monitoring devices in various play areas on the cruise ship, and integrate and analyze passenger flow data on the edge computing node. The risk assessment model on the server combines multi-source data to assess risks. When the passenger flow exceeds or is lower than the preset threshold, the APP pushes prompt information, and when the threshold is exceeded, flow limiting measures are implemented in the relevant area. Real-time monitoring of passenger flow data can effectively prevent safety accidents caused by excessive gathering of people and ensure the safety of tourists in the play area. According to the passenger flow, prompt information is pushed to tourists to help them plan their play route reasonably, reduce waiting time, and improve the play experience of tourists. The flow limiting measure when the threshold is exceeded can ensure that the play area is always within the safe carrying range and maintain a good play order. The risk assessment model combined with multi-source data can more comprehensively and accurately assess risks, providing scientific decision-making basis for cruise ship operation and management, optimizing resource allocation, and improving operation efficiency; Immersive interaction guidance emergency assistance module: Integrate related functions and initialize data in the official APP of the cruise ship. The APP calls the mobile phone positioning module to determine the tourist's location in combination with the cruise ship's three-dimensional map, uses path planning algorithms to plan the route, receives sensor data from the cruise ship central control system, and starts the emergency assistance mode when an emergency is detected. Through multi-modal interaction and AR technology, tourists are guided to evacuate. This module provides precise navigation services for tourists, making it easy for them to quickly find their destination and improving the convenience and comfort of cruise ship play. In emergency situations, the emergency assistance mode can be quickly started, and a safe and effective escape route can be planned based on real-time sensor data. Multi-modal interaction (voice, vibration, push notification) and AR technology enable tourists to clearly and accurately receive escape instructions in emergency situations, avoiding disorientation due to panic and greatly improving evacuation efficiency, reducing accident losses, and ensuring the safety of tourists. At the same time, the collected tourist location and behavior data can provide reference for the cruise ship operator to optimize layout and services; The cross-system coordination and linkage emergency response module integrates cruise ship fire fighting, medical treatment, security and other subsystems through a unified data interface, continuously monitors real-time data of each subsystem, analyzes and judges abnormal data to determine an emergency event, issues a coordinated response instruction according to a preset strategy, and tracks and adjusts the response process. The data barriers between the subsystems are broken, information sharing and coordinated work are realized, and the emergency response speed is significantly improved. In the event of an emergency, various resources can be quickly integrated, such as the fire fighting, medical treatment and security systems responding simultaneously when a fire occurs, each performing its own function, and effectively reducing the loss caused by the accident. Through continuous monitoring and analysis of data, the type, scale and influence range of the emergency event can be more accurately determined, so that the coordinated response instruction issued is more targeted and effective. The response process tracking and adjustment mechanism can flexibly optimize the emergency strategy according to the actual situation, ensuring the efficiency and scientificity of the emergency handling, and protecting the safety of cruise ship operation. The blockchain permission management and data storage module is based on a consortium chain mode to build a blockchain network, assigns digital identities and public and private key pairs to participating nodes, generates access permissions in the form of a smart contract according to visitor information, stores and certifies visitor-related data after hash operation, and can query and verify data through a blockchain browser. The distributed storage and encryption technology of the blockchain ensures the security of visitor data, preventing data tampering and theft. Even if some nodes have problems, the data remains complete and usable. The smart contract precisely controls visitor access permissions, reduces permission management vulnerabilities caused by human intervention, and improves the fairness and transparency of management. Hash operation and on-chain storage ensure data tamper resistance and traceability, providing reliable data basis for accident investigation and responsibility identification. Through the blockchain browser, authorized parties can quickly obtain accurate information, improving management efficiency and reducing trust costs.

[0006] Further, in the face recognition identity verification module, the activation function used by the convolutional neural network model is the ReLU function, which is trained through the stochastic gradient descent algorithm. The ReLU activation function can effectively alleviate the gradient vanishing problem, speed up the training of the model, and make the convolutional neural network more efficient in feature extraction. The stochastic gradient descent algorithm can quickly converge on large-scale data sets, helping to improve the accuracy and efficiency of face recognition identity verification and ensuring that visitor identities can be quickly and accurately verified.

[0007] Further, in the data access module, when interfacing with the personnel health management system, symmetric encryption and asymmetric encryption are combined to encrypt and transmit data. The combination of symmetric encryption and asymmetric encryption takes full advantage of the strengths of both encryption methods. Symmetric encryption is fast and can efficiently encrypt large amounts of data, while asymmetric encryption is highly secure and can be used for key exchange and identity verification. This combination ensures the security and integrity of personnel health information during transmission, preventing data leakage and tampering.

