Early warning safety protection method for double face recognition areas

Abstract

The invention discloses a double face recognition area early warning safety protection method, and relates to the technical field of face recognition, a face recognition camera, a database, a background server and a client are included, and the method has the advantages that under strong light, an intelligent light adjustment algorithm can automatically reduce the exposure, improve the contrast and enhance the color saturation; the face features are highlighted, the overexposure phenomenon of the face is avoided, under the backlight condition, the backlight compensation algorithm is started, the overall brightness of the face is improved by increasing the edge brightness and adjusting the hue, the accuracy of feature extraction is ensured, and in addition, based on the face detection technology of the deep neural network, through deep training of massive face image data, the accuracy of feature extraction is improved. According to the invention, the camera can learn the face features in various complex environments, and even if the face is partially shielded and is located at different angles or under poor light conditions, the system can accurately carry out detection and recognition, so that the stability and accuracy of face recognition are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of face recognition, and in particular to a dual face recognition area early warning security prevention method. Background Art

[0002] In today's society, various places have increasingly higher requirements for security precautions, and a more reliable and accurate face recognition technology is needed to cope with complex and changing environments and security challenges. Traditional face recognition technology often seems powerless in the face of these problems and cannot timely and effectively discover and warn of potential security threats. Therefore, developing a dual face recognition regional warning security precaution method that can overcome the defects of existing technologies has important practical significance and application value. This method can improve the accuracy and stability of face recognition, enhance the effect of security precautions, and provide more powerful protection for people's life and property safety.

[0003] The existing face recognition regional warning security prevention technology has certain defects. First, it is difficult to obtain clear face images under low light conditions. It is easy to overexposure or shadow under strong light and backlight, resulting in inaccurate extraction of face features, thus affecting the recognition effect. Then, the existing technology usually adopts a single face recognition mode, which cannot achieve targeted security monitoring, has low warning efficiency, is difficult to detect potential security threats in time, lacks accuracy in identifying security risks, and is difficult to accurately extract personnel behavior patterns and laws. Secondly, the existing security prevention systems are independent of each other, lack effective integration, have poor scalability and linkage, cannot form a collaborative defense security prevention system, and are difficult to deal with complex security situations. Finally, the database storage architecture and data sharding technology of the existing technology are imperfect, resulting in limited data storage capacity and access speed, insufficient data encryption and backup functions, and difficult to ensure data security. At the same time, the processing power and load balancing capabilities of the background server are weak, and the system stability is difficult to guarantee. To this end, we propose a dual face recognition regional warning security prevention method. Summary of the invention

[0004] The purpose of the present invention is to provide a dual face recognition area early warning security prevention method.

[0005] In order to solve the problems raised in the above background technology, the present invention provides the following technical solutions: a dual face recognition area early warning security prevention method, including a face recognition camera, a database, a backend server and a client, and the dual face recognition area early warning security prevention method includes the following specific steps:

[0006] Step 1: Divide the specific area and the non-specific area, and deploy face recognition cameras with anti-light interference capabilities and multi-angle face detection algorithms in both the specific area and the non-specific area to collect face image data. The face recognition camera can dynamically adjust the acquisition frame rate and extract facial feature data;

[0007] Step 2: Using a face recognition camera to capture face images in real time, and store the facial feature information of the identified face in a specific area in a database equipped with data encryption and backup functions;

[0008] Step 3: The backend server is connected to the face recognition cameras in the specific area and the non-specific area in real time, and interacts with the database in real time, and can analyze and process the face image data obtained by the face recognition camera in real time;

[0009] Step 4: When the facial features captured in the non-specific area do not appear in the specific area, the backend server will extract the database data for feature comparison and use the intelligent analysis algorithm to predict the movement trajectory of the person. When the movement trajectory is abnormal, the early warning mechanism will quickly send an alarm message;

[0010] Step 5: The triggered warning information will be received by the visualization interface and the management client, so that the management can take corresponding measures in time.

[0011] As a further solution of the present invention: the face recognition camera is provided with a face feature recognition unit, and the face feature recognition unit is provided with a face detection technology based on a deep neural network. Through deep training of massive face image data, the camera can learn face features in various environments, and automatically extract multi-level fine features of the face by using a convolutional neural network (CNN) structure, thereby improving the accuracy of face detection in complex environments. At the same time, infrared fill light technology is used to ensure that the camera can capture clear face images under low light conditions, reducing the interference of light on face recognition. The face recognition camera also has the function of dynamically adjusting the frame rate, and automatically adjusts the current acquisition frame rate according to the movement and changes of the people in the scene. The specific calculation formula of the current acquisition frame rate is as follows:

[0012] f=f0+k·(N-N0)

[0013] Where: f represents the current acquisition frame rate, f0 represents the initial frame rate, k represents the adjustment coefficient, N represents the activity parameters of the personnel in the current scene, and N0 represents the baseline personnel parameters preset by the management personnel according to the current scene;

[0014] The frame rate is automatically reduced when there is less human activity. When people move quickly or abnormal situations occur, the frame rate is increased to ensure that key information is captured, improving resource utilization efficiency and recognition accuracy.

