A facial recognition and security system for smart parks
By constructing a first evaluation model and a second evaluation model, and using a convolutional neural network to process face image frames at long distances and large angles, the problem of low efficiency in long-distance face recognition is solved, and fast and accurate passage judgment is achieved.
Patent Information
- Application Number
- CN202511273401.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-08
- Publication Date
- 2025-12-02
- Estimated Expiration
- 2045-09-08
AI Technical Summary
Existing technologies have low efficiency in facial recognition at long distances and large angles, leading to increased waiting times for people to pass through, and there is a lack of effective long-distance facial recognition solutions.
A first evaluation model and a second evaluation model are constructed. By acquiring face image frames and their similarity at different acquisition distances and angles, a convolutional neural network is used for training. The first evaluation model is used to identify similarity, and the second evaluation model is used to determine the target object, thus achieving accurate recognition at long distances and large angles.
Achieving timely and effective facial recognition at long distances and large angles reduces waiting time for people and improves traffic efficiency.
Smart Images
Figure CN120766336B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of facial recognition technology, specifically a facial intelligent recognition and security protection system for smart parks. Background Technology
[0002] Using facial recognition technology to manage people entering and exiting the park is a comprehensive security management solution designed specifically for the security needs of modern parks. It combines facial recognition algorithms, big data analysis, and intelligent monitoring technology to accurately identify and dynamically manage people within the park, effectively improving park security and management efficiency.
[0003] Due to objective limitations in image acquisition, the clarity of facial images decreases with increasing acquisition distance and angle, making them less suitable for facial recognition. Consequently, existing technologies often begin facial recognition only when personnel are standing in front of the acquisition device, inevitably resulting in waiting time and reduced efficiency in personnel passage. Therefore, existing technologies lack a solution for initiating facial recognition from a distance. To address this, this invention provides a facial intelligent recognition and security system for smart parks. Summary of the Invention
[0004] The purpose of this invention is to provide a facial intelligent recognition and security protection system for smart parks.
[0005] The objective of this invention can be achieved through the following technical solution: a facial recognition and security system for smart parks, comprising the following modules:
[0006] The data acquisition module is used to acquire and store the personal and facial information of people within the park;
[0007] The face analysis module is used to set up the acquisition unit, acquire face image frames of a single person at different acquisition distances and acquisition angles, and obtain the face information of the single person and the face similarity between the different face image frames.
[0008] The first evaluation module is used to construct a first evaluation model based on the face image frames of different people at different acquisition distances and acquisition angles and their face similarity.
[0009] The second evaluation module is used to obtain the similarity set and similarity change characteristics between the face image frames of a single person at different acquisition distances and acquisition angles and the face information of different people. The second evaluation model is constructed based on the similarity change characteristics between different face image frames of different people and different face information.
[0010] The face recognition module is used to acquire recognition image frames, combine them with the first evaluation model to obtain the corresponding recognition similarity and its similarity change features, and combine them with the second evaluation model to determine whether it is passable.
[0011] Furthermore, the process of obtaining and storing the personal and facial information of people within the park includes:
[0012] The personal information refers to the name, gender, age, occupation, mobile phone number, ID number, and ID photo of relevant personnel within the park; the facial information refers to the facial image and its image features used to identify relevant personnel within the park.
[0013] A personnel database is set up to bind the personal information of relevant personnel in the park with their corresponding facial information, and both are uploaded to the personnel database for storage.
[0014] Furthermore, the process of setting up a data acquisition unit to obtain facial image frames of a single person at different acquisition distances and angles includes:
[0015] The acquisition unit refers to a device used to acquire facial images and videos, and to acquire facial image data of relevant personnel within its acquisition range in real time.
[0016] The acquisition distance refers to the distance between the face of the relevant person and the acquisition unit, and the acquisition angle refers to the angle of the face of the relevant person relative to the acquisition unit, including the horizontal angle and the vertical angle.
[0017] Each image frame in the facial image data acquired from a single person in a single instance is labeled as a facial image frame of that single person, and the acquisition distance and acquisition angle of each facial image frame at the corresponding time are obtained.
