Access control method and system based on artificial intelligence and Internet of Things

Abstract

The invention provides an access control method and system based on artificial intelligence and Internet of Things, and is applied to an access control system, the access control system comprises a restricted access area door, and the restricted access area door is provided with a restricted acquisition camera; comprising the following steps: collecting a plurality of frames of walking images of a passing requesting person through the collection limiting camera; obtaining multiple frames of human body postures of the passing requesting person based on the multiple frames of walking images; the multiple frames of human body postures are sent to a human body walking dynamic analysis model, and the walking dynamic state of the passing requesting person is obtained; and if the walking dynamic state of the passing requesting person is a passing permission dynamic state, unlocking the access restriction area door, and performing biological feature recognition on the passing requesting person in the process that the passing requesting person moves to the access restriction area door. The problems that the door opening time is long and the efficiency is low due to the fact that a person needs to stand in front of a door to wait for recognition and unlocking in the existing biological recognition technology of fingerprint, iris, face recognition and vein recognition can be solved.

Description

Technical Field

[0001] The present invention relates to the technical field of the Internet of Things, and in particular, to an access control method and system based on artificial intelligence and the Internet of Things. Background Art

[0002] The access control system in the field of intelligent buildings is referred to as the Access Control System, abbreviated as ACS. It refers to the permission to prohibit access to the "door", which is the guard against the "door". Here, the "door" generally includes various channels that can be passed through, including the doors for people to pass through and the doors for vehicles to pass through.

[0003] The access control and security management system for entrances and exits is a new type of modern security management system that integrates microcomputer automatic identification technology and modern security management measures. The access control system can be divided into password recognition, card recognition, biometric recognition, etc. according to the access recognition method. Among them, biometric recognition controls access by verifying the biometric characteristics of personnel, including fingerprint type, palm type, iris type, face recognition type, and finger vein recognition type, etc. It has the characteristics of not needing to carry cards and high security. However, at present, the biometric recognition technologies such as fingerprint, iris, face recognition, and vein recognition have the problems that personnel need to stand in front of the door waiting for recognition and unlocking, resulting in a long door opening time and low efficiency. Summary of the Invention

[0004] In view of the above technical problems, the embodiments of the present application propose an access control method and system based on artificial intelligence and the Internet of Things, which can solve the problems that the current biometric recognition technologies such as fingerprint, iris, face recognition, and vein recognition require personnel to stand in front of the door waiting for recognition and unlocking, resulting in a long door opening time and low efficiency.

[0005] In a first aspect, the embodiments of the present application provide an access control method based on artificial intelligence and the Internet of Things, which is applied to an access control system. The access control system includes a door for restricting entry into an area, and a restricted acquisition camera is provided on the door for restricting entry into the area. The access control method based on artificial intelligence and the Internet of Things includes: Collecting multiple frames of walking images of a person requesting access through the restricted acquisition camera; Obtaining multiple frames of the human postures of the person requesting access based on the multiple frames of the walking images; Sending the multiple frames of the human postures to a human walking dynamic analysis model to obtain the walking dynamic of the person requesting access; If the walking dynamic of the person requesting access is an allowed access dynamic, unlocking the door for restricting entry into the area.

[0006] In some embodiments, obtaining multiple frames of the human postures of the person requesting access based on the multiple frames of the walking images includes: Perform human segmentation on the walking image to obtain a human walking image; Perform bone key point detection on the human walking image to obtain multiple bone key points of the requesting passerby; Stitch multiple of the bone key points into a human bone image to obtain the human posture of the requesting passerby.

[0007] In some embodiments, the unlocking of the restricted access area door when the walking dynamic of the requesting passerby is an allowed access dynamic includes: Calculate the dynamic similarity between the walking dynamic of the requesting passerby and the walking dynamic data in the dynamic access library; If the dynamic similarity is greater than the allowed access threshold and the walking dynamic data with a similarity greater than the allowed access threshold to the walking dynamic of the requesting passerby has an allowed access mark, then unlock the restricted access area door.