[0015] As a further solution of the present invention: the database adopts a distributed storage architecture and data sharding technology to disperse and store facial feature information on multiple nodes, thereby improving data storage capacity and access speed, and uses the Advanced Encryption Standard (AES) encryption algorithm to encrypt the stored data, and performs all-round encryption processing on the facial feature information data to prevent the data from being illegally stolen and tampered with. The database also has automatic data backup and rapid recovery functions, and backs up data to different secure storage media and remote servers according to different time intervals. The specific data backup time interval calculation formula is as follows:

[0016] T=T0·log N

[0017] Where: T represents the time interval for data backup, T0 represents the initial time interval, and N represents the amount of data;

[0018] When the database encounters hardware failure or data corruption, the data can be restored to ensure the normal operation of the system.

[0019] As a further solution of the present invention: the background server adopts a multi-core processor and parallel computing technology, can process multiple face recognition tasks at the same time, and uses a deep neural network (DNN) algorithm to perform feature comparison on face data. The algorithm can automatically learn the fine feature patterns of the face and optimize the model parameters through a large amount of training data, thereby improving the accuracy and efficiency of the comparison. The background server has an intelligent load balancing function, and monitors the server load in real time through a function, including CPU utilization and memory usage. The specific function calculation formula is as follows:

[0020] L = f(C, M)

[0021] Among them, L represents the load balancing distribution result, C represents CPU utilization, and M represents memory usage. It automatically and intelligently allocates computing resources according to the load of the background server, and assigns tasks to nodes with lighter loads for processing to ensure the stability of the system.

[0022] As a further solution of the present invention: the early warning mechanism adopts real-time data analysis and prediction technology to perform behavior analysis and trajectory prediction on unrecognized faces appearing in non-specific areas. By using big data analysis technology, a large amount of historical data and real-time data are deeply mined and analyzed to extract the behavior patterns and rules of personnel. Then, cluster analysis and association rule algorithms are used to identify potential security risks. The time series prediction model in the machine learning algorithm is used to predict the trajectory of personnel. The trajectory prediction is calculated by the following model:

[0023]

[0024] Where: P(t) represents the predicted trajectory at time t, t represents the specific face recognition time, n represents the number of data points obtained, X i (t) represents the face moving speed on the horizontal axis at time t, Y i (t) represents the face movement speed on the ordinate at time t, α and β represent the X obtained by the model based on historical data training. i (t) and Y i (t) weight parameter;

[0025] When an abnormal situation is detected, an early warning message is immediately issued. The early warning message includes captured video clips, data analysis reports and response suggestions to provide decision-making support for managers.

[0026] As a further solution of the present invention: the client adopts a visual interface and interactive operation design, and displays system data through charts, images and data, so that managers can view early warning data conveniently. The client is provided with remote control technology, so that managers can remotely control face recognition cameras through 5G networks to achieve real-time monitoring of different areas. The client also has intelligent voice recognition and fast intelligent response functions. Managers can interact with the client through voice commands, and the client can understand voice commands and respond accordingly. When the manager says "view the monitoring screen of the region-specific area" and "view the monitoring screen of the non-region-specific area", the client can automatically switch to the monitoring screen of the region-specific area and the non-specific area.

[0027] As a further solution of the present invention: the face recognition camera and the background server interact with each other using wireless communication technology, including 5G and WI FI 6 communication technology, to ensure that data can be transmitted quickly and stably, and adopt the LZ77 algorithm based on dictionary coding and the lossless compression algorithm based on Huffman coding. The LZ77 algorithm is used to find repeated character strings and use pointers to replace them to achieve compression. Huffman coding is used to assign codes of different lengths to different characters according to the frequency of their occurrence, thereby further improving the compression ratio and effectively reducing the amount of data without losing image quality.

[0028] As a further solution of the present invention: the face recognition camera is also provided with a multimodal face enhancement unit, which combines high dynamic range imaging (HDR) technology with an intelligent light adjustment algorithm. The HDR technology obtains image information under different brightness through multiple exposures, and then fuses these images, so that there is no overexposure under strong light, and the facial details can be clearly presented even under backlight. At the same time, the intelligent light adjustment algorithm analyzes the light intensity and direction in real time, and dynamically adjusts the exposure time and sensitivity parameters of image acquisition. When strong light is detected, the algorithm automatically reduces the exposure, improves the contrast, and enhances the color saturation to highlight the facial features. In the case of backlight, the algorithm enables the backlight compensation algorithm to increase the overall brightness of the face by increasing the edge brightness and adjusting the color tone, thereby ensuring the accuracy of feature extraction.

[0029] As a further solution of the present invention: the dual face recognition area warning security prevention method also includes an intelligent interaction unit, which is provided with an open architecture and a standardized interface, and can be seamlessly integrated with other security prevention systems. It adopts a general API interface and ONVI F protocol, and is integrated with the video surveillance system, the access control system and the alarm system through the API interface and the ONVI F protocol to achieve data sharing and interaction. The intelligent interaction unit has data sharing and linkage functions, and can perform real-time data interaction and collaborative work with other systems to form a security prevention system. When unauthorized entry is detected, the access control system can be automatically linked, the relevant channels can be closed, and alarm information can be sent to the alarm system at the same time to achieve collaborative defense between the systems.