[0018] Furthermore, the process of obtaining the facial information of a single person and the facial similarity between different facial image frames includes:
[0019] Based on the image features in a single face image frame of a single person, pose correction is performed on the single face image frame to make the face in a frontal pose. The AlexNet algorithm is used to extract the feature vectors of the face image in the single face image frame and the face information of the single person.
[0020] The cosine similarity between the feature vectors of the two is used as the face similarity between the two, and the face similarity between the face image in the face information of the single person and each face image frame is obtained.
[0021] Furthermore, the process of constructing the first evaluation model based on face image frames and their face similarity from different individuals at different acquisition distances and angles includes:
[0022] Acquire facial image frames of different individuals at different acquisition distances and angles, and obtain the facial similarity between the facial images of different individuals and their corresponding different facial image frames;
[0023] The first evaluation set is generated based on the facial images of different individuals, their facial image frames at different acquisition distances and angles, and facial similarity, and is divided into the first training set and the first test set.
[0024] Construct a first convolutional neural network, using the face images of different people in the first training set and their different face image frames as the input data of the first convolutional neural network, and using the corresponding face similarity in the first training set as the output data of the first convolutional neural network.
[0025] The first convolutional neural network is trained to obtain an initial first convolutional neural network. The initial first convolutional neural network is validated using a first test set. The initial first convolutional neural network whose output is less than or equal to a preset first test error threshold is used as the first evaluation model.
[0026] Furthermore, the process of obtaining the similarity set and similarity variation features between facial image frames of a single person at different acquisition distances and angles and facial information of different people includes:
[0027] When the initial face image frame of a single person is judged to be similar to multiple face images stored in the person database during the face recognition process, each face image that is in a similar state to the single person will be marked as a similar object of the single person.
[0028] The "initial" refers to the earliest corresponding time in the facial image data of a single person acquired in a single instance, and the judgment criterion for the similarity state is that the facial similarity between the initial facial image frame of a single person and the corresponding facial image is greater than or equal to a preset similarity threshold.
[0029] The facial similarity between a single person and a single similar object is obtained by acquiring different facial image frames of a single person and sorting the acquired facial similarity according to the chronological order of the corresponding facial image frames to obtain the similarity set between the single person and the single similar object.
[0030] Obtain the similarity set between the individual and each of its similar objects, and obtain the corresponding similarity change features, including mean, variance, and trend. Then bind the similarity set between the individual and its different similar objects and their similarity change features.
[0031] Furthermore, the process of constructing a second evaluation model based on the similarity variation characteristics of different facial image frames and different facial information of different individuals includes:
[0032] Retrieve similar objects that are in the same state as a single person among all similar objects, and mark the similar objects in the same state as the target objects of that single person.
[0033] The criterion for judging the same state is that the latest face similarity in the similarity set of the corresponding similar objects is greater than or equal to the preset same threshold. The latest refers to the latest time in the face image data of a single person in a single acquisition.
[0034] A second evaluation set is generated based on the different similarity sets of different individuals and their various similar objects, as well as the corresponding similarity change characteristics, and is divided into a second training set and a second test set.
[0035] Construct a second convolutional neural network, using the different similarity sets of different people and their similar objects in the second training set as the input data of the second convolutional neural network, and the target object in the similar objects of different people in the second training set as the output data of the second convolutional neural network.
[0036] The second convolutional neural network is trained to obtain an initial second convolutional neural network. The initial second convolutional neural network is validated using a second test set. The initial second convolutional neural network whose output is less than or equal to a preset second test error threshold is used as the second evaluation model.
[0037] Furthermore, the process of acquiring the image frame for recognition, combining the first evaluation model to obtain the corresponding recognition similarity and its similarity change features, and combining the second evaluation model to determine whether it is feasible includes:
[0038] In subsequent application scenarios, when the acquisition unit identifies a single person within its acquisition range, it continuously acquires multiple recognition image frames of that single person.
[0039] The initial recognition image frame and each face image stored in the personnel database are respectively input into the first evaluation model to obtain the corresponding recognition similarity. The face images that are greater than or equal to the preset similarity threshold are taken as the similar objects of the single person.
[0040] Using the first evaluation model, the recognition similarity between each recognition image frame of the single person and its single similar object is obtained, and the corresponding similarity set and similarity change features are obtained. The similarity set and similarity change features between the single person and each of its similar objects are obtained.