[0008] In some embodiments, the calculating the dynamic similarity between the walking dynamic of the requesting passerby and the walking dynamic data in the dynamic access library further includes: If the dynamic similarity is greater than the prohibited access threshold and less than or equal to the allowed access threshold, and the walking dynamic data with a similarity greater than the prohibited access threshold and less than or equal to the allowed access threshold to the walking dynamic of the requesting passerby has an allowed access mark, then extract personal feature data from the walking image; Calculate the personal feature similarity between the personal feature data and the feature data in the personal feature database; Obtain a personal dynamic access score based on the dynamic similarity and the personal feature similarity; If the personal dynamic access score is greater than the access score threshold, then unlock the restricted access area door.

[0009] In some embodiments, the obtaining a personal dynamic access score based on the dynamic similarity and the personal feature similarity further includes: If the personal dynamic access score is less than or equal to the access score threshold, then perform facial feature recognition on the requesting passerby; If the facial feature recognition passes, then unlock the restricted access area door.

[0010] In some embodiments, the access control method based on artificial intelligence and the Internet of Things further includes: Obtain the temporary access request of the requesting passerby; Set an allowed access mark and an allowed access time for the walking dynamic data in the dynamic access library corresponding to the requesting passerby based on the temporary access request.

[0011] In some embodiments, the access control method based on artificial intelligence and the Internet of Things further includes: Recognize multiple frames of walking images to obtain personal whereabouts data; Generate a personal event record based on the personal whereabouts data and the temporary pass request; The personal event record is sent to the management system.

[0012] In some embodiments, if the walking dynamics of the pedestrian requesting passage are dynamics of allowing passage, unlocking the restricted access area door further includes: Analyzing the walking dynamics of the pedestrian requesting passage to obtain the walking status of the pedestrian requesting passage; If the walking state is a state inconvenient to open the door, the restricted access area door is propped open by a supporting device.

[0013] In some embodiments, if the walking dynamics of the pedestrian requesting passage are dynamics of allowing passage, unlocking the restricted access area door further includes: The restricted acquisition camera is used to collect the real-time image of the person requesting to pass through; The real-time door entry image is analyzed, and if the person requesting access does not enter the restricted access area door, the time for unlocking the restricted access area door is extended.

[0014] In a second aspect, an embodiment of the present application provides an access control system based on artificial intelligence and the Internet of Things, which is applied to an access control system, wherein the access control system includes a restricted access area door, and a restricted acquisition camera is provided on the restricted access area door; The access control system based on artificial intelligence and the Internet of Things includes: A collection module, used for collecting multiple frames of walking images of the pedestrian requesting to pass through the restricted collection camera; An analysis module is used to obtain a plurality of frames of human postures of the pedestrian requesting to pass based on the plurality of frames of walking images; and send the plurality of frames of human postures to a human walking dynamics analysis model to obtain the walking dynamics of the pedestrian requesting to pass; The control module is used to unlock the restricted access area door if the walking dynamics of the pedestrian requesting passage are passage-permitting dynamics.

[0015] The present application provides an access control method and system based on artificial intelligence and the Internet of Things, which are applied to an access control system. The access control system includes a door for restricting access to an area, and a restricted acquisition camera is provided on the door for restricting access to the area. The method includes: collecting multiple frames of walking images of a person requesting access through the restricted acquisition camera; obtaining multiple frames of human postures of the person requesting access based on the multiple frames of walking images; sending the multiple frames of human postures to a human walking dynamic analysis model to obtain the walking dynamic of the person requesting access; if the walking dynamic of the person requesting access is an allowed access dynamic, unlocking the door for restricting access to the area, which can perform biometric identification on the person requesting access during the process of the person requesting access walking towards the door for restricting access to the area, and can solve the problem that the existing biometric identification technologies such as fingerprint, iris, face recognition, and vein recognition require personnel to stand in front of the door waiting for identification and unlocking, resulting in a long door opening time and low efficiency. Description of the Drawings

[0016] Hereinafter, the present invention will be described in more detail based on embodiments and with reference to the drawings.