[0008] Further, in the passenger flow monitoring and risk prompting module, the passenger flow analysis camera based on computer vision adopts YOLO series algorithm, and the risk assessment model adopts random forest algorithm or support vector machine algorithm; YOLO series algorithm has fast and accurate target detection capability, can analyze passenger flow situation in real time and efficiently, and provides accurate data for passenger flow monitoring. Random forest algorithm and support vector machine algorithm have high accuracy and stability in processing multi-source data for risk assessment, can more accurately assess passenger flow risk, and provide reliable basis for flow limiting and risk prompting decision-making.

[0009] Further, in the immersive interaction guiding emergency assistance module, internal positioning in the cruise ship mainly relies on Bluetooth positioning and Wi-Fi positioning, and the path planning algorithm adopts Dijkstra algorithm or algorithm; Bluetooth positioning and Wi-Fi positioning have good coverage and positioning accuracy in the internal of the cruise ship, and can accurately determine the position of the tourists. Dijkstra algorithm and algorithm have high efficiency and accuracy in path planning, and can plan the optimal route according to the position of the tourists and the layout of the cruise ship. In emergency situations, these technologies can quickly guide tourists to evacuate, improve the efficiency of emergency handling and the safety of tourists.

[0010] Further, in the cross-system collaborative linkage emergency response module, one of RESTful API and message queue technology is used for system integration, and data mining and machine learning algorithms are used for data fusion and analysis; RESTful API or message queue technology can realize efficient integration and data interaction between different subsystems, and ensure accurate and timely information transmission between subsystems. Data mining and machine learning algorithms can deeply analyze and fuse multi-source real-time data, improve the analysis and judgment ability of abnormal data, so as to more accurately determine emergency events and develop reasonable response strategies, improve the collaboration and effectiveness of emergency response.

[0011] Further, in the blockchain permission management and data storage module, Hyperledger Fabric is used as the blockchain platform, the consensus mechanism is PBFT (Practical Byzantine Fault Tolerance) mechanism, and SHA-256 hash algorithm is used for data storage; Hyperledger Fabric has high scalability and flexibility, and can meet the needs of the cruise visitor identity management system. PBFT (Practical Byzantine Fault Tolerance) mechanism can quickly reach consensus in a distributed environment, ensuring efficient operation of the system. SHA-256 hash algorithm has high security, can guarantee the non-tamperability and traceability of visitor-related data stored on the blockchain, and improve the credibility and security of data.

[0012] Further, it also includes a pre-warning information push optimization module, which intelligently optimizes the push method (such as push frequency, push channel priority) and content (such as detail level, urgency level identification) according to different types of risk pre-warning information, ensuring that visitors receive and understand pre-warning information in a timely and accurate manner; the pre-warning information push optimization module can optimize the push method and content according to the characteristics of different risk pre-warning information. This can improve the attention and effectiveness of pre-warning information, enabling visitors to receive and understand pre-warning information in a timely and accurate manner, thereby better responding to various risk situations, ensuring their safety and improving their play experience.

[0013] Further, it also includes an intelligent learning and adaptive adjustment module that continuously learns and analyzes historical visitor data, emergency event data, and system operation data to automatically optimize the parameter settings (such as face recognition threshold, risk assessment model weight) of each module to adapt to different operating scenarios and changes; the intelligent learning and adaptive adjustment module can automatically optimize the parameter settings of each module of the system through learning and analysis of various data. This enables the system to adaptively adjust according to different operating scenarios and changes, such as different visitor flow and different weather conditions, improving the accuracy and reliability of the system and better meeting actual operational needs.

[0014] Further, it also includes a multi-language interaction module that supports multi-language information display and interaction functions, making it convenient for visitors with different language backgrounds to use; the multi-language interaction module can meet the needs of visitors with different language backgrounds, making the system more widely applicable. Visitors can access information and interact in their familiar language, greatly improving the universality and friendliness of the system and providing better service experience for tourists from different countries and regions.

[0015] Compared with the prior art, the beneficial effects of the present application are: Precise and efficient identity verification: The high-definition camera equipped with active near-infrared technology, combined with the convolutional neural network model of deep learning and the Euclidean distance comparison algorithm, creatively solves the problem of face recognition in complex lighting environments. This technology is not affected by light interference and can accurately extract facial features, with a false recognition rate of less than 1%, and the entire verification process can be completed within a few seconds, greatly improving the accuracy and efficiency of identity verification and effectively preventing illegal personnel from boarding the ship, thereby building a strong defense for the safe operation of cruise ships.