[0041] The similarity set of the single person and each of their similar objects and the characteristics of their similarity changes are input into the second evaluation model to determine whether the single person's target object exists among the similar objects. If it exists, the single person is marked as passable; if it does not exist, it is marked as inaccessible.
[0042] Compared with the prior art, the beneficial effects of the present invention are:
[0043] This invention acquires facial image frames of the same person at different acquisition distances and angles, and obtains the facial similarity of each facial image frame based on the pre-stored facial information. Based on this, a first evaluation model can be constructed, which is beneficial for timely and effective facial recognition of faces at long distances or large angles in subsequent application scenarios, and obtaining their facial similarity.
[0044] By acquiring multiple similar objects of a single user at long distances or large angles, and obtaining the similarity set and similarity change features between the user and these similar objects, and based on the subtle differences between the target object of the single user and each similar object, a second evaluation model can be constructed to directly acquire the target object among multiple similar objects in subsequent application scenarios. This is beneficial for performing face recognition in advance at long distances and determining whether people can pass through, which can significantly reduce the waiting time for recognition and improve the efficiency of people passing through. Attached Figure Description
[0045] Figure 1 This is a schematic diagram of the modules of the present invention. Detailed Implementation
[0046] like Figure 1 As shown, a facial recognition and security system for smart parks includes the following modules:
[0047] The data acquisition module is used to acquire and store the personal and facial information of people within the park;
[0048] The face analysis module is used to set up the acquisition unit, acquire face image frames of a single person at different acquisition distances and acquisition angles, and obtain the face information of the single person and the face similarity between the different face image frames.
[0049] The first evaluation module is used to construct a first evaluation model based on the face image frames of different people at different acquisition distances and acquisition angles and their face similarity.
[0050] The second evaluation module is used to obtain the similarity set and similarity change characteristics between the face image frames of a single person at different acquisition distances and acquisition angles and the face information of different people. The second evaluation model is constructed based on the similarity change characteristics between different face image frames of different people and different face information.
[0051] The face recognition module is used to acquire recognition image frames, combine them with the first evaluation model to obtain the corresponding recognition similarity and its similarity change features, and combine them with the second evaluation model to determine whether it is passable.
[0052] It should be further explained that, in the specific implementation process, the process of obtaining and storing the personal and facial information of personnel within the park includes:
[0053] The personal information refers to the name, gender, age, occupation, mobile phone number, ID number, and ID photo of relevant personnel within the park. The facial information refers to the facial image and its image features used to identify relevant personnel within the park, such as contour features, facial features, and detail features.
[0054] Set up a personnel database, bind the personal information of relevant personnel in the park with their corresponding facial information, and upload both to the personnel database for storage;
[0055] The outline features include the shape, proportion, and size of the facial contour; the facial features include the shape, color, and spacing of the eyes; the shape, height, and size of the nose; the shape, position, and size of the ears; the shape, position, and density of the eyebrows; the shape, thickness, and size of the mouth; and the detail features include the position and shape of moles, freckles, and dimples on the face, as well as the distribution and depth of wrinkles.
[0056] It should be further explained that, in the specific implementation process, the process of setting up a data acquisition unit to obtain facial image frames of a single person at different acquisition distances and angles includes:
[0057] The acquisition unit refers to a device used to acquire facial images and videos, which is generally a camera. The acquisition unit acquires facial image data of relevant people within its acquisition range in real time. The acquisition range refers to the maximum range of image data that the acquisition unit can acquire.
[0058] The acquisition distance refers to the distance between the face of the relevant person and the acquisition unit. The acquisition angle refers to the angle of the face of the relevant person relative to the acquisition unit, including the horizontal angle and the vertical angle. The horizontal angle refers to the deflection angle of the face relative to the acquisition unit in the horizontal direction, and the vertical angle refers to the pitch angle of the face relative to the acquisition device in the vertical direction.
[0059] Taking the facial image data of a single person acquired in a single instance as an example, the "single instance" refers to the process from the person entering the acquisition range to leaving the acquisition range. The facial image data consists of a series of continuous image frames, which are arranged in a certain time order. Each image frame is marked as the facial image frame of the single person. The acquisition distance and acquisition angle of each facial image frame at the corresponding time are obtained, which can be obtained through existing technologies such as laser ranging and electronic compass.