[0017] Figure 1 is a flowchart of an access control method based on artificial intelligence and the Internet of Things provided by an embodiment of the present invention; Figure 2 is a schematic diagram of an access control system based on artificial intelligence and the Internet of Things provided by an embodiment of the present invention. Detailed Embodiments

[0018] The present invention will be further described below with reference to the drawings.

[0019] An access control system (ACS for short) in the field of intelligent buildings refers to the permission to prohibit access to a "door", which is a guard against the "door". Here, the "door" generally includes various passages that can be passed through, including doors for people to pass through, doors for vehicles to pass through, etc.

[0020] The access control and security management system for entrances and exits is a new type of modern security management system. It integrates microcomputer automatic identification technology and modern security management measures, and involves many new technologies such as electronics, machinery, optics, computer technology, communication technology, and biotechnology. It is an effective measure to solve the security prevention management of important department entrances and exits. It is applicable to various confidential departments, such as banks, hotels, parking lot management, computer rooms, confidential rooms, offices, intelligent communities, factories, etc. Nowadays, the access control system has long exceeded the simple management of doorways and keys, and has gradually developed into a complete access management system. It plays a greater role in administrative management work such as work environment security and personnel attendance management.

[0021] An access control system usually includes: an access control controller, an access control power supply, an identification device, an exit button, a computer, a network switch, an emergency button, an electric plug lock, etc. The access control system can be divided into password recognition, card recognition, biometric recognition, etc. according to the access recognition method. Among them, password recognition has problems such as easy password leakage, low security, no access record, and only one-way control; card recognition has problems such as card wear, short lifespan, easy card replication, and easy loss of card information due to external magnetic fields, making the card invalid; while biometric recognition technology has been widely used due to its characteristics of not requiring card carrying and high security. However, currently, biometric recognition technologies such as fingerprint, iris, face recognition, and vein recognition have problems of long opening time and low efficiency due to the need for personnel to stand in front of the door waiting for recognition and unlocking.

[0022] In a first aspect, as Figure 1 shown, in view of the above technical problems, an embodiment of the present application provides an access control method based on artificial intelligence and the Internet of Things, which is applied to an access control system. The access control system includes a door for restricting access to an area, and a restricted acquisition camera is provided on the door for restricting access to the area; The access control method based on artificial intelligence and the Internet of Things includes: S101: Collect multiple frames of walking images of the person requesting access through the restricted acquisition camera; S102: Obtain multiple frames of the body postures of the person requesting access based on the multiple frames of the walking images; S103: Send the multiple frames of the body postures to a human walking dynamic analysis model to obtain the walking dynamic of the person requesting access; S104: If the walking dynamic of the person requesting access is an allowed access dynamic, unlock the door for restricting access to the area.

[0023] It should be noted that the door for restricting access to the area can be an isolation door for the office area of company leaders, a floor access door of an office building, etc. An electric control lock is provided on the door for restricting access to the area. The electric control lock can be an electromagnetic lock, an anode lock, a cathode lock, etc. The specific type of the electric control lock can be selected according to the type of the door for restricting access to the area. The electric control lock is connected to an access control system (including a computer, a communication system, etc.).

[0024] It should be noted that a sensor can be set on the door for restricting access to the area or in front of the door for restricting access to the area. The sensor is used to detect the distance between the person requesting access and the door for restricting access to the area, and when the person requesting access enters a preset acquisition range, start the restricted acquisition camera to perform image acquisition.

