Method and system for opening a door lock

Abstract

The application discloses a door unlocking method and system of a door lock, and the method comprises the following steps: according to the event that the door lock detects that the self power is lower than a preset threshold, an auxiliary identification processing linked list containing the request source door lock to the end door lock on both sides is constructed through step-by-step transmission; according to the event that any door lock in the auxiliary identification processing linked list identifies the face feature of the owner of the request source door lock, the door lock is used as a group leader to create a tracking identification group; according to the snapshot time data of the face of the owner detected by the door locks in the tracking identification group in sequence, a time point of arrival at the request source door lock is predicted; according to the predicted time point of arrival, the door unlocking instruction is sent to the request source door lock by the group leader door lock, the tracking identification group is dissolved by the group leader door lock after the door is unlocked, and the normal identification mode of each door lock is restored. By using the embodiment of the application, the user can be continuously authenticated and actively served without feeling, and the robustness and user experience of the intelligent door lock system are improved.

Description

Technical Field

[0001] This invention belongs to the field of smart door lock technology, and in particular to a method and system for unlocking a door lock. Background Technology

[0002] Currently, the smart lock systems widely used in hotels and apartments rely on independent biometric identification for authentication and door opening. However, when the lock enters a low-power protection mode due to battery power falling below a threshold, disabling core identification functions, users will be unable to open the door normally, leading to interruption and a degraded user experience. Existing solutions are mostly reactive, such as using physical keys or guiding users to the front desk for assistance, failing to proactively restore service without the user's awareness. Furthermore, the recognition range of a single lock is limited, making it impossible to track and predict the user's movement within the building, and there is a lack of collaborative coordination between systems. Therefore, existing technologies suffer from service interruptions under special operating conditions such as low battery levels, reactive and inefficient responses, and a lack of global collaborative scheduling capabilities. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for unlocking door locks to overcome the shortcomings of the prior art, enabling seamless and continuous authentication and proactive service for users, thereby improving the robustness and user experience of smart door lock systems.

[0004] One embodiment of this application provides a method for unlocking a door lock, the method comprising:

[0005] Low battery detection and auxiliary network construction: Based on the event that the door lock detects that its own battery is lower than a preset threshold, unnecessary functional components are turned off, and auxiliary identification request messages are unicasted to adjacent door locks. An auxiliary identification processing chain list containing the request source door lock to the two end door locks is constructed through hierarchical transmission.

[0006] Target detection and tracking group establishment: Based on the event that any lock in the auxiliary recognition processing chain recognizes the homeowner's face feature of the requesting source lock, a tracking and recognition group is created with that lock as the group leader. The tracking and recognition group includes all locks on the path from the lock that recognizes the homeowner's face to the requesting source lock.

[0007] Movement tracking and arrival prediction: Based on the capture time data of the homeowner's face detected sequentially by the door locks of the tracking and identification team members, the team leader door lock calculates the average time interval between the homeowner passing through each door lock and predicts the time point when the homeowner arrives at the request source door lock;

[0008] Remote door opening and status recovery: Based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock, requesting the source door lock to perform a timed door opening operation. After the door opening is completed, the group leader door lock disbands the tracking and identification group and restores the normal identification mode of each door lock.

[0009] Optionally, the low-power detection and auxiliary network construction includes:

[0010] Low battery detection and power saving mode activation: Based on the trigger event that the door lock power sensor detects that the power is lower than the preset power threshold, an alarm notification is sent to the operation and maintenance system, and the camera, intelligent algorithm module and structured light camera are turned off at the same time;

[0011] Auxiliary Request Message Encapsulation: Based on the locally stored left and right neighbor door lock information, encapsulate an auxiliary recognition request message containing the request source door lock ID, request source door lock IP, transmission direction marker, and a list of homeowner face feature vectors;

[0012] The auxiliary network is constructed step by step: Based on the processing results of the auxiliary recognition request messages received by the adjacent door locks, the following operations are performed: save the homeowner's face feature vector of the requesting source door lock to the local feature vector table; update the auxiliary recognition processing linked list and add the node information of this door lock; forward the updated auxiliary recognition request message to the next adjacent door lock;

[0013] Auxiliary network integrity verification: Based on the event that the message has been transmitted to the end lock and cannot be transmitted further, verify that the auxiliary identification processing chain list has been constructed.

[0014] Optionally, the establishment of the target detection and tracking group includes:

[0015] Homeowner face recognition trigger: The process of creating a tracking and recognition group is triggered when a face image is captured by any door lock camera and matched with the facial feature vector of the homeowner of the requesting door lock.

[0016] Tracking and identification group initialization: Determine the complete path from the door lock that identifies the homeowner's face to the requesting source door lock based on the auxiliary identification processing chain, and create a tracking and identification group information table containing all door lock IDs on the path based on the door lock that identifies the homeowner's face;

[0017] Member status confirmation: Based on the response results of the group leader lock sending the group creation request message to all member locks, perform the following processing: For locks that respond normally, mark them as active members; for locks that do not respond, implement the retry mechanism until the maximum number of retries is reached; for locks that never respond, remove them from the tracking and identification group.

[0018] Tracking mode activation: When the number of active group members reaches a preset threshold, the tracking and recognition mode of all group members' door locks is activated, and the reporting object of the homeowner's facial features is set as the group leader's door lock.

[0019] Optionally, the travel tracking and arrival prediction include:

[0020] Sequential movement detection: Based on the events of the group members' door locks capturing the homeowner's face in the order of the path, the group leader's door lock receives and records the capture timestamps of each door lock;

[0021] Time interval calculation: Based on the difference in timestamps of two consecutive group members' door locks, calculate the average time interval between the homeowner passing through a single door lock;

[0022] Anomaly handling: Based on the detection of non-sequential movement events, perform the following anomaly handling: request the abnormal door lock to provide a captured image for re-identification and verification; investigate the reason for the failure to identify the missing door lock; adjust the movement path status according to the verification results;

[0023] Arrival Time Prediction: Based on the timestamp of the last member's door lock and the average time interval, calculate the estimated time when the homeowner will arrive at the requesting source door lock.