[0060] It should be further explained that, in the specific implementation process, the process of obtaining the facial similarity between the facial information of a single person and its different facial image frames includes:
[0061] The core principle of facial recognition technology is to convert a face image into a set of representative feature vectors, and then measure the similarity between faces by calculating the distance between these feature vectors (such as cosine similarity). The closer the distance, the higher the similarity.
[0062] Taking the facial information of a single person and a single facial image frame as an example, the image features in the single facial image frame are obtained, and the pose correction of the single facial image frame is performed according to the image features so that the face is in a standard frontal pose.
[0063] The AlexNet algorithm is used to extract the feature vectors of the face image in the single face image frame and the face information of the single person. The cosine similarity between the feature vectors of the two is used as the face similarity between the two. The same method is used to obtain the face similarity between the face image in the face information of the single person and each of its face image frames.
[0064] It should be further explained that, in the specific implementation process, the process of constructing the first evaluation model based on the facial image frames and their facial similarity from different personnel at different acquisition distances and angles includes:
[0065] Using the same method as described above for obtaining face image frames of a single person at different acquisition distances and angles, face image frames of different people at different acquisition distances and angles were obtained respectively.
[0066] Using the same method as described above for obtaining the facial similarity between a single person's facial image and its different facial image frames, the facial similarity between different people's facial images and their corresponding different facial image frames is obtained respectively;
[0067] The first evaluation set is generated based on the facial images of different individuals, their facial image frames at different acquisition distances and angles, and facial similarity, and is divided into the first training set and the first test set.
[0068] Construct a first convolutional neural network, using the face images of different people in the first training set and their different face image frames as the input data of the first convolutional neural network, and using the corresponding face similarity in the first training set as the output data of the first convolutional neural network.
[0069] The first convolutional neural network is trained to obtain an initial first convolutional neural network. The initial first convolutional neural network is validated using a first test set. The initial first convolutional neural network whose output is less than or equal to a preset first test error threshold is used as the first evaluation model.
[0070] It should be further explained that, in the specific implementation process, the process of obtaining the similarity set and similarity change features between facial image frames of a single person at different acquisition distances and angles and facial information of different people includes:
[0071] The farther the acquisition distance and the larger the acquisition angle, the lower the clarity of the corresponding face image frame. This will cause the initial face image frame of a single person and multiple face images stored in the person database to be judged as similar in the face recognition process. Each face image that is in a similar state to the single person will be marked as a similar object of the single person.
[0072] The "initial" refers to the earliest corresponding time in the facial image data of a single person acquired in a single instance, and the judgment criterion for the similarity state is that the facial similarity between the initial facial image frame of a single person and the corresponding facial image is greater than or equal to a preset similarity threshold.
[0073] Taking a single person and a single similar object as an example, the facial similarity between the single person and the single similar object is obtained from different facial image frames of the single person. The obtained facial similarity is sorted according to the time sequence of the corresponding facial image frames to obtain the similarity set between the single person and the single similar object.
[0074] Obtain the similarity set between the individual and each of its similar objects. Since a single similarity set is a set of values arranged in chronological order, it is possible to obtain the similarity change characteristics based on its changes, including mean, variance, trend (increasing, decreasing, or remaining flat), and bind the similarity set between the individual and its different similar objects and their corresponding similarity change characteristics.
[0075] It should be further explained that, in the specific implementation process, the process of constructing the second evaluation model based on the similarity variation characteristics of different facial image frames and different facial information of different individuals includes:
[0076] Among all the similar objects corresponding to a single person, there is one similar object that is in the same state as the single person. That is, among all the similar objects of the single person, there is one that is the single person himself.
[0077] The criterion for judging the same state is that the latest face similarity in the similarity set of the corresponding similar objects is greater than or equal to the preset same threshold, and the similar objects in the same state are marked as target objects. The latest refers to the latest time in the face image data of a single person in a single acquisition.
[0078] A second evaluation set is generated based on the different similarity sets of different individuals and their various similar objects, as well as the corresponding similarity change characteristics, and is divided into a second training set and a second test set.
[0079] Construct a second convolutional neural network, using the different similarity sets of different people and their similar objects in the second training set as the input data of the second convolutional neural network, and the target object in the similar objects of different people in the second training set as the output data of the second convolutional neural network.