[0025] It should be noted that due to the differences in physical conditions, habits, and genes of each person, the dynamic postures of each person during walking are not the same (such as stride length, arm swing amplitude, body forward inclination degree, body sway amplitude and frequency, head sway amplitude and frequency, etc.). Therefore, by using the human walking dynamic analysis model to analyze multiple frames of the human postures, the walking dynamic of the requesting passerby is obtained (obtaining the walking dynamic of the requesting passerby can determine who is the person going to and approaching the restricted access area door); if the walking dynamic of the requesting passerby is an allowed access dynamic, the restricted access area door is unlocked, and biometric identification of the requesting passerby can be performed during the process of the requesting passerby approaching the restricted access area door, thereby avoiding the requesting passerby waiting for identification and unlocking in front of the door and improving the access efficiency. Among them, the human walking dynamic analysis model is obtained through training with a large amount of data, and the human walking dynamic analysis model can be a deep learning model, a neural network model, etc. In some embodiments, obtaining multiple frames of the human postures of the requesting passerby based on the multiple frames of the walking images includes: Performing human segmentation on the walking images to obtain human walking images; Performing bone key point detection on the human walking images to obtain multiple bone key points of the requesting passerby; Stitching the multiple bone key points into a human bone image to obtain the human posture of the requesting passerby.

[0026] It should be noted that by performing human segmentation on the walking images to obtain human walking images, the human body in the walking images can be separated from the background, which is convenient for subsequent detection and analysis of the human walking images. The bone key points include part points such as ankles, knees, hip joints, wrists, elbows, shoulders, and cervical vertebrae. By stitching the multiple bone key points into a human bone image according to the human body sequence, the human posture of the requesting passerby can be obtained.

[0027] In some embodiments, if the walking dynamic of the requesting passerby is an allowed access dynamic, then unlocking the restricted access area door includes: Calculating the dynamic similarity between the walking dynamic of the requesting passerby and the walking dynamic data in the dynamic access library; If the dynamic similarity is greater than the allowed access threshold and the walking dynamic data with a similarity greater than the allowed access threshold to the walking dynamic of the requesting passerby has an allowed access mark, then unlock the restricted access area door.

[0028] It should be noted that if the dynamic similarity is greater than the allowable passage threshold and the walking dynamic data with a similarity greater than the allowable passage threshold to the walking dynamics of the requesting passerby has an allowable passage mark, it indicates that the requesting passerby is an internal person and has the permission to enter the restricted access area door. If the dynamic similarity is greater than the allowable passage threshold, but the walking dynamic data with a similarity greater than the allowable passage threshold to the walking dynamics of the requesting passerby does not have an allowable passage mark, it indicates that although the requesting passerby is an internal person, they do not have the permission to enter the restricted access area door, thus realizing access management based on permissions. Among them, the allowable passage threshold can be 90%, or can be set as needed. The dynamic passage library stores the walking dynamic data of all employees of the company, or all employees of a department, or regular entrants and exits of the office building.

[0029] In some embodiments, calculating the dynamic similarity between the walking dynamics of the requesting passerby and the walking dynamic data in the dynamic passage library further includes: If the dynamic similarity is greater than the prohibited passage threshold and less than or equal to the allowable passage threshold, and the walking dynamic data with a similarity greater than the prohibited passage threshold and less than or equal to the allowable passage threshold to the walking dynamics of the requesting passerby has an allowable passage mark, personal characteristics are extracted from the walking image to obtain personal characteristic data; Calculate the personal characteristic similarity between the personal characteristic data and the characteristic data in the personal characteristic database; Obtain a personal dynamic passage score based on the dynamic similarity and the personal characteristic similarity; If the personal dynamic passage score is greater than the passage score threshold, unlock the restricted access area door.