[0024] Optionally, the remote door opening and status restoration includes:

[0025] Remote door opening command transmission: Based on the calculated estimated arrival time, the group leader door lock sends an opening command containing the estimated time to the requesting source door lock;

[0026] Scheduled door opening execution: Based on the processing result of the door opening command received from the requesting door lock, a scheduled door opening task is set and the door opening operation is executed at the expected time.

[0027] State restoration and resource release: Based on the door opening completion event, perform the following cleanup operations: The group leader door lock sends a tracking and identification group disbandment message to all group members; the group member door locks restore the local reporting mode of the homeowner's facial features; and release the memory resources related to the tracking and identification group.

[0028] Voice prompts: Based on the successful door opening event, request the source door lock to announce the successful door opening and low battery reminder.

[0029] Optionally, the method further includes an anomaly detection and adaptive adjustment mechanism:

[0030] Adaptive threshold adjustment: If the group leader's door lock detects that a member's door lock has failed to be recognized due to an excessively high similarity threshold, a threshold adjustment suggestion is sent to that member's door lock.

[0031] Dynamic threshold configuration: Based on the processing results of the group member's door lock reception threshold adjustment suggestion, only the similarity threshold of the auxiliary recognition mode is adjusted, while the local recognition threshold remains unchanged;

[0032] Tracking timeout handling: If the tracking and identification group does not receive a new identification message within a preset time, the tracking and identification group will be automatically disbanded and resources will be released.

[0033] Turnback behavior recognition: Based on the detected event of the homeowner repeatedly appearing at the door lock on the path, the direction of travel is recalculated and the tracking recognition group status is updated.

[0034] Another embodiment of this application provides a door lock unlocking system, the system comprising:

[0035] The detection module is used for low battery detection and auxiliary network construction: based on the event that the door lock detects that its own battery is lower than a preset threshold, it shuts down unnecessary functional components and sends an auxiliary identification request message to the adjacent door locks by unicast. It constructs an auxiliary identification processing chain list containing the request source door lock to the two end door locks by passing the message level by level.

[0036] A module is established for target detection and tracking group establishment: based on the event that any lock in the auxiliary recognition processing chain recognizes the homeowner's face feature of the request source lock, the lock is used as the group leader to create a tracking and recognition group, which includes all locks on the path from the lock that recognizes the homeowner's face to the request source lock;

[0037] The tracking module is used for movement tracking and arrival prediction: based on the capture time data of the homeowner's face detected sequentially by the door locks of the tracking and identification group members, the group leader door lock calculates the average time interval between the homeowner passing through each door lock and predicts the time point of arrival at the request source door lock;

[0038] The door opening module is used for remote door opening and status recovery: based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock, requesting the source door lock to perform a timed door opening operation. After the door opening is completed, the group leader door lock disbands the tracking and identification group and restores the normal identification mode of each door lock.

[0039] Another embodiment of this application provides a storage medium storing a computer program, wherein the computer program is configured to execute the method described in any of the preceding claims when running.

[0040] Another embodiment of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the method described in any of the preceding claims.

[0041] Compared with existing technologies, the present invention provides a door lock opening method that, based on an event where the door lock detects that its own battery level is below a preset threshold, constructs an auxiliary identification processing chain list containing the requesting source door lock to the two end door locks through a hierarchical transmission; based on an event where any door lock in the auxiliary identification processing chain identifies the homeowner's facial features of the requesting source door lock, that door lock acts as the group leader to create a tracking identification group; based on the capture time data of the homeowner's face detected sequentially by the member door locks in the tracking identification group, the arrival time of the requesting source door lock is predicted; based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock; after the door is opened, the group leader door lock disbands the tracking identification group and restores the normal identification mode of each door lock, thereby achieving seamless, continuous identity verification and proactive service for the user, improving the robustness and user experience of the smart door lock system. Attached Figure Description

[0042] Figure 1 Hardware structure block diagram of a computer terminal for a door lock unlocking method provided in an embodiment of the present invention;

[0043] Figure 2 A flowchart illustrating a method for unlocking a door lock according to an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the location distribution of door locks in a hotel apartment, provided as an embodiment of the present invention.

[0045] Figure 4 This is a schematic diagram of a door lock opening system provided in an embodiment of the present invention. Detailed Implementation

[0046] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0047] Serviced apartments are a popular choice for accommodation, but their doors often use electronic locks with facial recognition or similar methods. This frequently leads to low battery issues, causing residents to be unable to open the door upon returning home, resulting in significant inconvenience and posing a challenge to timely maintenance. Solving this problem to ensure proper recognition and door opening is a key issue encountered during the operation of serviced apartments.

[0048] This invention first provides a method for unlocking a door lock, which can be applied to electronic devices, such as computer terminals, specifically ordinary computers.

[0049] The following detailed explanation uses a computer terminal as an example. Figure 1 This is a hardware structure block diagram of a computer terminal for a door lock unlocking method provided in an embodiment of the present invention. (See diagram below.) Figure 1As shown, the computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.

[0050] Non-volatile storage media can store operating systems and computer programs. These computer programs include program instructions that, when executed, cause the processor to perform any method of unlocking the door.

[0051] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0052] Internal memory provides an environment for the execution of computer programs stored in non-volatile storage media. When executed by a processor, the computer program enables the processor to perform any method of unlocking a door.

[0053] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0054] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0055] See Figures 2-3 The present invention provides a method for unlocking a door lock, which may include the following steps:

[0056] S201, Low power detection and auxiliary network construction: Based on the event that the door lock detects that its own power is lower than a preset threshold, it shuts down unnecessary functional components and sends an auxiliary identification request message to the adjacent door locks by unicast. The auxiliary identification processing chain list containing the request source door lock to the two end door locks is constructed by passing the message level by level.