[0016] Comprehensive real-time data support: Using standard network communication protocols such as TCP / IP, Modbus TCP, HTTP / HTTPS, etc., the efficient access of multi-source data is creatively realized. Real-time access to cruise ship location, equipment operation, weather, and personnel health information can provide comprehensive basis for route planning, equipment maintenance, severe weather response, and medical security, preventing potential safety hazards in advance and ensuring navigation safety and personnel health.

[0017] Intelligent passenger flow and risk control: Passenger flow monitoring devices are deployed in various play areas on the cruise ship, and risk assessment models are built using algorithms such as random forest or support vector machine. By monitoring passenger flow data in real time, risks can be accurately assessed, prompt information can be pushed, and flow control measures can be implemented. This creative initiative effectively prevents safety accidents caused by excessive gathering of people, improves the tourist experience, and provides scientific decision-making basis for cruise ship operation and management, optimizing resource allocation.

[0018] Efficient emergency response and guidance: The cruise ship official APP integrates multiple functions, uses Bluetooth and Wi-Fi positioning technology, combines with cruise ship three-dimensional map and path planning algorithms to achieve precise navigation. In emergency situations, tourists can be guided to evacuate through multi-modal interaction and AR technology. The cross-system collaborative emergency response module breaks down data barriers, realizes information sharing and collaborative work among subsystems, greatly shortens the emergency response time, improves the scientificity and efficiency of emergency handling, and ensures the safety of tourists.

[0019] Safe and reliable data management: Based on the alliance chain mode, a blockchain network is built, and public and private key pairs, smart contracts, and SHA-256 hash algorithms are used to creatively ensure the safety, tamper resistance, and traceability of visitor data. Precise control of visitor access rights reduces human intervention vulnerabilities, improves management fairness and transparency, facilitates authorized parties to query and verify data, and improves management efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 A principle block diagram of a visitor identity management system with precise and efficient face recognition; Figure 2 A principle block diagram of a face recognition identity verification module in a visitor identity management system with precise and efficient face recognition; Figure 3 A principle block diagram of a data access module in a visitor identity management system with precise and efficient face recognition; Figure 4 A principle block diagram of a passenger flow monitoring and risk prompting module in a visitor identity management system with precise and efficient face recognition; Figure 5For a kind of visitor identity management system of accurate and efficient face recognition, the principle block diagram of immersive interaction guide emergency auxiliary module; Figure 6 For a kind of visitor identity management system of accurate and efficient face recognition, the principle block diagram of cross-system coordination linkage emergency response module; Figure 7 For a kind of visitor identity management system of accurate and efficient face recognition, the principle block diagram of early warning information push optimization module. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0022] Please refer to Figures 1-7 , the present application provides a technical solution: A week-long voyage of a certain luxury cruise ship carries 2000 passengers and 500 crew members, and a visitor identity management system of accurate and efficient face recognition is applied comprehensively, including: The face recognition identity verification module operates as follows: I. Implementation steps 1. Device deployment and image acquisition: Install high-definition active near-infrared cameras at key locations such as boarding ports, regional entrances, safe deposit box storage, machine rooms, etc. The cameras actively emit near-infrared light to capture visitor facial images and digitize them, and transmit them to edge computing nodes through Ethernet or Wi-Fi.

[0023] 2. Feature extraction: The edge computing node runs a ResNet convolutional neural network model, which enhances expression ability through ReLU activation function after multi-layer convolution and pooling operation, and finally extracts a 128-dimensional face feature vector.

[0024] 3. Identity comparison and verification feedback: The feature vector is transmitted to the cloud server and compared with the visitor face database for similarity through Euclidean distance. The formula is: , (here n=128).

[0025] If the distance is <0.5 (preset threshold), the verification is successful, and the visitor information is pushed to the staff terminal; otherwise, manual verification is prompted.

[0026] II. Key notes 1. Convolution operation formula: ResNet convolution layer operation is: ; For output features, For activation function (such as ReLU function, ), by stochastic gradient descent algorithm training, adjusting the weight w and bias b to accurately extract features.