[0030] By adopting the above technical solution, compared with the prior art, the beneficial effects of the present invention are:

[0031] 1. The present invention can play an important role in low-light conditions through the infrared fill light technology equipped with the face recognition camera. When the ambient light is insufficient, the infrared fill light technology can be automatically started to provide an additional light source for the camera, so that the camera can capture clear face images. At the same time, combined with the HDR technology in the multimodal face enhancement unit, image information under different brightness is obtained and fused through multiple exposures, and the intelligent light adjustment algorithm analyzes the light intensity and direction in real time, and dynamically adjusts the exposure time, sensitivity and other parameters of image acquisition. Under strong light, the intelligent light adjustment algorithm will automatically reduce the exposure, improve the contrast, and enhance the color saturation, highlight the face features, and avoid overexposure of the face. In the case of backlight, the algorithm activates the backlight compensation algorithm, and improves the overall brightness of the face by increasing the edge brightness and adjusting the color tone, ensuring the accuracy of feature extraction. In addition, the face detection technology based on the deep neural network, after deep training of massive face image data, enables the camera to learn the face features in various complex environments. Even if the face is partially blocked, at different angles or in poor light conditions, the system can accurately detect and identify, greatly improving the stability and accuracy of face recognition;

[0032] 2. The present invention realizes more targeted security monitoring by adopting a face recognition mode combining specific areas with non-specific areas. Face recognition cameras with anti-light interference capabilities and multi-angle face detection algorithms are deployed in both specific areas and non-specific areas, which can comprehensively and accurately collect face image data. When the facial features captured in the non-specific area do not appear in the specific area, the background server will quickly extract database data for feature comparison and trigger the early warning mechanism. This mode effectively improves the early warning efficiency and can timely discover potential security threats. At the same time, the early warning mechanism adopts real-time data analysis and prediction technology to perform behavioral analysis and trajectory prediction on unrecognized faces. By using big data analysis technology and machine learning algorithms, it can deeply mine and analyze a large amount of historical data and real-time data, extract personnel's behavior patterns and laws, and thus more accurately identify potential security risks. When an abnormal situation is detected, an early warning information including intercepted video clips, data analysis reports and response suggestions is immediately issued, providing comprehensive decision-making support for managers and further enhancing security prevention capabilities;

[0033] 3. The present invention is provided with an open architecture and a standardized interface through an intelligent interactive unit, and can be seamlessly integrated with other security systems. By adopting a universal API interface and ONVI F protocol, data sharing and interaction are achieved with a video surveillance system, an access control system, and an alarm system. It has a powerful linkage function. When an unauthorized person is detected entering, the access control system can be automatically linked to close the relevant channels, and an alarm message can be sent to the alarm system at the same time, forming a collaborative defense security system. The strong scalability and linkage greatly improve the overall security prevention capability and the ability to respond to emergencies.

[0034] 4. The present invention improves data storage capacity and access speed by adopting distributed storage architecture and data sharding technology in the database, and at the same time uses advanced encryption standards to encrypt data to ensure data security. The automatic data backup and rapid recovery functions ensure that the system can quickly resume normal operation when encountering a failure. The background server adopts multi-core processors and parallel computing technology, which can handle multiple face recognition tasks at the same time. The intelligent load balancing function automatically allocates computing resources according to the server load situation to ensure stable and reliable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Schematic diagram of the method steps in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0037] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0038] Embodiment 1:

[0039] In today's society, various places have increasingly higher requirements for security precautions, and a more reliable and accurate face recognition technology is needed to cope with complex and changing environments and security challenges. Traditional face recognition technology often seems powerless in the face of these problems and cannot timely and effectively discover and warn of potential security threats. Therefore, developing a dual face recognition regional warning security precaution method that can overcome the defects of existing technologies has important practical significance and application value. This method can improve the accuracy and stability of face recognition, enhance the effect of security precautions, and provide more powerful protection for people's life and property safety.

[0040] The existing face recognition regional warning security prevention technology has certain defects. First, it is difficult to obtain clear face images under low light conditions. It is easy to overexposure or shadow under strong light and backlight, resulting in inaccurate extraction of face features, thus affecting the recognition effect. Then, the existing technology usually adopts a single face recognition mode, which cannot achieve targeted security monitoring, has low warning efficiency, is difficult to detect potential security threats in time, lacks accuracy in identifying security risks, and is difficult to accurately extract personnel behavior patterns and laws. Secondly, the existing security prevention systems are independent of each other, lack effective integration, have poor scalability and linkage, cannot form a collaborative defense security prevention system, and are difficult to deal with complex security situations. Finally, the database storage architecture and data sharding technology of the existing technology are imperfect, resulting in limited data storage capacity and access speed, insufficient data encryption and backup functions, and difficult to ensure data security. At the same time, the processing power and load balancing capabilities of the background server are weak, and the system stability is difficult to guarantee. To this end, we propose a dual face recognition regional warning security prevention method.