[0057] This solution primarily addresses the issue of low battery locks on facial recognition door locks in serviced apartment settings. It utilizes the tracking and recognition of nearby door locks to assist in unlocking the low-battery lock. The core implementation idea is as follows:

[0058] When a door lock (designated as door lock A) detects that its battery is low, it sends a unicast message to the door locks on its left and right sides requesting assistance in identification. The message contains the homeowner's facial feature vector. Upon receiving the message, the left and right door locks save the homeowner's facial feature vector from door lock A and activate the assistance identification mode. The message continues to be transmitted from the left door to the left and from the right door to the right, until the far left or far right, and all receiving door locks switch to the assistance identification mode accordingly.

[0059] If a door lock (let's assume it's door lock E, and there are three doors, D, C, and B, between door lock A and door lock E) detects the face of homeowner A at door lock A, then door lock E, as the group leader, triggers the creation of a tracking and recognition group. This allows all door locks between E and A to track and recognize the face in sequence, making it easier to more accurately identify A's face and accurately estimate the time when the homeowner arrives at their own door lock.

[0060] If A passes through the locks from E to A in sequence, the time A will reach lock A can be predicted based on the capture time of the last lock and the average time taken to pass each lock, and lock A can be notified to open the door. If A does not pass through the locks in sequence or does not reach the vicinity of lock A, the door will not be opened.

[0061] When anomalies occur during A's movement, such as incorrect identification sequence, repeated appearance of personnel at a certain door lock, or failure to detect personnel, we investigate and confirm these anomalies to ensure identification accuracy. This includes an adaptive threshold adjustment mechanism: if a door lock (door lock C) fails to accurately identify the target personnel due to its own high threshold setting, the tracking and identification team leader will re-identify the personnel and propose adjusting the identification threshold for door lock C based on their own identification performance. Ultimately, door lock C will adopt this suggestion.

[0062] Specifically, the low-power detection and auxiliary network construction includes:

[0063] Low battery detection and power-saving mode activation: Based on the trigger event that the door lock's battery level is lower than a preset threshold (e.g., 5%) detected by the door lock's battery sensor, an alarm notification is sent to the operation and maintenance system, and the camera, intelligent algorithm module, and structured light camera are simultaneously turned off; Auxiliary request message encapsulation: Based on the locally stored information of the left and right neighboring door locks, an auxiliary recognition request message is encapsulated, including the request source door lock ID, request source door lock IP, transmission direction marker, and a list of homeowner face feature vectors; Auxiliary network hierarchical construction: Based on the processing results of the auxiliary recognition request messages received by neighboring door locks, the following operations are performed: Save the homeowner's face feature vector of the request source door lock to the local feature vector table; Update the auxiliary recognition processing chain list, adding the node information of this door lock; Forward the updated auxiliary recognition request message to the next adjacent door lock; Auxiliary network integrity confirmation: Based on the event that the message has been transmitted to the last door lock and cannot be transmitted further, the construction of the auxiliary recognition processing chain list is confirmed to be complete.

[0064] The preset conditions include:

[0065] ① The door locks of the hotel apartments are all within the same network segment (such as 192.168.0.0 / 24);

[0066] ② All door locks have been entered with complete local information, such as IP address, room number, and left and right neighbors; the method for saving left and right neighbor information is shown in Table 1 below:

[0067] Table 1: Static Information Table of Door Locks' Left and Right Neighbors

[0068] ;

[0069] ③ All door locks can parse and recognize the messages and instructions involved in this solution;

[0070] ④ The facial information of all homeowners in the house has been entered into the corresponding door locks;

[0071] ⑤ The door locks use chips with the same technical standard, which can achieve feature vector compatibility recognition between different door locks. That is to say, each door lock uses the same facial feature vector extraction network model.

[0072] Low battery detection and auxiliary request propagation: After detecting a low battery signal, door lock A notifies the repair department and shuts down its non-essential power-consuming devices, notifying the surrounding door locks to perform auxiliary identification. After receiving the message, the surrounding door locks start the auxiliary identification mode.

[0073] When door lock A detects that its battery level is below a threshold (usually set to 5%), it pushes a low battery message to the apartment maintenance personnel's mobile app. Upon receiving the message, the app displays a notification stating "Door lock A's battery level is below the threshold; please charge it promptly." After sending the message to the app, the system enters power-saving mode (this is to ensure that the equipment can operate for as long as possible and prevent situations where the homeowner is unable to open the door due to insufficient power upon arrival), turning off non-essential door-opening functions such as the camera, smart algorithm module, and structured light camera. Synchronously, unicast auxiliary identification request messages to the left and right neighboring door locks stored locally (the working principle of the left and right door locks is the same; the following explanation takes the left door lock direction as an example, and the left door lock neighbor relationship list is formed by passing messages one by one through the left door locks). The content includes: request assistance in identification command word, transmission direction mark (left direction), currently processed door lock ID (here IDA), currently processed door lock IP (IPA), request source ID (i.e., the door lock ID that sent the initial request when the battery was low, here IDA), request source IP (here IPA), and a list of facial feature information (containing the facial feature vectors of all owners in this room and the corresponding person IDs of the features stored by this door lock).