[0027] 2. Euclidean distance and threshold: the smaller the distance, the higher the identity matching possibility. Threshold 0.5 is determined through multiple scene tests, balancing the false recognition rate (non-matching misjudgment as matching) and the rejection rate (matching misjudgment as non-matching), avoiding threshold too low to increase the rejection rate or too high to increase the false recognition rate.

[0028] III. Application example A business passenger who often takes the cruise ship is captured by the camera at the boarding port when boarding, and the face image is transmitted to the edge computing node. The ResNet model generates a 128-dimensional feature vector, which is compared with the cloud database, and the Euclidean distance is less than 0.5, and the verification is successful. The staff terminal receives the name, cabin number and other information, and guides the boarding, the whole process takes less than 5 seconds.

[0029] The advantages of the present application are as follows: high accuracy: false recognition rate <1%, low rejection rate, prevent illegal personnel from entering. High efficiency: complete verification within seconds, reduce queuing, improve experience and operation efficiency. Strong adaptability: active near-infrared technology ensures stable work in strong light, low light or even dark environment. Good integration: seamless docking with boarding management, room allocation and other systems, realizing data sharing.

[0030] Data access module: I. Docking steps 1. Cruise ship trajectory tracking system docking: based on TCP / IP protocol, through Ethernet interface to build data channel, transmit cruise ship latitude, longitude, speed, direction and other data in JSON format every second. The edge node first performs data cleaning (remove outliers, repeated values), and then uploads to the cloud to support route planning and risk assessment.

[0031] 2. Equipment operation monitoring system docking: ModbusTCP protocol is used to connect power, ventilation, elevator and other systems, and the data (voltage, air volume, etc.) is polled periodically according to the device address (such as power 01, ventilation 02). After calibration by the edge node, it is uploaded to realize equipment fault warning.

[0032] 3. Weather data acquisition: request real-time and future 24-hour weather data (temperature, wind speed, etc.) from service providers through HTTP / HTTPS protocol, parameters include latitude and longitude and route area, upload after parsing JSON / XML format data, support route adjustment.

[0033] 4. Personnel health management system interface: based on HTTPS protocol, using AES (symmetric encryption) and RSA (asymmetric encryption) combined way to transfer health data (body temperature, vaccine records, etc.), after digital certificate verification, edge node cache backup and upload, to ensure medical and epidemic prevention needs.

[0034] II. Key notes Protocol details: TCP / IP establishes connection through three-way handshake, data contains source / destination IP (such as 192.168.1.101 and 192.168.1.100) and port number; ModbusTCP reads register data with function code 03; HTTPS request parameters include geographic location (latitude=20.0&longitude=110.0) etc.

[0035] Encryption mechanism: when transmitting health data, first exchange AES key with RSA, then encrypt data with AES to prevent leakage and tampering.

[0036] III. Application examples On the third day of navigation, the system obtains meteorological data that 15m / s strong wind and 3m sea wave will appear in the island area within 6 hours, and monitors that the deck ventilation opening is open. The command center closes the ventilation opening and fine-tunes the route accordingly, avoiding equipment damage and safety accidents.

[0037] The module advantages of the present application are as follows: comprehensiveness: integrating navigation, equipment, weather, and health multi-source data to support accurate decision-making; real-time: high-frequency collection and transmission ensure information timeliness, gaining time for emergency response; compatibility: standard protocol adapts to multiple systems, facilitating extension and upgrade; security: encryption technology ensures data security, complying with privacy regulations.

[0038] Passenger flow monitoring and risk prompting module: I. Implementation steps 1. Device deployment and data collection: deploy devices in theaters, water parks, etc.: open areas use computer vision cameras, narrow passages use infrared counters. Devices collect personnel data (cameras capture images, infrared record passage) in real time, and transmit to edge computing nodes in JSON format through Ethernet / Wi-Fi.

[0039] 2. Data integration and preliminary analysis: edge nodes integrate data: camera data uses YOLO algorithm (based on convolutional neural network) to identify the number of personnel, infrared data directly counts, and correlates time and area to form structured data (such as "20XX-XX-XX10:00 Water Park Entrance 50 People").

[0040] 3. Risk assessment and early warning release: data is transmitted to the server, and the risk assessment model (random forest or support vector machine) combines multiple source data such as cruise trajectory and weather to assess the risk. When the threshold is exceeded (such as 800 people in the water park threshold), the APP pushes the prompt and limits the flow (such as the gate from 20 people / minute to 10 people / minute); combined with weather data, additional warnings are issued (such as rain and anti-skid prompts).