[0041] Therefore, in order to effectively solve the above problems, the present application proposes a dual face recognition area warning security prevention method, as shown in the accompanying drawings of the specification. Figure 1 As shown, including a face recognition camera, a database, a backend server and a client, the double face recognition area early warning security prevention method includes the following specific steps:

[0042] Step 1: Divide the specific area and the non-specific area, and deploy face recognition cameras with anti-light interference capabilities and multi-angle face detection algorithms in both the specific area and the non-specific area to collect face image data. The face recognition camera can dynamically adjust the acquisition frame rate and extract facial feature data;

[0043] Step 2: Using a face recognition camera to capture face images in real time, and store the facial feature information of the identified face in a specific area in a database equipped with data encryption and backup functions;

[0044] Step 3: The backend server is connected to the face recognition cameras in the specific area and the non-specific area in real time, and interacts with the database in real time, and can analyze and process the face image data obtained by the face recognition camera in real time;

[0045] Step 4: When the facial features captured in the non-specific area do not appear in the specific area, the backend server will extract the database data for feature comparison and use the intelligent analysis algorithm to predict the movement trajectory of the person. When the movement trajectory is abnormal, the early warning mechanism will quickly send an alarm message;

[0046] Step 5: The triggered warning information will be received by the visualization interface and the management client, so that the management can take corresponding measures in time;

[0047] Specific workflow: A dual face recognition area warning security prevention method mainly consists of face recognition cameras, databases, background servers and clients. First, specific areas and non-specific areas are clearly divided, and powerful face recognition cameras are deployed in both areas. These cameras not only have the ability to resist light interference and can work stably under various complex light conditions, but also have multi-angle face detection algorithms, which can accurately collect face image data from different angles. In addition, the camera can dynamically adjust the acquisition frame rate according to actual conditions to ensure efficient data acquisition and accurate extraction of facial feature data. The background server maintains real-time connection with face recognition cameras in specific areas and non-specific areas, and interacts with the database in real time, so that the face image data obtained by the face recognition camera can be analyzed and processed in real time. When the facial features captured in the non-specific area do not appear in the specific area, the background server will quickly extract the database data for feature comparison and immediately trigger the early warning mechanism. The triggered early warning information will be transmitted to the visualization interface and the manager's client in a timely manner, so that the manager can quickly know and take corresponding measures.

[0048] Furthermore, with the help of the infrared fill light technology equipped with the face recognition camera, it can play an effective role in low-light conditions. When the ambient light is insufficient, the infrared fill light technology will automatically start to provide the camera with additional light source to ensure that it can capture clear facial images. At the same time, combined with the HDR technology in the multimodal face enhancement unit, image information at different brightness is obtained and fused through multiple exposures, and the intelligent light adjustment algorithm analyzes the light intensity and direction in real time, and dynamically adjusts the exposure time, sensitivity and other parameters of image acquisition. In strong light, the intelligent light adjustment algorithm will automatically reduce the exposure, increase the contrast and enhance the color saturation to highlight the facial features and avoid overexposure. In the case of backlight, the algorithm will enable the backlight compensation algorithm to increase the overall brightness of the face by increasing the edge brightness and adjusting the color tone to ensure the accuracy of feature extraction. In addition, the face detection technology based on deep neural networks, after deep training of massive face image data, enables the camera to learn facial features in various complex environments. Even if the face is partially blocked, at different angles or in poor lighting conditions, the system can accurately detect and identify, greatly improving the stability and accuracy of face recognition.

[0049] Embodiment 2:

[0050] Based on the first embodiment, as shown in the accompanying drawings of the specification Figure 1As shown, the face recognition camera is provided with a face feature recognition unit, which is provided with face detection technology based on a deep neural network. Through deep training of massive face image data, the camera can learn face features in various environments, and automatically extract multi-level fine features of the face by using a convolutional neural network (CNN) structure, thereby improving the accuracy of face detection in complex environments. At the same time, infrared fill light technology is used to ensure that the camera can capture clear face images under low light conditions, reducing the interference of light on face recognition. The face recognition camera also has the function of dynamically adjusting the frame rate, and automatically adjusts the current acquisition frame rate according to the movement and changes of the people in the scene. The specific calculation formula of the current acquisition frame rate is as follows:

[0051] f=f0+k·(N-N0)

[0052] Where: f represents the current acquisition frame rate, f0 represents the initial frame rate, k represents the adjustment coefficient, N represents the activity parameters of the personnel in the current scene, and N0 represents the baseline personnel parameters preset by the management personnel according to the current scene;

[0053] The frame rate is automatically reduced when there are fewer people moving around. When people move quickly or abnormal situations occur, the frame rate is increased to ensure that key information is captured, improve resource utilization efficiency and recognition accuracy. The database uses a distributed storage architecture and data sharding technology to store facial feature information in multiple nodes to increase data storage capacity and access speed. The Advanced Encryption Standard (AES) encryption algorithm is used to encrypt the stored data, and facial feature information data is fully encrypted to prevent data from being illegally stolen and tampered with. The database also has automatic data backup and rapid recovery functions, and backs up data to different secure storage media and remote servers at different time intervals. The specific data backup time interval calculation formula is as follows:

[0054] T=T0·log N

[0055] Where: T represents the time interval for data backup, T0 represents the initial time interval, and N represents the amount of data;