[0074] After receiving an auxiliary identification request message from lock A, the left lock of lock A (i.e., lock B) parses all fields in the message. Based on the source direction of the auxiliary identification request message (in this example, lock B finds that the message comes from its right neighbor, lock A), and the transmission direction (in this example, lock B finds that the transmission direction is from right to left), it creates a left-direction auxiliary identification processing linked list 2 and saves the request source information to table 2. At the same time, it adds its own lock information to table 2 as "Left Door 1" (the meaning in table 2 is: the request source is lock A, its ID is ID-A, and its IP is IP-A; the first node in the generated neighbor list, "Left Door 1," is this lock ID-B, and its IP address is IP-B). Then, it stores the parsed feature vectors into a local feature vector table (table 3), forming a vector information table containing the face of the homeowner and the face requiring auxiliary identification. After saving, door lock B adds the feature vector to be identified from door lock A to its own recognition feature vector, restarts the face recognition module, and enters the room + auxiliary recognition mode:

[0075] Table 2 Left-directed auxiliary recognition processing chain (Door lock B)

[0076] ;

[0077] Table 3. Door Lock Feature Vector Table (Door Lock B)

[0078] ;

[0079] The auxiliary recognition request is further transmitted to the left door lock of door lock B (i.e., door lock C). The message content includes: the request for auxiliary recognition command, the currently processed door lock ID (ID-B), the currently processed door lock IP (IP-B), the complete information in Table 2, and the face feature information table (containing all personnel feature vectors and the corresponding personnel IDs recorded in door lock A). After receiving the message, door lock C performs similar processing, saves it to the local left-side auxiliary recognition processing list, and adds the information of door lock C as left door 2, as shown in Table 4 below:

[0080] Table 4 Left-directed auxiliary recognition processing chain (door lock C)

[0081] ;

[0082] After door lock C completes the face recognition mode switch, it sends all the contents of the table to the next door lock in the message it sends, until the door lock receiving the message does not have a left-side door lock configured (ultimately forming a left-direction auxiliary recognition processing linked list such as ABCDEFG...), as shown in Table 5:

[0083] Table 5 Left-directed auxiliary recognition processing chain (door lock F)

[0084] ;

[0085] Similarly, messages for the right-side lock are passed along this path until the last lock without a right-side lock is reached (forming a right-side neighbor relationship table). Based on this, both locks on either side of lock A have obtained the facial feature information of lock A, as well as the number and order of locks between themselves and lock A.

[0086] S202, Target Detection and Tracking Group Establishment: Based on the event that any lock in the auxiliary recognition processing chain identifies the homeowner's facial features of the requesting lock, a tracking and recognition group is created with that lock as the group leader. The tracking and recognition group includes all locks along the path from the lock that identified the homeowner's face to the requesting lock. Specifically, the establishment of the target detection and tracking group includes:

[0087] Homeowner Face Recognition Trigger: The tracking and recognition group creation process is triggered when any door lock camera captures a face image and matches the homeowner's face feature vector to the requesting door lock. Tracking and Recognition Group Initialization: The complete path from the door lock that recognized the homeowner's face to the requesting door lock is determined based on the auxiliary recognition processing chain. A tracking and recognition group information table containing the IDs of all door locks along the path is created by the door lock that recognized the homeowner's face. Member Status Confirmation: Based on the response of the group leader door lock to all member door locks' group establishment request messages, the following processing is performed: For door locks that respond normally, they are marked as active members; for door locks that do not respond, a retry mechanism is implemented until the maximum number of retries is reached; for door locks that never respond, they are removed from the tracking and recognition group. Tracking Mode Activation: Based on the event that the number of active members reaches a preset threshold, the tracking and recognition mode of all member door locks is activated, and the reporting target for the homeowner's face feature is set to the group leader door lock.

[0088] Among them, the target detection and tracking identification group is established: After a door lock (let's say E) in the auxiliary identification processing chain detects the face of the homeowner target A at door lock A, door lock E establishes a tracking identification group (including the complete path from door lock E to door lock A), and the tracking identification group is managed by E.

[0089] When lock E captures a face and matches the facial feature vector of a certain owner from lock A, it is considered that the target of tracking and identification has appeared (assuming the target person is A). According to the left auxiliary identification processing chain (Table 5), it can be seen that there are three locks D, C and B between lock E and lock A. First, tracking and identification group information is created locally to record the working status of the locks of the group members and the information of the target person's passage through the locks of the group members.

[0090] Once completed, door lock E sends a "Message to Establish Tracking and Identification Group" to the three door locks DCB. The message includes: identification group creation command word, tracking and identification group ID, tracking and identification group leader ID (ID-E), auxiliary identification source ID (ID-A), group member IDs (ID-D, ID-C, ID-B), identification object ID (A's ID), and identification object feature vector (A's feature vector).

[0091] After receiving the message, door locks D, C, and B locate the row containing the object (A's ID) in the door lock feature vector table (refer to Table 3, but this is each door lock's local data). They then modify the reporting object in the table to the IP address of door lock E—IP-E—and resend the message to the identification module. After processing, they send a confirmation message back to door lock E.

[0092] After receiving the message, lock E considers it to be in normal working condition. If some locks do not reply to the message in a timely manner, abnormal handling is required. Assuming locks D and B reply, lock E records their status as normal in its local table. However, if lock C does not reply, lock E resends the "Establish Tracking and Identification Group" message every second. If C still does not reply after three resends, its status is considered abnormal, and it is removed from the tracking and identification group.

[0093] If two or more tracking and identification group members respond normally, lock E considers the tracking and identification group to have been successfully created and begins tracking and identification. If fewer than two tracking and identification locks respond, the tracking and identification group creation is considered unsuccessful, tracking and identification stops, lock E releases the locally created tracking and identification group information, and logs it. It also sends a tracking and identification group release message to the member locks that have reported successful creation. The message includes the tracking and identification group release command word, the tracking and identification group ID, and the tracking and identification group leader ID (ID-E). After receiving the message, a member lock (e.g., B) modifies the reporting object in its feature vector table to "local" and reissues the tracking and identification group (after detecting lock A's feature vector, it can create the tracking and identification group independently), and replies to lock E with a confirmation message.