[0080] The second convolutional neural network is trained to obtain an initial second convolutional neural network. The initial second convolutional neural network is validated using a second test set. The initial second convolutional neural network whose output is less than or equal to a preset second test error threshold is used as the second evaluation model.
[0081] It should be further explained that, in the specific implementation process, the process of acquiring the recognition image frame, obtaining the corresponding recognition similarity and its similarity change features by combining the first evaluation model, and determining whether it is feasible by combining the second evaluation model includes:
[0082] In subsequent application scenarios, when the acquisition unit identifies a single person within its acquisition range, it continuously acquires multiple facial image frames of that single person, which are recorded as recognition image frames.
[0083] The initial recognition image frame and each face image stored in the personnel database are respectively input into the first evaluation model to obtain the corresponding recognition similarity. The face images that are greater than or equal to the preset similarity threshold are taken as the similar objects of the single person.
[0084] Using the first evaluation model, the recognition similarity between each recognition image frame of the single person and its single similar object is obtained, and the corresponding similarity set and similarity change features are obtained. The similarity set and similarity change features between the single person and each of its similar objects are obtained.
[0085] The similarity set between the single person and each of their similar objects, along with their similarity change characteristics, is input into the second evaluation model to determine whether the single person's target object exists among the similar objects. If it exists, the single person is determined to be a relevant person stored in the personnel database and is marked as passable. If it does not exist, the single person is determined not to be a relevant person stored in the personnel database and is marked as inaccessible.
[0086] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A facial recognition and security system for smart parks, characterized in that, It includes the following modules: a data acquisition module, used to acquire and store the personal and facial information of people in the park; The face analysis module is used to set up acquisition units to acquire face image frames of a single person at different acquisition distances and angles, and to obtain the face similarity between the single person's face information and different face image frames. The first evaluation module is used to construct a first evaluation model based on the face image frames of different people at different acquisition distances and angles and their face similarities. The second evaluation module is used to acquire the similarity set and similarity change characteristics between the face image frames of a single person at different acquisition distances and angles and the face information of different people, and to construct a second evaluation model based on the similarity set and the aforementioned similarity change characteristics between different face image frames of different people and different face information, by constructing a second convolutional neural network, dividing the training set and test set, and training and validating. The face recognition module is used to acquire recognition image frames, combine the first evaluation model to obtain the corresponding recognition similarity and its similarity change characteristics, and combine the second evaluation model to determine whether passage is permitted. The similarity change features include: mean, variance, and trend, which bind the similarity set of a single person with different similar objects and their similarity change features; The second evaluation model is an initial second convolutional neural network with an error threshold less than or equal to a preset second test error threshold. The second convolutional neural network is constructed by taking the different similarity sets of different people and their similar objects in the training set and the similarity change features as the input data of the second convolutional neural network, and taking the target object in the similar objects of different people in the training set as the output data of the second convolutional neural network.
2. The facial recognition and security system for smart parks according to claim 1, characterized in that, The process of obtaining and storing personal and facial information includes: The personal information refers to the name, gender, age, occupation, mobile phone number, ID number, and ID photo of relevant personnel within the park; the facial information refers to the facial image and its image features used to identify relevant personnel within the park. A personnel database is set up to bind the personal information of relevant personnel in the park with their corresponding facial information, and both are uploaded to the personnel database for storage.
3. The facial recognition and security system for smart parks according to claim 2, characterized in that, The process of acquiring different facial image frames of a single person includes: The acquisition unit refers to a device used to acquire facial images and videos, and to acquire facial image data of relevant personnel within its acquisition range in real time. The acquisition distance refers to the distance between the face of the relevant person and the acquisition unit, and the acquisition angle refers to the angle of the face of the relevant person relative to the acquisition unit, including the horizontal angle and the vertical angle. Each image frame in the facial image data acquired from a single person in a single instance is labeled as a facial image frame of that single person, and the acquisition distance and acquisition angle of each facial image frame at the corresponding time are obtained.