[0056] When the database encounters hardware failure or data corruption, it can restore data to ensure the normal operation of the system. The backend server uses multi-core processors and parallel computing technology to handle multiple face recognition tasks at the same time, and uses the deep neural network (DNN) algorithm to perform feature comparison on face data. The algorithm can automatically learn the fine feature patterns of the face and optimize the model parameters through a large amount of training data, thereby improving the accuracy and efficiency of the comparison. The backend server has an intelligent load balancing function, which monitors the server load in real time through functions, including CPU utilization and memory usage. The specific function calculation formula is as follows:

[0057] L = f(C, M)

[0058] Among them, L represents the load balancing distribution result, C represents CPU utilization, and M represents memory usage. It automatically and intelligently allocates computing resources according to the load of the background server, and assigns tasks to nodes with lighter loads for processing to ensure the stability of the system. The early warning mechanism uses real-time data analysis and prediction technology to perform behavioral analysis and trajectory prediction on unrecognized faces that appear in non-specific areas. By using big data analysis technology, a large amount of historical data and real-time data are deeply mined and analyzed to extract the behavior patterns and rules of personnel. Then, cluster analysis and association rule algorithms are used to identify potential security risks. The time series prediction model in the machine learning algorithm is used to predict the trajectory of personnel. The trajectory prediction is calculated through the following model:

[0059]

[0060] Where: P(t) represents the predicted trajectory at time t, t represents the specific face recognition time, n represents the number of data points obtained, X i (t) represents the face moving speed on the horizontal axis at time t, Y i (t) represents the face movement speed on the ordinate at time t, α and β represent the X obtained by the model based on historical data training. i (t) and Y i (t) weight parameter;

[0061] When an abnormal situation is detected, an early warning message is immediately issued, which includes intercepted video clips, data analysis reports and response suggestions to provide decision-making support for managers;

[0062] Specific workflow: Face recognition cameras use deep neural networks to detect faces, and use infrared fill light technology and multimodal face enhancement units to ensure accurate recognition under different lighting conditions. They also have the function of dynamically adjusting the frame rate. Face recognition cameras can add face liveness detection technology in actual applications. This technology uses an algorithm model to make judgments by deeply analyzing multi-dimensional features such as face texture, skin color, and expression. It detects subtle features such as changes in facial micro-expressions, gloss and elasticity of the skin, to identify whether the face is real and alive, effectively preventing deception methods such as photos and videos, and can greatly improve the security and accuracy of face recognition.

[0063] The database adopts distributed storage and data sharding technology, uses the AES encryption algorithm, and automatically calculates the backup time interval according to the amount of data to ensure data security and rapid recovery. The database can use advanced blockchain technology for data storage according to different application scenarios. Blockchain has unique characteristics such as decentralization and non-tamperability. During the data storage process, the blockchain stores data in multiple nodes to form a distributed ledger to ensure data security and integrity. At the same time, any modification to the data needs to be verified by multiple nodes, making it almost impossible for the data to be tampered with, improving the credibility and traceability of the data. In addition, the encryption technology of the blockchain can also effectively protect the privacy of the data and prevent data leakage;

[0064] The backend server uses multi-core processors and parallel computing technology, uses DNN algorithms for feature comparison, monitors load conditions through functions, and intelligently allocates computing resources to ensure stability. The backend server can introduce an artificial intelligence-assisted decision-making system, which can deeply analyze face recognition data and historical warning information. By using machine learning algorithms and big data analysis technology, it can mine and analyze large amounts of data, extract valuable information and patterns, and predict potential security risks based on face recognition data in different time periods and regions, and provide corresponding intelligent decision-making suggestions. These suggestions can help managers make more accurate decisions faster, improve emergency response capabilities, and respond to various security threats in a timely and effective manner;

[0065] The early warning mechanism conducts behavioral analysis and trajectory prediction on unrecognized faces in non-specific areas, and uses big data analysis and machine learning algorithms to issue early warning information when abnormalities are detected. The early warning mechanism can be closely linked with patrol personnel. When abnormal situations are detected, the early warning system can quickly send instructions to patrol personnel through calling devices, and can quickly go to the scene of the incident, and feedback relevant information to management personnel in real time, so that further treatment measures can be taken in time to improve the timeliness of protection;

[0066] Furthermore, a face recognition mode combining specific areas and non-specific areas is adopted to achieve more targeted security monitoring. Face recognition cameras with anti-light interference capabilities and multi-angle face detection algorithms are deployed in both specific areas and non-specific areas, which can comprehensively and accurately collect face image data. When the facial features captured in the non-specific area do not appear in the specific area, the background server will quickly extract the database data for feature comparison and trigger the early warning mechanism. This mode effectively improves the early warning efficiency and can timely detect potential security threats. At the same time, the early warning mechanism uses real-time data analysis and prediction technology to perform behavioral analysis and trajectory prediction on unrecognized faces. By using big data analysis technology and machine learning algorithms, it can deeply mine and analyze a large amount of historical data and real-time data, extract personnel behavior patterns and laws, and thus more accurately identify potential security risks. When an abnormal situation is detected, it will immediately issue a warning information including intercepted video clips, data analysis reports and response suggestions, providing comprehensive decision-making support for managers and further enhancing security prevention capabilities.