[0094] S203, Movement Tracking and Arrival Prediction: Based on the capture time data of the homeowner's face detected sequentially by the door locks of the tracking and identification team members, the team leader door lock calculates the average time interval between the homeowner passing through each door lock to predict the arrival time of the requesting source door lock; specifically, the movement tracking and arrival prediction includes:

[0095] Sequential Movement Detection: Based on the events of the homeowner's face being captured by the group members' door locks in the order of the path, the group leader's door lock receives and records the capture timestamps of each door lock; Time Interval Calculation: Based on the difference in capture timestamps between two consecutive group members' door locks, calculate the average time interval for the homeowner to pass through a single door lock; Abnormal Sequence Handling: Based on the detection of non-sequential movement events, perform the following abnormal handling: request the abnormal door lock to provide a captured image for re-identification and verification; investigate the reason for the failure to recognize the missing door lock; adjust the movement path status based on the verification results; Arrival Time Prediction: Based on the capture timestamp of the last group member's door lock and the average time interval, calculate the estimated time point at which the homeowner will arrive at the requesting source door lock.

[0096] Among them, movement tracking and time prediction: This step explains that after E establishes the tracking and identification group, if door lock E finds that the target is moving in the order of door locks in the auxiliary identification processing chain, then door lock E notifies door lock A to open the door.

[0097] After the tracking and identification group is created, all locks within the group enter the tracking and identification state. If person A passes by lock D, and lock D captures and identifies the face as person A, lock D immediately sends a person passing report message to lock E. The message includes: its own (capturing lock) ID (ID-D), the ID of the identified person (person A), and the capture time. Upon receiving the message, lock E parses the new capturing lock ID, person ID, and capture time, and adjusts the status of lock D in the tracking and identification group to "passed." Similarly, when lock E receives messages from locks C and B, it adjusts the passing status of locks C and B to "passed." At this point, lock E has detected that person A has passed all locks in the tracking and identification group, and according to the auxiliary identification processing chain, lock B is the left neighbor of lock A. Therefore, lock E believes that person A is about to reach lock A. When lock E receives the message from lock B, lock E finds the time points when person A passes through locks E, D, C, and B, and calculates the average time difference for person A passing through each lock (if C is not online, then point C should also be included in the calculation of the average time difference, that is, the time from D to B is divided into two equal segments for calculation). The average time difference is added to the time captured by lock B to obtain the estimated time point when person A arrives at lock A. Lock E then sends an open message to lock A, the message content of which includes: open instruction, person identification result ID, estimated arrival time, tracking identification group ID, tracking identification group leader ID, etc.

[0098] After receiving the unlock message from lock E, lock A parses the various fields, sets a timer for unlocking based on the expected arrival time, and records other information locally for later review. Lock A then replies with an unlock confirmation message to lock E. Upon receiving the message, lock E begins dismantling the tracking and identification group. First, it writes the contents of the local tracking and identification group's memory to local storage and releases the relevant memory information. Simultaneously, it sends an exit tracking and identification group message to locks D, C, and B, containing the exit command word and the tracking and identification group ID. Upon receiving the exit message, locks D, C, and B record the relevant memory information for their local tracking and identification group, release the memory, and simultaneously re-inform the target lock of the feature vector results in their local auxiliary identification mode (after detecting lock A's feature vector, it can autonomously create a tracking and identification group), and reply with a confirmation message to lock E.

[0099] Lock A opens the door at a set time and announces an opening notification voice and a low battery voice message to inform the homeowner (A) that the door has been opened, thus preventing the situation where the door is open but A is unaware of it. It also reminds A to charge the device or replace the battery.

[0100] S204, Remote Door Opening and Status Restoration: Based on the predicted arrival time, the group leader lock sends an opening command to the requesting source lock. The requesting source lock performs a timed opening operation. After the opening is completed, the group leader lock disbands the tracking and identification group and restores the normal identification mode of each lock. Specifically, the remote door opening and status restoration includes: Sending a remote door opening command: Based on the calculated predicted arrival time, the group leader lock sends an opening command containing the predicted time to the requesting source lock; Executing a timed opening: Based on the processing result of the opening command received by the requesting source lock, a timed opening task is set and the opening operation is executed at the predicted time; Status restoration and resource release: Based on the door opening completion event, the following cleanup operations are performed: The group leader lock sends a tracking and identification group disbandment message to all group members; Group member locks restore the local reporting mode of the homeowner's facial features; Release the relevant memory resources of the tracking and identification group; Voice prompt broadcast: Based on the successful door opening event, the requesting source lock broadcasts a successful door opening and a low battery reminder voice message.

[0101] Furthermore, the method also includes an anomaly detection and adaptive adjustment mechanism: adaptive threshold adjustment: based on the event that the group leader's door lock detects a member's door lock recognition failure due to an excessively high similarity threshold, a threshold adjustment suggestion is sent to that member's door lock; dynamic threshold configuration: based on the processing result of the member's door lock receiving the threshold adjustment suggestion, only the similarity threshold of the auxiliary recognition mode is adjusted, while the local recognition threshold remains unchanged; tracking timeout handling: based on the event that the tracking recognition group does not receive a new recognition message within a preset time, the tracking recognition group is automatically disbanded and resources are released; reversal behavior recognition: based on the event that the homeowner's door lock repeatedly appears on the path, the direction of travel is recalculated and the status of the tracking recognition group is updated.

[0102] The anomaly detection and recovery mechanism addresses several issues. While the above process operates under normal conditions, real-world applications may encounter various anomalies requiring handling, including incorrect identification order, duplicate appearances of individuals on a particular door lock, and undetected individuals. We address each anomaly individually.

[0103] 1) Incorrect recognition order:

[0104] After the tracking and identification group is created in the above process, if the order in which person A passes by door lock E is not in accordance with the order set by the tracking and identification group, for example, in the above process, the identification order for person A should be EDCB, but if E receives the message from door lock D and then directly receives the message from door lock B without receiving the person passing by report message from door lock C, door lock E will perform two steps: first, further verify the correctness of B's ​​identification; second, investigate the reason why C was not detected.