[0041] II. Key notes 1. YOLO algorithm convolution formula: the core is , wherein is the output feature, is the activation function (such as ReLU function, ), which identifies personnel by adjusting weights and bias .

[0042] 2. Support vector machine (SVM) optimization formula to solve the optimal classification hyperplane: , The constraint condition is: , and the nonlinear data is mapped to a high-dimensional space through a kernel function (such as ).

[0043] 3. Basic data types include real-time passenger flow, threshold, cruise trajectory, weather (temperature, wind speed), equipment operation (elevator status), etc., providing input for risk assessment.

[0044] III. Application examples On the fourth day of the summer vacation at 11 o'clock, the water park passenger flow reached 700 people (close to the 800 people threshold), and the system pushed the congestion prompt to more than 200 people going there, and adjusted the gate speed from 20 people / minute to 10 people / minute; combined with weather data, the system pushed the anti-skid warning of rain in 1 hour to more than 300 people in the field. The final passenger flow stabilized at 750 people, avoiding safety accidents.

[0045] The modules of the present application have the following advantages to improve safety: real-time monitoring and flow reduction to reduce congestion risk and prevent clustering accidents. Optimize experience: push people more or less prompts, reduce queuing, and improve satisfaction. Scientific decision-making: multi-source data support assessment to help adjust performances, cleaning and other plans. Efficiency and cost reduction: automated monitoring reduces manual labor and avoids facility damage and operational disruption.

[0046] Immersive interactive guidance and emergency assistance module: I. Implementation steps 1. Function integration and data initialization: The cruise ship's official APP integrates emergency assistance functions, with built-in 3D maps, facility distribution, safety exits and emergency plans. It establishes a data connection with the central control system to obtain real-time equipment status (such as smoke sensor data) and personnel distribution information.

[0047] 2. Navigation Guidance Service Response: When a passenger initiates navigation, the app uses Bluetooth and Wi-Fi positioning combined with a 3D map to determine the precise location. The algorithm plans the optimal route and provides guidance through 3D visualization and voice prompts (such as "Go straight for 50 meters and turn left, the restaurant is on your right").

[0048] 3. Emergency Scenario Perception and Response: Receives sensor data from the central control system in real time. When an emergency event (such as a fire) is detected, the system immediately activates the emergency mode, avoids dangerous areas and densely populated areas, and replans an escape route to the nearest safe exit.

[0049] 4. Multimodal interactive guidance implementation: In an emergency, information is delivered through voice (urgent prompts), vibration, and push notifications. AR technology overlays escape arrows on the camera screen to intuitively guide evacuation and reduce panic.

[0050] II. Key Explanations 1. Positioning and route planning: Bluetooth (signal strength triangulation) and Wi-Fi (access point signal fingerprint matching) enable accurate indoor positioning; The core formula of the algorithm is: .in, This indicates the nodes passed from the starting point. The estimated total cost to the destination. Indicates the distance from the starting point to the node. The actual cost (such as the distance traveled, the time spent, etc.). From node The estimated cost to the destination (a heuristic function, typically using Manhattan distance or Euclidean distance) ensures the route is optimal.

[0051] 2. AR technology principle: The camera captures scene images, extracts feature points such as corners and matches them with the cruise ship's 3D model. After calculating the phone's posture, the virtual escape guide is superimposed on the real scene to achieve intuitive navigation.

[0052] III. Application Examples On the third night of the voyage, a fire broke out in the entertainment area, affecting more than 100 passengers. APP immediately switched to emergency mode, pushed alerts to surrounding passengers, and re-planned escape routes. A passenger's original route was blocked, and the APP prompted "follow the right channel to the rear safety exit" through voice prompts and vibration reminders. After opening the camera, the AR arrow clearly pointed the direction, and eventually all passengers were safely evacuated within 5 minutes, with no casualties.

[0053] The modules in the present application have the following advantages: improving emergency efficiency: starting evacuation within 1 minute in fire cases, significantly speeding up and reducing losses compared to traditional methods. Optimizing experience: precise navigation reduces route-finding time and improves convenience. Enhancing safety: multi-modal guidance (voice, AR, etc.) ensures that all passengers respond quickly and improves evacuation success rate. Facilitating management: optimizing facility layout (such as adding popular area dining) through navigation data.

[0054] Cross-system coordinated emergency response module: I. Implementation steps 1. System integration: Use RESTful API unified interface to integrate fire (FS001), medical (MS001), security (SS001) and other subsystems, communicate through HTTP / HTTPS protocol, and develop emergency event codes (such as fire E001, sudden illness E002) and response specifications.