[0067] Embodiment three:

[0068] Based on the second embodiment, as shown in the accompanying drawings of the specification Figure 1As shown, the client adopts a visual interface and interactive operation design, and displays system data through charts, images and data, which is convenient for managers to view early warning data. The client is equipped with remote control technology, allowing managers to remotely control face recognition cameras through 5G networks to achieve real-time monitoring of different areas. The client also has intelligent voice recognition and fast intelligent response functions. Managers can interact with the client through voice commands, and the client can understand voice commands and respond accordingly. When the manager says "view the monitoring screen of the specific area of ​​the region" and "view the monitoring screen of the non-specific area of ​​the region", the client can automatically switch to the monitoring screen of the specific area and the non-specific area of ​​the region. Wireless communication technology is used to interact between the face recognition camera and the background server, including 5G and WI FI 6 communication technology to ensure fast and stable data transmission. It adopts the LZ77 algorithm based on dictionary coding and the lossless compression algorithm based on Huffman coding. The LZ77 algorithm is used to find repeated strings and replace them with pointers to achieve compression. Huffman coding is used to assign different lengths of codes to different characters according to the frequency of their occurrence, thereby further improving the compression ratio and effectively reducing the amount of data without losing image quality. The face recognition camera is also equipped with a multi-modal face enhancement unit, which combines high dynamic range imaging (HDR) technology with an intelligent light adjustment algorithm. HDR technology obtains image information at different brightness levels through multiple exposures, and then fuses these images to make the image more vivid. There will be no overexposure under strong light, and the facial details can be clearly presented under backlight. At the same time, the intelligent light adjustment algorithm analyzes the light intensity and direction in real time, and dynamically adjusts the exposure time and sensitivity parameters of image acquisition. When strong light is detected, the algorithm automatically reduces the exposure, improves the contrast, and enhances the color saturation to highlight the facial features. In the case of backlight, the algorithm enables the backlight compensation algorithm to increase the overall brightness of the face by increasing the edge brightness and adjusting the color tone to ensure the accuracy of feature extraction. The dual face recognition area warning security prevention method also includes an intelligent interaction unit. The intelligent interaction unit is provided with an open architecture and a standardized interface, and can be seamlessly integrated with other security prevention systems. It adopts a general API interface and ONVI F protocol, and is integrated with the video surveillance system, the access control system and the alarm system through the API interface and the ONVI F protocol to realize data sharing and interaction. The intelligent interaction unit has data sharing and linkage functions, and can perform real-time data interaction and collaborative work with other systems to form a security prevention system. When unauthorized personnel are detected entering, the access control system can be automatically linked to close the relevant channels, and an alarm message is sent to the alarm system at the same time to realize collaborative defense between the systems.

[0069] Specific workflow: The client displays system data through a visual interface, has remote control technology and intelligent voice recognition functions, and can switch monitoring screens according to voice commands. The client can add advanced face recognition unlocking function, which is based on a high-precision face recognition algorithm. Users only need to point their faces at the camera, and the system can quickly identify the user's identity. After the verification is passed, the client will automatically unlock and provide relevant information to the user, greatly improving the convenience and safety of operation;

[0070] The face recognition camera and the backend server use wireless communication technology, and use LZ77 and Huffman coding algorithms for data compression transmission to ensure fast and stable data transmission. Quantum encryption communication technology can be introduced between the face recognition camera and the backend server. This technology makes full use of the unique characteristics of quantum mechanics to ensure that the security and confidentiality of communication reach an extremely high level, and can effectively prevent data from being illegally stolen or tampered with;

[0071] The multimodal face enhancement unit in the face recognition camera combines HDR technology and intelligent light adjustment algorithm to ensure clear presentation of facial details and accurate feature extraction under different lighting conditions. The face recognition camera can be equipped with an emotion recognition module as needed. This module uses advanced expression analysis algorithms to analyze facial expression features in detail. By capturing and analyzing expression changes in key parts such as eyes and mouth, it can accurately determine the emotional state of people. When abnormal emotions such as tension and anxiety are detected, the system will issue early warning prompts to remind managers to pay close attention. This function helps to discover potential safety hazards in advance.

[0072] The intelligent interactive unit is integrated with other security systems through an open architecture and standardized interfaces to achieve data sharing and linkage. When unauthorized personnel are detected, the relevant systems are automatically linked. The intelligent interactive unit can be deeply integrated with the intelligent security system to achieve comprehensive remote control and management of security equipment. Through the client, users can remotely and accurately control the angle, focal length and other parameters of the camera, adjust the monitoring range in real time, and use intelligent algorithms to automatically adjust the direction of the camera to ensure that key areas are always within the monitoring range. At the same time, the monitoring screens of different cameras can be flexibly switched according to the real-time situation, which improves the flexibility and pertinence of security prevention.