[0105] First, lock E verifies the recognition result of lock B. Lock E sends an image request message to lock B. The message content includes: image request command, lock E's ID (ID-E), lock E's address (IP-E), and the image capture time (from the recognition time reported by B). After receiving the message, lock B parses out the command requirements and fields, retrieves the stored image of the corresponding time point and event based on the time, and replies to lock E with the image. After obtaining the image, lock E re-extracts the feature vector and performs a further similarity comparison with the feature vector of the locally captured image of lock A. If the result is still lock A (i.e., the similarity exceeds the threshold, for example, 80%), it is assumed that lock A has arrived at lock B's door. If lock E's check result is not lock A (i.e., the similarity is less than the threshold), it is assumed that lock B has made a mistake in recognition, and it continues to wait for a new recognition result (it is very likely that lock A has not actually arrived at lock B yet).

[0106] In the above steps, if the similarity exceeds the threshold, it means that lock B has correctly identified A, and A has indeed arrived at lock B. This implies that lock C has not correctly identified A, and it is necessary to determine whether the similarity threshold of lock C is set too high. Objectively, it is inappropriate to set the same similarity threshold for every lock. Some locks have good lighting conditions, so the similarity threshold can be set higher, while others have average lighting conditions, so the similarity threshold can be set lower. Therefore, lock E needs to further check the confirmation result of lock C. Lock E sends an anomaly verification message to lock C. The message content includes: anomaly verification message command word, lock E's ID (ID-E) and tracking and identification group ID, and anomaly check time period (between the capture time of lock D and lock B). After receiving the message, lock C begins to check whether the device has the following situations: ① A face has been detected, but the similarity is not high, and there are no other anomalies during the period; ② Within the capture time of locks D and B, there are facial recognition results of the homeowner or other people using lock C; ③ There are anomalies in the camera. After completing the detection, door lock C sends an error verification feedback message to door lock E. The message includes: error code (i.e., ①②③ above), door lock ID-C, and tracking and recognition group ID. If it is case ①, the captured image and capture time information are also required. After receiving the feedback message, if door lock E confirms that door lock C has cases ② and ③, it considers door lock C to be normal and notifies door lock A to open the door. If it confirms that door lock C has case ①, it re-extracts the feature vector from the image in the message. If door lock E's recognition result is A, it adjusts the pass result of door lock C in the tracking and recognition group to "passed," and the tracking and recognition group recognizes normally and notifies door lock A to open the door. At the same time, door lock E sends a "similarity threshold adjustment message" to door lock C (since door lock E's recognition result is A, but door lock C's recognition result is incorrect, it means that door lock C's face recognition threshold setting is too high, and it is recommended that door lock C adjust the threshold). The message includes: threshold adjustment command word, suggested device ID, and suggested adjustment range. After receiving the threshold adjustment suggestion, door lock C parses the content and lowers the feature values ​​related to the auxiliary recognition mode and tracking recognition group to the corresponding threshold. However, it does not adjust the local feature vector threshold of door lock C (the reason for not adjusting the local feature vector threshold is that local personnel often capture frontal faces when comparing, while when assisting other door locks in comparison, they capture more side faces, so adjustments need to be made according to the actual capture needs).

[0107] 2) Target personnel disappear:

[0108] The disappearance of the target person here refers to a situation where the person passes through some door locks but is not detected by the others. This can be caused by several possible reasons: ① Incorrect capture or recognition; ② The person moves between doors midway; ③ The person stays for an extended period of time. To address this, we set a timeout threshold for the tracking and recognition group (e.g., 300 seconds). Once this threshold is exceeded, follow-up recognition ceases. Specifically, when establishing the tracking and recognition group, a maximum target waiting time threshold is set. Each time a message indicating that a target has been identified by a member of the tracking and recognition group is received, the threshold is reset to its initial value (300 seconds). If no new identification message is received within the threshold time, door lock E, after making its local record, directly releases the memory information of the tracking and recognition group and notifies the group members to release their data.

[0109] 3) The target appears repeatedly on a certain door lock:

[0110] If a target repeatedly appears at a certain lock, it means that the target may turn back. Therefore, the direction is determined by combining the next lock where the target appears. For example, in the above identification process, when the target passes through lock E, the pass flag of all tracking identification group members (locks D, C, and B) from lock E to lock A (the requesting identification source device) is set to pass-through. If lock D detects the target, locks C and B are set to pass-through. (Under this mechanism, even if the target passes through some tracking identification group members and then walks back, it can be processed normally without causing identification confusion. This is because if lock C detects the target after lock D, it is considered that the person is continuing to walk towards lock A. If lock E detects the target again, it is considered that the person is leaving. At this time, no further processing is required. The tracking identification group will automatically release after the waiting time exceeds the threshold (e.g., set to 300 seconds).

[0111] This invention utilizes a perimeter door lock tracking and identification method to reduce the rate of power loss in low-power door locks by disabling unnecessary components. This effectively reduces the occurrence of door lock failures in serviced apartments due to power outages, decreases emergency maintenance work caused by charging needs, improves the user experience for apartment guests, and lowers maintenance costs, thereby reducing labor costs. Beneficial effects:

[0112] 1. Effectively solves the problem of door locks failing to open due to rapid power consumption when the battery is low;

[0113] 2. Effectively reduces the timeliness requirements for maintenance when the door lock battery is low, thereby reducing the maintenance costs of serviced apartments;

[0114] 3. Through door lock coordination, we ensure safe and efficient door opening, guaranteeing a good experience during your stay.

[0115] Protection point:

[0116] 1. Working mechanism of the tracking and detection team. When a target appears, after a nearby door lock E detects the face of the homeowner A, E initiates the establishment of a tracking and identification team to track and identify the target in sequence. Based on the average time, the team predicts the time it will take for the person to finally reach the target door lock and finally realizes the entire process of opening the door.