[0055] 2. Real-time data monitoring: continuously collect subsystem data: fire (smoke density ≥ 50 ppm, temperature > 60℃ is abnormal), security (monitoring video stream, personnel trajectory), medical (inventory, on-duty information). Edge computing devices analyze video streams in real time and use YOLO algorithm to identify abnormal behavior (such as lingering in restricted areas).

[0056] 3. Emergency event determination: after the event is triggered by abnormal data, determine the type and level based on multi-source information: fire is divided into scales according to smoke density, temperature and spread trend; security events are identified according to permissions and behavior patterns to identify threats (such as unauthorized access to machine rooms).

[0057] 4. Collaborative instruction issuance: send instructions according to preset strategies: fire starts sprinkler, security blocks area, medical prepares first aid supplies; each subsystem synchronously executes operations.

[0058] 5. Response tracking and adjustment: monitor the execution in real time, if deviating from the expected (such as fire extinguishing equipment failure), immediately adjust the strategy (supplement equipment, increase personnel), ensure efficient emergency.

[0059] II. Key notes 1. Interface and algorithm: RESTful API enables data interaction through methods such as POST (e.g., sending fire data to " / fire-system / alarm"); YOLO algorithm convolution formula: ; wherein is the output feature, is the activation function, and personnel behavior is identified through training.

[0060] 2. Emergency strategy basis: Based on historical data modeling, such as matching the number of fire extinguishing equipment, evacuation time, and medical resources for fire response, to ensure the accuracy of instructions.

[0061] III. Application example On the fourth day of navigation in the early morning, the kitchen smoke concentration reached 80 ppm and the temperature was 70°C, triggering a fire response. The module determined that it was a small-scale but had the risk of spreading, and immediately instructed: fire fighting to start alarm and sprinkler (part of the fire extinguisher failure after the device is dispatched), security to block the area within 5 minutes, and medical treatment to prepare materials within 10 minutes. Finally, the fire was extinguished in 30 minutes without casualties.

[0062] The advantages of the module in the present application are as follows: speed response: 1 minute start response in fire case, greatly shortening the time compared with traditional way. Strengthen cooperation: break the data barrier, fire, security, medical treatment synchronous action, improve the overall efficiency. Scientific decision-making: multi-source data support event judgment, instruction is more targeted. Reduce cost and increase efficiency: precise allocation of resources reduces waste, reduces facility damage and operating costs.

[0063] Early warning information push optimization module: I. Implementation steps 1. Early warning information reception and analysis: The module is connected with passenger flow, equipment failure, and meteorological monitoring systems, and receives early warning information (including area, data, threshold, level, etc.) in real time. After analysis, the key information is extracted to lay the foundation for subsequent processing.

[0064] 2. Analysis and evaluation and strategy development: Determine the type of warning (passenger flow, safety risk, etc.), and evaluate the trend and fluctuation of passenger flow warning analysis and the impact range of safety warning based on historical data. According to this, develop a push strategy - high-level safety warning with forced pop-up (30 seconds / time), low-level passenger flow warning with ordinary message (15 minutes / time).

[0065] 3. Content generation and format optimization: Generate content according to the strategy: high-level warning with flashing icon (such as red exclamation mark) and "please take action immediately" prompt, low-level with mild expression. Adapt to the terminal (APP adjusts text layout, display screen optimizes layout) to ensure clear display.

[0066] 4. Push execution and feedback collection: Push through APP, broadcast, and display screen, monitor delivery status. Collect feedback (view rate, action situation), analyze effect to optimize subsequent strategy (such as adjust push time).

[0067] II. Key notes 1. Early warning level evaluation algorithm: Taking passenger flow early warning as an example, the formula is When and is "high"; when and , it is determined as "medium"; when , it is determined as "low".

[0068] 2. Push strategy decision tree: Taking early warning level as root node, combining type to output strategy: high-level safety warning uses forced pop-up window (30 seconds / time), high-level passenger flow warning uses APP top (1 minute / time), and historical effect data is trained and optimized.

[0069] 3. Core data: Including monitoring system early warning data (passenger flow, equipment, etc.), cruise ship basic data (regional layout), historical push data (effect feedback), supporting algorithm operation and strategy formulation.