[0030] It should be noted that if the dynamic similarity is greater than the no - access threshold and less than or equal to the access - permitted threshold, and the walking - dynamic data with a similarity greater than the no - access threshold and less than or equal to the access - permitted threshold to the walking dynamics of the requesting access person has an access - permitted mark, it indicates that it is difficult to determine whether the requesting access person can enter the restricted - access area door solely based on the walking dynamics of the requesting access person. However, at this time, a preliminary candidate has been determined (i.e., the person corresponding to the walking - dynamic data in the dynamic access library). And each person has some personal distinctive features (i.e., the personal - feature data), which are relatively easy to collect and identify, such as a large mole on the face, the position of the hairline, the height of the nose bridge, etc. By obtaining a personal dynamic access score based on the dynamic similarity and the personal - feature similarity, and unlocking the restricted - access area door when the personal dynamic access score is greater than the access - score threshold, the requesting access person can be accurately identified. Among them, obtaining the personal dynamic access score based on the dynamic similarity and the personal - feature similarity can specifically be: the dynamic similarity × the personal - feature similarity, or the dynamic similarity + the personal - feature similarity, or other relevant calculation methods. This application does not make specific limitations on this.

[0031] In some embodiments, obtaining the personal dynamic access score based on the dynamic similarity and the personal - feature similarity further includes: If the personal dynamic access score is less than or equal to the access - score threshold, facial feature recognition is performed on the requesting access person; If the facial feature recognition passes, the restricted - access area door is unlocked.

[0032] It should be noted that when the personal dynamic access score is less than or equal to the access - score threshold, usually the requesting access person does not meet the conditions for entering the restricted - access area door (or does not have the corresponding permission). However, to avoid misjudgment, the requesting access person can be reminded to perform facial feature recognition. If the facial feature recognition fails, the restricted - access area door is locked to prohibit the requesting access person from entering. Among them, other biometric recognition methods can also be used for further verification, such as fingerprint recognition, iris recognition, etc. This application does not make specific limitations on this.

[0033] In some embodiments, the access control method based on artificial intelligence and the Internet of Things further includes: Obtain the temporary access request of the requesting access person; Based on the temporary access request, set the access - permitted mark and the access - permitted time for the walking - dynamic data in the dynamic access library corresponding to the requesting access person.

[0034] It should be noted that during the company's office activities, employees often need to report matters to their leaders or request their leaders to approve matters. At this time, they usually need to enter the restricted access area door to report and request approval, but employees do not have the authority to enter the restricted access area door under normal circumstances. The embodiment of the present application sets the mark of the employee's walking dynamic data in the dynamic access library (i.e., the allowed access mark) and the allowed access time according to the temporary access request filed by the employee (the temporary access request includes the requested matter and the requested access time), so as to meet the temporary access needs of the employee, and the allowed access mark and the allowed access time can be cancelled after the employee makes the corresponding report and requests for approval, thereby ensuring the safety of access. Among them, employees who have the right to enter the restricted access area door are usually set with a long-term allowed access mark, and the temporary access request of the person requesting access can be obtained through the enterprise's internal management system.

[0035] It should be noted that when judging whether the person requesting access has the authority to enter the restricted access area, a search may be performed in the walking dynamic data with the allowed access mark. If the person requesting access is not found in the walking dynamic data with the allowed access mark, it means that the person requesting access does not have the authority to enter the restricted access area door, thereby improving the efficiency of searching and comparing.

[0036] In some embodiments, the access control method based on artificial intelligence and the Internet of Things further includes: Recognize multiple frames of walking images to obtain personal whereabouts data; Generate a personal event record based on the personal whereabouts data and the temporary pass request; The personal event record is sent to the management system.

[0037] It should be noted that employee attendance is usually required within an enterprise. In order to avoid attendance errors caused by employees leaving their workstations for a long time, personal matter records are generated based on the personal whereabouts data and the temporary pass request, and the personal matter records are sent to the management system. This can accurately understand and record the whereabouts and matters of the employees, and effectively avoid attendance errors. The management system can be an internal enterprise management system, a personnel management system, etc.