4. A facial recognition and security system for smart parks according to claim 3, characterized in that, The process of obtaining facial similarity between different facial image frames includes: Based on the image features in a single face image frame of a single person, pose correction is performed on the single face image frame to make the face in a frontal pose. The AlexNet algorithm is used to extract the feature vectors of the face image in the single face image frame and the face information of the single person. The cosine similarity between the feature vectors of the two is used as the face similarity between the two, and the face similarity between the face image in the face information of the single person and each face image frame is obtained.
5. A facial recognition and security system for smart parks according to claim 4, characterized in that, The process of constructing the first evaluation model includes: Acquire facial image frames of different individuals at different acquisition distances and angles, and obtain the facial similarity between the facial images of different individuals and their corresponding different facial image frames; The first evaluation set is generated based on the facial images of different individuals, their facial image frames at different acquisition distances and angles, and facial similarity, and is divided into the first training set and the first test set. Construct a first convolutional neural network, using the face images of different people in the first training set and their different face image frames as the input data of the first convolutional neural network, and using the corresponding face similarity in the first training set as the output data of the first convolutional neural network. The first convolutional neural network is trained to obtain an initial first convolutional neural network. The initial first convolutional neural network is validated using a first test set. The initial first convolutional neural network whose output is less than or equal to a preset first test error threshold is used as the first evaluation model.
6. A facial recognition and security system for smart parks according to claim 5, characterized in that, The process of obtaining the similarity set and similarity change features of a single person includes: When the initial face image frame of a single person is judged to be similar to multiple face images stored in the person database during the face recognition process, each face image that is in a similar state to the single person will be marked as a similar object of the single person. The "initial" refers to the earliest corresponding time in the facial image data of a single person acquired in a single instance. The criterion for judging the similarity state is that the facial similarity between the initial facial image frame of a single person and the corresponding facial image is greater than or equal to a preset similarity threshold. The facial similarity between a single person and a single similar object is obtained by acquiring different facial image frames of a single person and sorting the acquired facial similarity according to the chronological order of the corresponding facial image frames to obtain the similarity set between the single person and the single similar object. Obtain the similarity set between the individual and each of its similar objects, and obtain the corresponding similarity change features, including mean, variance, and trend. Then bind the similarity set between the individual and its different similar objects and their similarity change features.
7. A facial recognition and security system for smart parks according to claim 6, characterized in that, The process of constructing the second evaluation model includes: Retrieve similar objects that are in the same state as a single person among all similar objects, and mark the similar objects in the same state as the target objects of that single person. The criterion for judging the same state is that the latest face similarity in the similarity set of the corresponding similar objects is greater than or equal to the preset same threshold. The latest refers to the latest time in the face image data of a single person in a single acquisition. A second evaluation set is generated based on the different similarity sets of different individuals and their various similar objects, as well as the corresponding similarity change characteristics, and is divided into a second training set and a second test set. Construct a second convolutional neural network, using the different similarity sets of different people and their similar objects in the second training set as the input data of the second convolutional neural network, and the target object in the similar objects of different people in the second training set as the output data of the second convolutional neural network. The second convolutional neural network is trained to obtain an initial second convolutional neural network. The initial second convolutional neural network is validated using a second test set. The initial second convolutional neural network whose output is less than or equal to a preset second test error threshold is used as the second evaluation model.
8. A facial recognition and security system for smart parks according to claim 7, characterized in that, The process of acquiring and recognizing image frames, obtaining their recognition similarity and similarity change features, and determining whether they are passable includes: In subsequent application scenarios, when the acquisition unit identifies a single person within its acquisition range, it continuously acquires multiple recognition image frames of that single person. The initial recognition image frame and each face image stored in the personnel database are respectively input into the first evaluation model to obtain the corresponding recognition similarity. The face images that are greater than or equal to the preset similarity threshold are taken as the similar objects of the single person. Using the first evaluation model, the recognition similarity between each recognition image frame of the single person and its single similar object is obtained, and the corresponding similarity set and similarity change features are obtained. The similarity set and similarity change features between the single person and each of its similar objects are obtained. The similarity set of the single person and each of their similar objects and the characteristics of their similarity changes are input into the second evaluation model to determine whether the single person's target object exists among the similar objects. If it exists, the single person is marked as passable; if it does not exist, it is marked as inaccessible.
Citation Information
Patent Citations
Intelligent identification system for access control
CN118155269A
Monitoring device and monitoring method
JP2018061114A