[0070] III. Application example On the third day of the voyage in the afternoon, the water park passenger flow reached the threshold of 85%, and the fluctuation coefficient was 1.3 (high level). The module pushed the forced pop-up window (yellow warning icon + "long queue" prompt), 1 minute / time, and the display screen played in a loop. The result showed that 80% of the passengers who received the push changed their route, and the push effect was significant.

[0071] In the present application, the module has the following advantages: improving attention: high-level warning view rate increased by 30%+, ensuring timely information reach. Enhance safety protection: emergency warning quickly push, gain time for emergency. Optimize experience: travel planning satisfaction increased by 15%, reduce queuing time. Improve operational efficiency: feedback data help optimize threshold and strategy, fine management.

[0072] Intelligent learning and adaptive adjustment module: The system continuously collects and analyzes historical visitor data, emergency events, and system operation data to realize automatic optimization of parameters: such as when the face recognition misrecognition rate of a certain area is high, the threshold is automatically fine-tuned from 0.5 to 0.45, and the convolutional neural network model is retrained to improve accuracy; at the same time, according to the passenger flow law of season and voyage, the risk assessment model weight is optimized to enhance the accuracy of safety risk prediction.

[0073] Multilingual interaction module: When foreign tourists open the APP, the system automatically detects the phone language setting (such as English), and switches the interface (including navigation prompts, warning information, and operation instructions) to the corresponding language; the information display screen in the public area of the cruise ship (elevator, corridor) also supports multilingual switching to meet the information acquisition needs of visitors with different language backgrounds.

[0074] Core explanation of the blockchain permission management and data storage module: I. Generating and deploying access permissions based on visitor information 1. Information collection and organization: Collect visitor information such as name, ID number, itinerary, etc. before boarding, and verify the format and completeness.

[0075] 2. Role and permission definition: Preset roles such as ordinary tourists and crew family members, and assign permissions according to the "least privilege principle" (such as ordinary tourists can only access public areas and reserved rooms).

[0076] 3. Smart contract writing and deployment: Write contracts in languages such as Solidity, receive visitor information and verify its authenticity (such as interface comparison with identity agencies), extract corresponding permissions according to roles, associate visitor unique identifiers (such as ID card hash values), and deploy to the blockchain to ensure that permissions cannot be tampered with and are automatically executed.

[0077] II. Visitor data hashing and on-chain storage 1. Data screening and preparation: Determine the storage range (identity information, boarding records, etc.), and desensitize sensitive information (such as part of the ID number) to protect privacy.

[0078] 2. Hash operation execution: Use the SHA-256 algorithm to operate on the screened data block to generate a unique "digital fingerprint" (hash value). For example, a certain visitor information data block gets a hash value like "56a9f8d23f54792b93e756f8c32d1e49876543210987654321098765432109876" after operation, and small changes in data will result in significant differences in hash values.

[0079] 3. On-chain storage operation: Package the hash value and metadata (visitor identifier, storage time, etc.) into a transaction request and send it to the blockchain network. After the node verifies the legality of the transaction, it writes the record to a new block, achieving data tamper-proof storage. When verifying data integrity, re-operate and compare the hash value.

[0080] Advantages of modules in the invention: Intelligent learning improves system adaptability; multilingual function enhances universality; blockchain technology ensures data security and permission transparency, providing reliable support for cruise ship safety management.

[0081] To sum up: in the present application, through the face recognition identity verification module, the active near-infrared high-definition camera is used to collect facial images, combined with deep learning convolutional neural network and Euclidean distance comparison, the precise and efficient visitor identity verification function is realized, and the cruise ship safety is ensured. The data access module is used, the standard network communication protocol and multiple systems are connected, multiple source data is obtained, comprehensive data support is provided for cruise ship operation management, and various decisions are assisted. The passenger flow monitoring and risk prompting module is used, the monitoring equipment is installed in each play area, the data is analyzed combined with the risk assessment model, the passenger flow real-time monitoring, risk early warning and flow control are realized, the passenger experience and operation safety are improved. Relying on the immersive interactive guidance emergency auxiliary module and the cross-system collaborative linkage emergency response module, the APP integrated function is combined with positioning, path planning technology and multi-system collaboration, the rapid response and efficient emergency treatment in emergency situation are realized. Through the blockchain permission management and data evidence module, based on the alliance chain mode, the smart contract and hash operation are used, the data security and the fair and transparent permission management are ensured. With the aid of the early warning information pushing optimization module, the intelligent learning and self-adaptive adjustment module and the multi-language interaction module, the early warning optimization, the system parameter automatic adjustment and the multi-language service are realized, and the system intelligent level and universality are improved.