[0038] In some embodiments, if the walking dynamics of the pedestrian requesting passage are dynamics of allowing passage, unlocking the restricted access area door further includes: Analyzing the walking dynamics of the pedestrian requesting passage to obtain the walking status of the pedestrian requesting passage; If the walking state is a state inconvenient to open the door, the restricted access area door is propped open by a supporting device.

[0039] It should be noted that, since employees (or office workers, people requesting access) sometimes need to carry some items to enter the restricted access area door, the people requesting access are usually in a state where it is inconvenient to open the door (such as holding a box with both hands). At this time, if the restricted access area door is simply unlocked, it is not convenient for the people requesting access to enter. Therefore, the restricted access area door can be propped open by the supporting device to facilitate entry. The supporting device is arranged on the restricted access area door and is used to prop open the restricted access area door. The supporting device can be a cylinder pushing a connecting rod structure, or a motor driving structure, etc., and the present application does not make specific limitations on this.

[0040] In some embodiments, if the walking dynamics of the pedestrian requesting passage are dynamics of allowing passage, unlocking the restricted access area door further includes: The restricted acquisition camera is used to collect the real-time image of the person requesting to pass through; The real-time door entry image is analyzed, and if the person requesting access does not enter the restricted access area door, the time for unlocking the restricted access area door is extended.

[0041] It should be noted that sometimes the person requesting access does not enter the restricted-access area door in time when the restricted-access area door is unlocked, and the unlocking time of the restricted-access area door expires. At this time, the restricted-access area door is re-locked, resulting in the need to re-identify the person requesting access in order to unlock the restricted-access area door again. By analyzing the real-time entry image, if the person requesting access does not enter the restricted-access area door, the time for unlocking the restricted-access area door is extended, which can avoid the trouble of re-detection and verification caused by the re-locking of the restricted-access area door.

[0042] In summary, the embodiment of the present application provides an access control method based on artificial intelligence and the Internet of Things, which is applied to an access control system, wherein the access control system includes a restricted access area door, and the restricted access area door is provided with a restricted acquisition camera; comprising: acquiring multiple frames of walking images of a person requesting passage through the restricted acquisition camera; obtaining multiple frames of human postures of the person requesting passage based on the multiple frames of the walking images; sending the multiple frames of human postures to a human walking dynamics analysis model to obtain the walking dynamics of the person requesting passage; if the walking dynamics of the person requesting passage is a dynamics of allowing passage, unlocking the restricted access area door, and being able to perform biometric identification of the person requesting passage in the process of the person requesting passage walking towards the restricted access area door, which can solve the problem that the current biometric identification technology through fingerprint, iris, facial recognition, and vein recognition requires people to stand in front of the door waiting for identification and waiting for unlocking, thereby resulting in a long door opening time and low efficiency.

[0043] Second aspect, as Figure 2 shown, an access control system based on artificial intelligence and the Internet of Things provided by an embodiment of the present application is applied to an access control system. The access control system includes a restricted access area door, and a restricted acquisition camera is provided on the restricted access area door; The access control system based on artificial intelligence and the Internet of Things includes: An acquisition module 210, configured to acquire multiple frames of walking images of a person requesting access through the restricted acquisition camera; An analysis module 220, configured to obtain multiple frames of body postures of the person requesting access based on the multiple frames of walking images; send the multiple frames of body postures to a human walking dynamic analysis model to obtain the walking dynamic of the person requesting access; A control module 230, configured to unlock the restricted access area door if the walking dynamic of the person requesting access is an allowed access dynamic.

[0044] In some embodiments, obtaining multiple frames of body postures of the person requesting access based on the multiple frames of walking images includes: Performing human segmentation on the walking images to obtain human walking images; Performing bone key point detection on the human walking images to obtain multiple bone key points of the person requesting access; Stitching the multiple bone key points into a human bone image to obtain the body posture of the person requesting access.