[0073] Furthermore, the intelligent interactive unit is equipped with an open architecture and standardized interfaces, which can be seamlessly integrated with other security systems. By adopting a universal API interface and ONVI F protocol, it can share and interact with video surveillance systems, access control systems and alarm systems. It has a powerful linkage function. When unauthorized entry is detected, it can automatically link the access control system to close the relevant channels and send alarm information to the alarm system at the same time, forming a collaborative defense security system with strong scalability and linkage, which greatly improves the overall security prevention capabilities and the ability to respond to emergencies. The database adopts a distributed storage architecture and data sharding technology to improve data storage capacity and access speed. At the same time, it uses advanced encryption standards to encrypt data to ensure data security. The automatic data backup and rapid recovery functions ensure that the system can quickly resume normal operation when encountering a fault. The background server adopts a multi-core processor and parallel computing technology, which can process multiple face recognition tasks at the same time. The intelligent load balancing function can automatically allocate computing resources according to the server load to ensure stable and reliable operation of the system.

[0074] At the same time, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.

[0075] Some aspects of the present application may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may be referred to as "data blocks", "modules", "engines", "units", "components" or "systems". The processor may be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DAPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors or combinations thereof. In addition, various aspects of the present application may be expressed as computer products located in one or more computer-readable media, which include computer-readable program codes. For example, computer-readable media may include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, tapes ...), optical disks (e.g., compact disks CDs, digital versatile disks DVDs ...), smart cards, and flash memory devices (e.g., cards, sticks, key drives ...).

[0076] A computer-readable medium may include a propagated data signal containing computer program code, such as in baseband or as part of a carrier wave. The propagated signal may have a variety of manifestations, including electromagnetic, optical, etc., or a suitable combination. A computer-readable medium may be any computer-readable medium other than a computer-readable storage medium, which may be connected to an instruction execution system, device or apparatus to communicate, propagate or transmit a program for use. The program code on the computer-readable medium may be propagated through any suitable medium, including radio, cable, fiber optic cable, radio frequency signal, or similar medium, or a combination of any of the above mediums.

[0077] Similarly, it should be noted that in order to simplify the description disclosed in this application, so as to help understand one or more embodiments of the invention, in the above description of the embodiments of the present application, sometimes multiple features are merged into one embodiment, drawings or descriptions thereof. However, this disclosure method does not mean that the features required by the object of the present application are more than the features mentioned in the claims. In fact, the features of the embodiment are less than all the features of the single embodiment disclosed above. In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "approximately", "approximately" or "substantially" in some examples. Unless otherwise specified, "approximately", "approximately" or "substantially" indicate that the number allows a variation of ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the characteristics required by individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining the digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of the present application are approximate values, in specific embodiments, such numerical settings are as accurate as possible within the feasible range.

[0078] Although the present invention is disclosed as above in terms of preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art may make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the scope of protection defined by the claims of the present invention.

Claims

1. A dual face recognition area early warning security prevention method, characterized by: The method comprises a face recognition camera, a database, a backend server and a client, and the dual face recognition area early warning security prevention method comprises the following specific steps: Step 1: Divide the specific area and the non-specific area, and deploy face recognition cameras with anti-light interference capabilities and multi-angle face detection algorithms in both the specific area and the non-specific area to collect face image data. The face recognition camera can dynamically adjust the acquisition frame rate and extract facial feature data; Step 2: Using a face recognition camera to capture face images in real time, and store the facial feature information of the identified face in a specific area in a database equipped with data encryption and backup functions; Step 3: The backend server is connected to the face recognition cameras in the specific area and the non-specific area in real time, and interacts with the database in real time, so as to be able to analyze and process the face image data obtained by the face recognition camera in real time; Step 4: When the facial features captured in the non-specific area do not appear in the specific area, the backend server will extract the database data for feature comparison and use the intelligent analysis algorithm to predict the movement trajectory of the person. When the movement trajectory is abnormal, the early warning mechanism will quickly send an alarm message; Step 5: The triggered warning information will be received by the visualization interface and the management client, so that the management can take corresponding measures in time.

2. A dual face recognition area early warning security prevention method according to claim 1, characterized in that: The face recognition camera is provided with a face feature recognition unit, which is provided with a face detection technology based on a deep neural network. Through deep training of massive face image data, the camera can learn face features in various environments, and automatically extract multi-level fine features of the face by using a convolutional neural network (CNN) structure, thereby improving the accuracy of face detection in complex environments. At the same time, infrared fill light technology is used to ensure that the camera can capture clear face images under low light conditions, reducing the interference of light on face recognition. The face recognition camera also has the function of dynamically adjusting the frame rate, and automatically adjusts the current acquisition frame rate according to the movement and changes of the people in the scene. The specific calculation formula of the current acquisition frame rate is as follows: f=f0+k·(N-N0) Where: f represents the current acquisition frame rate, f0 represents the initial frame rate, k represents the adjustment coefficient, N represents the activity parameters of the personnel in the current scene, and N0 represents the baseline personnel parameters preset by the management personnel according to the current scene; The frame rate is automatically reduced when there is less human activity. When people move quickly or abnormal situations occur, the frame rate is increased to ensure that key information is captured, improving resource utilization efficiency and recognition accuracy.