[0117] 2. Adaptive Threshold Adjustment Mechanism. If a door lock (door lock C) fails to accurately identify a target person due to its own high threshold setting, the tracking and identification team leader will re-identify the person and propose an adjustment to the identification threshold for door lock C based on their own identification results. This adjustment will ultimately be adopted by door lock C.

[0118] 3. Dynamic assisted identification processing of the linked list transmission and formation. When the door lock detects that its own power is low, it establishes a complete neighbor relationship table by passing the information from the left and right door locks one by one, and realizes the assisted identification process.

[0119] As can be seen, based on the event that the door lock detects that its own battery level is lower than a preset threshold, an auxiliary identification processing chain list is constructed by passing the information level by level, including the requesting source door lock and the two end door locks. When any door lock in the auxiliary identification processing chain identifies the homeowner's facial features of the requesting source door lock, that door lock acts as the group leader and creates a tracking identification group. Based on the capture time data of the homeowner's face detected sequentially by the member door locks in the tracking identification group, the arrival time of the requesting source door lock is predicted. Based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock. After opening the door, the group leader door lock disbands the tracking identification group and restores the normal identification mode of each door lock. This enables seamless, continuous identity verification and proactive service for the user, improving the robustness and user experience of the smart door lock system.

[0120] Another embodiment of the present invention provides a door lock unlocking system, see [link to relevant documentation]. Figure 4 The system may include:

[0121] Detection module 401 is used for low power detection and auxiliary network construction: based on the event that the door lock detects that its own power is lower than a preset threshold, it shuts down unnecessary functional components and sends an auxiliary identification request message to the adjacent door locks via unicast. It constructs an auxiliary identification processing chain list containing the request source door lock to the two end door locks by passing the message level by level.

[0122] Module 402 is established for target detection and tracking group establishment: based on the event that any lock in the auxiliary recognition processing chain recognizes the homeowner's face feature of the request source lock, the lock is used as the group leader to create a tracking recognition group, which includes all locks on the path from the lock that recognizes the homeowner's face to the request source lock;

[0123] Tracking module 403 is used for movement tracking and arrival prediction: based on the capture time data of the homeowner's face detected sequentially by the door locks of the tracking and identification group members, the group leader door lock calculates the average time interval between the homeowner passing through each door lock and predicts the time point of arrival at the request source door lock;

[0124] The door opening module 404 is used for remote door opening and status recovery: based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock, requesting the source door lock to perform a timed door opening operation. After the door opening is completed, the group leader door lock disbands the tracking and identification group and restores the normal identification mode of each door lock.

[0125] This invention also provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0126] This invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0127] Specifically, the aforementioned electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the aforementioned processor, and the input / output device is connected to the aforementioned processor.

[0128] The above description, based on the embodiments shown in the figures, details the structure, features, and effects of the present invention. The above description is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or equivalent embodiments modified to have equivalent changes, that do not exceed the spirit covered by the specification and figures, should be within the protection scope of the present invention.

Claims

1. A method for unlocking a door lock, characterized in that, The method includes: Low battery detection and auxiliary network construction: Based on the event that the door lock detects that its own battery is lower than a preset threshold, unnecessary functional components are turned off, and auxiliary identification request messages are unicasted to adjacent door locks. An auxiliary identification processing chain list containing the request source door lock to the two end door locks is constructed through hierarchical transmission. Target detection and tracking group establishment: Based on the event that any lock in the auxiliary recognition processing chain identifies the homeowner's facial features of the requesting source lock, a tracking and recognition group is created with that lock as the group leader. The tracking and recognition group includes all locks along the path from the lock that identified the homeowner's face to the requesting source lock. The establishment of the target detection and tracking group includes: Homeowner face recognition triggering: Based on the event that any lock's camera captures a facial image and matches the homeowner's facial feature vector of the requesting source lock, the tracking and recognition group creation process is triggered; Tracking and recognition group initialization: Based on the auxiliary recognition processing chain, the path from the lock that identified the homeowner's face to the requesting source lock is determined. The complete path of the door lock is established by creating a tracking and identification group information table containing the IDs of all door locks along the path from the door lock that recognizes the homeowner's face; Member status confirmation: Based on the response result of the group leader door lock sending the group establishment request message to all member door locks, the following processing is performed: For door locks that respond normally, they are marked as active members; For door locks that do not respond, a retry mechanism is implemented until the maximum number of retries is reached; For door locks that never respond, they are removed from the tracking and identification group; Tracking mode activation: Based on the event that the number of active members reaches a preset threshold, the tracking and identification mode of all member door locks is activated, and the reporting object of the homeowner's facial features is set to the group leader door lock; Movement tracking and arrival prediction: Based on the capture time data of the homeowner's face detected sequentially by the door locks of the tracking and identification team members, the team leader door lock calculates the average time interval between the homeowner passing through each door lock and predicts the time point when the homeowner arrives at the request source door lock; Remote door opening and status recovery: Based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock, requesting the source door lock to perform a timed door opening operation. After the door opening is completed, the group leader door lock disbands the tracking and identification group and restores the normal identification mode of each door lock.

2. The method according to claim 1, characterized in that, The low-power detection and auxiliary network construction includes: Low battery detection and power saving mode activation: Based on the trigger event that the door lock power sensor detects that the power is lower than the preset power threshold, an alarm notification is sent to the operation and maintenance system, and the camera, intelligent algorithm module and structured light camera are turned off at the same time; Auxiliary Request Message Encapsulation: Based on the locally stored left and right neighbor door lock information, encapsulate an auxiliary recognition request message containing the request source door lock ID, request source door lock IP, transmission direction marker, and a list of homeowner face feature vectors; The auxiliary network is constructed step by step: Based on the processing results of the auxiliary recognition request messages received by the adjacent door locks, the following operations are performed: save the homeowner's face feature vector of the requesting source door lock to the local feature vector table; update the auxiliary recognition processing linked list and add the node information of this door lock; forward the updated auxiliary recognition request message to the next adjacent door lock; Auxiliary network integrity verification: Based on the event that the message has been transmitted to the end lock and cannot be transmitted further, verify that the auxiliary identification processing chain list has been constructed.