[0045] In some embodiments, if the walking dynamic of the person requesting access is an allowed access dynamic, then unlocking the restricted access area door includes: Calculating the dynamic similarity between the walking dynamic of the person requesting access and the walking dynamic data in the dynamic access library; If the dynamic similarity is greater than the allowed access threshold and the walking dynamic data whose similarity with the walking dynamic of the person requesting access is greater than the allowed access threshold has an allowed access mark, then unlock the restricted access area door.

[0046] In some embodiments, calculating the dynamic similarity between the walking dynamic of the person requesting access and the walking dynamic data in the dynamic access library further includes: If the dynamic similarity is greater than the prohibited access threshold and less than or equal to the allowed access threshold, and the walking dynamic data whose similarity with the walking dynamic of the person requesting access is greater than the prohibited access threshold and less than or equal to the allowed access threshold has an allowed access mark, then extract personal characteristics from the walking images to obtain personal characteristic data; Calculating the personal characteristic similarity between the personal characteristic data and the characteristic data in the personal characteristic database; Obtaining a personal dynamic traffic score based on the dynamic similarity and the personal characteristic similarity; If the personal dynamic access score is greater than the access score threshold, the restricted access area door is unlocked.

[0047] In some embodiments, obtaining a personal dynamic traffic score based on the dynamic similarity and the personal characteristic similarity further includes: If the individual dynamic access score is less than or equal to the access score threshold, facial feature recognition is performed on the person requesting access; If the facial feature recognition is passed, the restricted access area door is unlocked.

[0048] In some embodiments, the control module 230 is further configured to: Obtaining a temporary pass request from the pass requester; Based on the temporary pass request, a pass permission flag and a pass permission time are set for the walking dynamic data in the dynamic pass database corresponding to the person requesting pass.

[0049] In some embodiments, the control module 230 is further configured to: Recognize multiple frames of walking images to obtain personal whereabouts data; Generate a personal event record based on the personal whereabouts data and the temporary pass request; The personal event record is sent to the management system.

[0050] In some embodiments, if the walking dynamics of the pedestrian requesting passage are dynamics of allowing passage, unlocking the restricted access area door further includes: Analyzing the walking dynamics of the pedestrian requesting passage to obtain the walking status of the pedestrian requesting passage; If the walking state is a state inconvenient to open the door, the restricted access area door is propped open by a supporting device.

[0051] In some embodiments, if the walking dynamics of the pedestrian requesting passage are dynamics of allowing passage, unlocking the restricted access area door further includes: The restricted acquisition camera is used to collect the real-time image of the person requesting to pass through; The real-time door entry image is analyzed, and if the person requesting access does not enter the restricted access area door, the time for unlocking the restricted access area door is extended.

[0052] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (devices, systems), and / or computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 or means for implementing the functions specified in one or more of the blocks.

[0053] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 or means for implementing the functions specified in one or more of the blocks.

[0054] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 or means for implementing the functions specified in one or more of the blocks.

[0055] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present application. It should be understood that the above are only specific embodiments of the present application and are not used to limit the protection scope of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An access control method based on artificial intelligence and the Internet of Things, characterized in that, Applied to an access control system, the access control system includes a door for restricting access to an area, and a restricted acquisition camera is provided on the door for restricting access to the area; The access control method based on artificial intelligence and the Internet of Things includes: Collecting multiple frames of walking images of the person requesting access through the restricted acquisition camera; Obtaining multiple frames of the human postures of the person requesting access based on the multiple frames of the walking images; Sending the multiple frames of the human postures to a human walking dynamic analysis model to obtain the walking dynamic of the person requesting access; If the walking dynamic of the person requesting access is an allowed access dynamic, unlock the door for restricting access to the area.

2. The access control method based on artificial intelligence and the Internet of Things according to claim 1, characterized in that, The obtaining multiple frames of the human postures of the person requesting access based on the multiple frames of the walking images includes: Performing human segmentation on the walking images to obtain human walking images; Performing bone key point detection on the human walking images to obtain multiple bone key points of the person requesting access; Stitching the multiple bone key points into a human bone image to obtain the human posture of the person requesting access.