3. The dual face recognition area warning security prevention method according to claim 1, characterized in that: The database adopts a distributed storage architecture and data sharding technology to store facial feature information in multiple nodes in a dispersed manner, thereby improving data storage capacity and access speed. The Advanced Encryption Standard (AES) encryption algorithm is used to encrypt the stored data, and facial feature information data is fully encrypted to prevent data from being illegally stolen and tampered with. The database also has automatic data backup and rapid recovery functions, and backs up data to different secure storage media and remote servers at different time intervals. The specific data backup time interval calculation formula is as follows: T=T0·log N Where: T represents the time interval for data backup, T0 represents the initial time interval, and N represents the amount of data; When the database encounters hardware failure or data corruption, the data can be restored to ensure the normal operation of the system.

4. The dual face recognition area early warning security prevention method according to claim 1, characterized in that: The backend server uses a multi-core processor and parallel computing technology, can handle multiple face recognition tasks at the same time, and uses a deep neural network (DNN) algorithm to perform feature comparison on face data. The algorithm can automatically learn the fine feature patterns of the face and optimize the model parameters through a large amount of training data, thereby improving the accuracy and efficiency of the comparison. The backend server has an intelligent load balancing function, and monitors the server load in real time through functions, including CPU utilization and memory usage. The specific function calculation formula is as follows: L=f(C,M) Among them, L represents the load balancing distribution result, C represents CPU utilization, and M represents memory usage. It automatically and intelligently allocates computing resources according to the load of the background server, and assigns tasks to nodes with lighter loads for processing to ensure the stability of the system.

5. The dual face recognition area early warning security prevention method according to claim 1, characterized in that: The early warning mechanism uses real-time data analysis and prediction technology to perform behavioral analysis and trajectory prediction on unrecognized faces that appear in non-specific areas. By using big data analysis technology, a large amount of historical data and real-time data are deeply mined and analyzed to extract the behavioral patterns and rules of personnel. Then, cluster analysis and association rule algorithms are used to identify potential security risks. The time series prediction model in the machine learning algorithm is used to predict the trajectory of personnel. The trajectory prediction is calculated using the following model: Where: P(t) represents the predicted trajectory at time t, t represents the specific face recognition time, n represents the number of data points obtained, X i (t) represents the face moving speed on the horizontal axis at time t, Y i (t) represents the face movement speed on the ordinate at time t, α and β represent the X obtained by the model based on historical data training. i (t) and Y i (t) weight parameter; When an abnormal situation is detected, an early warning message is immediately issued. The early warning message includes captured video clips, data analysis reports and response suggestions to provide decision-making support for managers.

6. The dual face recognition area early warning security prevention method according to claim 1, characterized in that: The client adopts a visual interface and interactive operation design, and displays system data through charts, images and data, making it convenient for managers to view early warning data. The client is equipped with remote control technology, so that managers can remotely control face recognition cameras through the 5G network to achieve real-time monitoring of different areas. The client also has intelligent voice recognition and fast intelligent response functions. Managers can interact with the client through voice commands, and the client can understand voice commands and respond accordingly. When the manager says "view the monitoring screen of the area-specific area" and "view the monitoring screen of the non-area-specific area", the client can automatically switch to the monitoring screen of the area-specific area and the non-specific area.

7. The dual face recognition area early warning security prevention method according to claim 1, characterized in that: The face recognition camera and the backend server interact with each other using wireless communication technology, including 5G and WIFI 6 communication technology, to ensure that data can be transmitted quickly and stably. The LZ77 algorithm based on dictionary coding and the lossless compression algorithm based on Huffman coding are adopted. The LZ77 algorithm is used to find repeated character strings and use pointers to replace them to achieve compression. Huffman coding is used to assign codes of different lengths to different characters according to the frequency of their occurrence, thereby further improving the compression ratio and effectively reducing the amount of data without losing image quality.

8. The dual face recognition area early warning security prevention method according to claim 2, characterized in that: The face recognition camera is also provided with a multimodal face enhancement unit, which combines high dynamic range imaging (HDR) technology with an intelligent light adjustment algorithm. The HDR technology obtains image information under different brightness through multiple exposures, and then fuses these images, so that there is no overexposure under strong light, and the facial details can be clearly presented even under backlight. At the same time, the intelligent light adjustment algorithm analyzes the light intensity and direction in real time, and dynamically adjusts the exposure time and sensitivity parameters of image acquisition. When strong light is detected, the algorithm automatically reduces the exposure, improves the contrast, and enhances the color saturation to highlight the facial features. In the case of backlight, the algorithm enables the backlight compensation algorithm to increase the overall brightness of the face by increasing the edge brightness and adjusting the color tone, thereby ensuring the accuracy of feature extraction.

9. The dual face recognition area early warning security prevention method according to claim 1, characterized in that: The dual face recognition area warning security prevention method also includes an intelligent interaction unit. The intelligent interaction unit is provided with an open architecture and a standardized interface, and can be seamlessly integrated with other security prevention systems. It adopts a general API interface and ONVIF protocol, and is integrated with a video surveillance system, an access control system, and an alarm system through the API interface and the ONVIF protocol to achieve data sharing and interaction. The intelligent interaction unit has data sharing and linkage functions, and can perform real-time data interaction and collaborative work with other systems to form a security prevention system. When unauthorized entry is detected, the access control system can be automatically linked, relevant channels can be closed, and alarm information can be sent to the alarm system at the same time to achieve collaborative defense between systems.

Images