[0082] The above is only an embodiment of the present application, and does not limit the patent range of the present application, and any equivalent structure or equivalent process transformation using the content of the present application specification and drawings, or direct or indirect application in other related technical fields, are also included in the patent protection range of the present application.

Claims

1. A precise and efficient visitor identity management system based on facial recognition, characterized in that: include: Facial recognition identity verification module: High-definition cameras equipped with active near-infrared technology are deployed at various entrances and key areas of the cruise ship to capture facial images; edge The computing node runs a convolutional neural network model to extract features and verifies the identity of visitors by comparing them with the visitor face database in the cloud server using Euclidean distance. Data access module: Utilizing standard network communication interfaces and TCP / IP, ModbusTCP, and HTTP / HTTPS protocols, it connects to cruise ship trajectory tracking, equipment operation monitoring, meteorological data, and personnel health management systems to acquire multi-source data; Passenger flow monitoring and risk alert module: Passenger flow monitoring equipment is installed in various recreational areas of the cruise ship, edge computing nodes integrate and analyze passenger flow data, and the server is equipped with a risk assessment model that combines multi-source data to assess risks; Immersive interactive guidance emergency assistance module: The official cruise app integrates location and route planning functions, using mobile phone location and cruise ship 3D map to determine the location of passengers; when an emergency is detected, it guides evacuation through multimodal interaction and AR technology; Cross-system collaborative emergency response module: It integrates the cruise ship's fire protection, medical, and security subsystems using a unified data interface, continuously monitors real-time data, analyzes abnormal data to identify emergency events, issues collaborative response instructions based on preset strategies, and tracks and adjusts them accordingly; Blockchain access control and data storage module: Based on the consortium blockchain model, a blockchain network is built, digital identities and public-private key pairs are allocated, smart contracts are generated to deploy access permissions, and visitor data is hashed and stored on the blockchain.

2. The accurate and efficient facial recognition visitor identity management system as described in claim 1, characterized in that: In the face recognition authentication module, the convolutional neural network model uses the ReLU activation function and is trained using the stochastic gradient descent algorithm.

3. The visitor identity management system with accurate and efficient facial recognition as described in claim 1, characterized in that: In the data access module, when interfacing with the personnel health management system, a combination of symmetric and asymmetric encryption is used to encrypt and transmit the data.

4. The visitor identity management system with accurate and efficient facial recognition as described in claim 1, characterized in that: In the passenger flow monitoring and risk warning module, the passenger flow analysis camera based on computer vision adopts the YOLO series algorithm, and the risk assessment model adopts one of the random forest algorithm and support vector machine algorithm. When the passenger flow exceeds or falls below the preset threshold, a prompt message is pushed through the APP. When the threshold is exceeded, flow restriction measures are implemented in the relevant area.

5. The visitor identity management system with accurate and efficient facial recognition as described in claim 1, characterized in that: In the immersive interactive guidance emergency assistance module, positioning within the cruise ship primarily relies on Bluetooth and Wi-Fi positioning, and the path planning algorithm employs Dijkstra's algorithm. One type of algorithm.

6. The visitor identity management system with accurate and efficient facial recognition as described in claim 1, characterized in that: The cross-system collaborative emergency response module uses either RESTful API or message queue technology for system integration, and employs data mining and machine learning algorithms for data fusion and analysis.

7. The visitor identity management system with accurate and efficient facial recognition as described in claim 1, characterized in that: The blockchain permission management and data storage module uses Hyperledger as the blockchain platform, the consensus mechanism is Practical Byzantine Fault Tolerance, and the data storage uses the SHA-256 hash algorithm.

8. The accurate and efficient facial recognition visitor identity management system as described in claim 1, characterized in that: It also includes a warning information push optimization module, which intelligently optimizes the push method and content according to different types of risk warning information to ensure that visitors receive and understand the warning information in a timely and accurate manner.

9. The accurate and efficient facial recognition visitor identity management system as described in claim 1, characterized in that: It also includes an intelligent learning and adaptive adjustment module, which automatically optimizes the parameter settings of each module through continuous learning and analysis of historical visitor data, emergency event data and system operation data, in order to adapt to different operating scenarios and changes.

10. A visitor identity management system with accurate and efficient facial recognition as described in claim 1, characterized in that: It also includes a multilingual interaction module, which supports information display and interactive functions in multiple languages, making it convenient for visitors from different language backgrounds to use.

Images