3. The access control method based on artificial intelligence and the Internet of Things according to claim 1, wherein The if the walking dynamic of the person requesting access is an allowed access dynamic, then unlocking the door for restricting access to the area includes: Calculating the dynamic similarity between the walking dynamic of the person requesting access and the walking dynamic data in the dynamic access library; If the dynamic similarity is greater than the allowed access threshold and the walking dynamic data with a similarity greater than the allowed access threshold to the walking dynamic of the person requesting access has an allowed access mark, then unlock the door for restricting access to the area.

4. The access control method based on artificial intelligence and the Internet of Things according to claim 3, characterized in that, The calculating the dynamic similarity between the walking dynamic of the person requesting access and the walking dynamic data in the dynamic access library further includes: If the dynamic similarity is greater than the prohibited access threshold and less than or equal to the allowed access threshold, and the walking dynamic data with a similarity greater than the prohibited access threshold and less than or equal to the allowed access threshold to the walking dynamic of the person requesting access has an allowed access mark, then extract personal characteristic data from the walking images; Calculating the personal characteristic similarity between the personal characteristic data and the characteristic data in the personal characteristic database; Obtaining a personal dynamic access score based on the dynamic similarity and the personal characteristic similarity; If the personal dynamic access score is greater than the access score threshold, then unlock the door for restricting access to the area.

5. The access control method based on artificial intelligence and the Internet of Things according to claim 4, characterized in that, The obtaining a personal dynamic access score based on the dynamic similarity and the personal characteristic similarity further includes: If the personal dynamic access score is less than or equal to the access score threshold, then perform facial feature recognition on the person requesting access; If the facial feature recognition passes, then unlock the door for restricting access to the area.

6. The access control method based on artificial intelligence and the Internet of Things according to claim 3, characterized in that It further includes: Obtaining the temporary access request of the person requesting access; Setting an allowed access mark and an allowed access time for the walking dynamic data in the dynamic access library corresponding to the person requesting access based on the temporary access request.

7. The access control method based on artificial intelligence and the Internet of Things according to claim 6, characterized in that, It further includes: Identifying the multiple frames of the walking images to obtain personal whereabouts data; Generating a personal matter record based on the personal whereabouts data and the temporary access request; Sending the personal matter record to the management system.

8. The access control method based on artificial intelligence and the Internet of Things according to claim 1, wherein If the walking dynamics of the pedestrian requesting passage are that passage is allowed, unlocking the restricted access area door also includes: Analyzing the walking dynamics of the pedestrian requesting passage to obtain the walking status of the pedestrian requesting passage; If the walking state is a state inconvenient to open the door, the restricted access area door is propped open by a supporting device.

9. The access control method based on artificial intelligence and the Internet of Things according to any one of claims 1-8, characterized in that, If the walking dynamics of the pedestrian requesting passage are that passage is allowed, unlocking the restricted access area door also includes: The restricted acquisition camera is used to collect the real-time image of the person requesting to pass through; The real-time door entry image is analyzed, and if the person requesting access does not enter the restricted access area door, the time for unlocking the restricted access area door is extended.

10. An access control system based on artificial intelligence and the Internet of Things, characterized in that, Applied to an access control system, the access control system includes a restricted access area door, and the restricted access area door is provided with a restricted acquisition camera; The access control system based on artificial intelligence and the Internet of Things includes: A collection module, used for collecting multiple frames of walking images of the pedestrian requesting to pass through the restricted collection camera; An analysis module is used to obtain a plurality of frames of human postures of the pedestrian requesting to pass based on the plurality of frames of walking images; and send the plurality of frames of human postures to a human walking dynamics analysis model to obtain the walking dynamics of the pedestrian requesting to pass; The control module is used to unlock the restricted access area door if the walking dynamics of the pedestrian requesting passage are passage-permitting dynamics.

Images