3. The method according to claim 2, characterized in that, The travel tracking and arrival prediction include: Sequential movement detection: Based on the events of the group members' door locks capturing the homeowner's face in the order of the path, the group leader's door lock receives and records the capture timestamps of each door lock; Time interval calculation: Based on the difference in timestamps of two consecutive group members' door locks, calculate the average time interval between the homeowner passing through a single door lock; Anomaly handling: Based on the detection of non-sequential movement events, perform the following anomaly handling: request the abnormal door lock to provide a captured image for re-identification and verification; investigate the reason for the failure to identify the missing door lock; adjust the movement path status according to the verification results; Arrival Time Prediction: Based on the timestamp of the last member's door lock and the average time interval, calculate the estimated time when the homeowner will arrive at the requesting source door lock.

4. The method according to claim 3, characterized in that, The remote door opening and status restoration include: Remote door opening command transmission: Based on the calculated estimated arrival time, the group leader door lock sends an opening command containing the estimated time to the requesting source door lock; Scheduled door opening execution: Based on the processing result of the door opening command received from the requesting door lock, a scheduled door opening task is set and the door opening operation is executed at the expected time. State restoration and resource release: Based on the door opening completion event, perform the following cleanup operations: The group leader door lock sends a tracking and identification group disbandment message to all group members; the group member door locks restore the local reporting mode of the homeowner's facial features; and release the memory resources related to the tracking and identification group. Voice prompts: Based on the successful door opening event, request the source door lock to announce the successful door opening and low battery reminder.

5. The method according to claim 4, characterized in that, The method also includes an anomaly detection and adaptive adjustment mechanism: Adaptive threshold adjustment: If the group leader's door lock detects that a member's door lock has failed to be recognized due to an excessively high similarity threshold, a threshold adjustment suggestion is sent to that member's door lock. Dynamic threshold configuration: Based on the processing results of the group member's door lock reception threshold adjustment suggestion, only the similarity threshold of the auxiliary recognition mode is adjusted, while the local recognition threshold remains unchanged; Tracking timeout handling: If the tracking and identification group does not receive a new identification message within a preset time, the tracking and identification group will be automatically disbanded and resources will be released. Turnback behavior recognition: Based on the detected event of the homeowner repeatedly appearing at the door lock on the path, the direction of travel is recalculated and the status of the tracking and recognition group is updated.

6. A door lock unlocking system, characterized in that, The system includes: The detection module is used for low battery detection and auxiliary network construction: based on the event that the door lock detects that its own battery is lower than a preset threshold, it shuts down unnecessary functional components and sends an auxiliary identification request message to the adjacent door locks by unicast. It constructs an auxiliary identification processing chain list containing the request source door lock to the two end door locks by passing the message level by level. A module is established for target detection and tracking group establishment: based on the event that any lock in the auxiliary recognition processing chain recognizes the homeowner's facial features of the requesting source lock, a tracking and recognition group is created with that lock as the group leader. The tracking and recognition group includes all locks along the path from the lock that recognized the homeowner's face to the requesting source lock. The establishment of the target detection and tracking group includes: homeowner face recognition triggering: based on the event that any lock's camera captures a face image and matches the homeowner's facial feature vector of the requesting source lock, the tracking and recognition group creation process is triggered; tracking and recognition group initialization: based on the auxiliary recognition processing chain, the locks that recognized the homeowner's face are determined... The complete path to the requesting source lock is used to create a tracking and identification group information table containing the IDs of all locks along the path, based on the lock that recognizes the homeowner's face. Member status confirmation: Based on the response results of the group leader lock sending the group creation request message to all member locks, the following processing is performed: For locks that respond normally, they are marked as active members; for locks that do not respond, a retry mechanism is implemented until the maximum number of retries is reached; for locks that never respond, they are removed from the tracking and identification group; Tracking mode activation: Based on the event that the number of active members reaches a preset threshold, the tracking and identification mode of all member locks is activated, and the reporting object for the homeowner's facial features is set to the group leader lock. The tracking module is used for movement tracking and arrival prediction: based on the capture time data of the homeowner's face detected sequentially by the door locks of the tracking and identification group members, the group leader door lock calculates the average time interval between the homeowner passing through each door lock and predicts the time point of arrival at the request source door lock; The door opening module is used for remote door opening and status recovery: based on the predicted arrival time, the group leader door lock sends an opening command to the requesting source door lock, requesting the source door lock to perform a timed door opening operation. After the door opening is completed, the group leader door lock disbands the tracking and identification group and restores the normal identification mode of each door lock.

7. The system according to claim 6, characterized in that, The detection module is specifically used for: Low battery detection and power saving mode activation: Based on the trigger event that the door lock power sensor detects that the power is lower than the preset power threshold, an alarm notification is sent to the operation and maintenance system, and the camera, intelligent algorithm module and structured light camera are turned off at the same time; Auxiliary Request Message Encapsulation: Based on the locally stored left and right neighbor door lock information, encapsulate an auxiliary recognition request message containing the request source door lock ID, request source door lock IP, transmission direction marker, and a list of homeowner face feature vectors; The auxiliary network is constructed step by step: Based on the processing results of the auxiliary recognition request messages received from adjacent door locks, the following operations are performed: the homeowner's face feature vector of the requesting source door lock is saved to the local feature vector table; Update the auxiliary identification processing chain list and add the node information of this door lock; forward the updated auxiliary identification request message to the next adjacent door lock; Auxiliary network integrity verification: Based on the event that the message has been transmitted to the end lock and cannot be transmitted further, verify that the auxiliary identification processing chain list has been constructed.

8. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method of any one of claims 1-5 when it is run.

9. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method of any one of claims 1-5.

Images