Intelligent door lock joint defense method and system based on repeated behavior recognition and community-level anonymous sharing and intelligent door lock

Abstract

The invention discloses an intelligent door lock joint defense method and system based on repeated behavior recognition and community-level anonymous sharing and an intelligent door lock, and the method comprises the steps: collecting target data of strangers, extracting behavior abstract data, calculating the abnormal frequency of the strangers, and if the abnormal frequency exceeds an abnormal behavior threshold value, judging whether the strangers are abnormal or not; marking a stranger as a concerned object; refusing a remote unlocking instruction, starting a voice alarm, synchronizing a risk event to a user APP interface, prompting a risk score, starting a high-frequency video recording mode, and encrypting and storing behavior fragments; receiving a user feedback result, performing risk level labeling on the concerned object, and synchronously uploading the feedback result to the community sharing pool; when the concerned object appears in front of the door of the other resident again and the matching is successful, triggering a local high warning mode; and if the accumulated risk value of the concerned object exceeds a preset threshold value, bringing the concerned object into a community blacklist candidate pool. According to the invention, multi-user cooperative identification of risk personnel is realized, and the detection precision of potential risks is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent door lock control, and particularly relates to an intelligent door lock joint defense method and system based on repeated behavior recognition and community-level anonymous sharing, an intelligent door lock and a computer readable storage medium. BACKGROUND

[0002] At present, a home security system gradually evolves from a traditional mechanical lock to an intelligent door lock, a video doorbell, a visual intercom, an intelligent monitoring device and the like, and improves home security and use convenience. A typical intelligent door lock device integrates remote unlocking, fingerprint recognition, password verification, visitor reminding and the like, and forms a preliminary security automation system after linkage with a smart phone and a smart home system.

[0003] However, most of the existing intelligent door lock systems are based on a single trigger response mechanism: when a visitor presses a doorbell, approaches a door lock or lingers in front of a door, the system will immediately notify a user or record a video. However, for some potential risk personnel (such as a footmark thief, a malicious salesman or a long-tailed follower) with exploratory and multiple approaching behaviors, a traditional system often cannot form effective identification and continuous monitoring. In addition, since the devices of each family operate relatively independently, there is a lack of cross-user data sharing and behavior linkage mechanism, so even if a user finds a potential risk, it is difficult to timely inform neighbors or a community management party, resulting in that the overall protection of the security system is relatively dispersed and the anti-risk ability is weak.

[0004] Therefore, the prior art still needs to be improved and developed. SUMMARY

[0005] The main purpose of the present application is to provide an intelligent door lock joint defense method and system based on repeated behavior recognition and community-level anonymous sharing, an intelligent door lock and a computer readable storage medium, which aims to solve the problem that the intelligent door lock lacks a linkage mechanism when performing risk identification in the prior art, resulting in that the overall protection of the security system is relatively dispersed and the anti-risk ability is weak.

[0006] To achieve the above purpose, the present application provides an intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, which comprises the following steps:

[0007] When a stranger in front of a door is identified, target data is collected, behavior summary data is extracted according to the target data, an abnormal frequency of the appearance of the stranger is calculated, and if the abnormal frequency exceeds an abnormal behavior threshold, the stranger is marked as an object of attention;

[0008] When the stranger is marked as a focus object, reject the remote unlocking instruction, start the voice warning, synchronize the risk event to the user APP interface, prompt the risk score, start the high-frequency recording mode and encrypt the behavior fragment;

[0009] After prompting the user about the focus object, receive the user's feedback results based on the user feedback mechanism, label the risk level of the focus object, and synchronize the feedback results to the community sharing pool to participate in the formation of the community risk portrait;

[0010] When the focus object appears in front of other residents' doors again, automatic matching is performed, and if the matching is successful, the local high alert mode is triggered;

[0011] If the cumulative risk value of the focus object exceeds the preset threshold, the focus object is included in the community blacklist candidate pool.

[0012] Optionally, the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, wherein when a stranger appears in front of the door, target data is collected, behavior summary data is extracted from the target data, the abnormal frequency of the stranger's appearance is calculated, and if the abnormal frequency exceeds the abnormal behavior threshold, the stranger is marked as a focus object, specifically including:

[0013] When the front-end recognition device recognizes that a stranger appears in front of the door, target data is collected, wherein the target data includes video stream and displacement trajectory data;

[0014] The edge computing module is used to pre-process and analyze the video stream and the displacement trajectory data to extract behavior summary data;

[0015] The abnormal frequency of the stranger's appearance is calculated according to the behavior summary data, and it is judged whether the abnormal frequency exceeds the abnormal behavior threshold;

[0016] If the abnormal frequency exceeds the abnormal behavior threshold, the stranger is marked as a focus object.

[0017] Optionally, the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, wherein the behavior summary data includes personnel residence time, action category, motion trajectory and walking frequency.

[0018] Optionally, the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, wherein the calculation of the risk score is:

[0019] RiskScore=w1×F+w2×B+w3×U;

[0020] Wherein, RiskScore represents the risk score, F represents the frequency score, B represents the behavior characteristic score, U represents the historical feedback influence factor, w1, w2 and w3 represent different weights.

[0021] Optionally, the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, wherein the risk level includes non-risk, high risk and ignore this time and continue to observe.

[0022] Optionally, the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, wherein the local high alert mode includes strong warning playing, rapid shooting and uploading, and synchronization to the property and community management background.

[0023] Optionally, the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, wherein the cumulative risk value is calculated as follows:

[0024] CumulativeRisk=Σ(Fi×Wi);

[0025] Wherein, CumulativeRisk represents the cumulative risk value, Fi is the risk score of each occurrence, and Wi is the weight related to the number of households and the building propagation range.

[0026] In addition, in order to achieve the above purpose, the application also provides an intelligent door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing, wherein the intelligent door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing comprises:

[0027] A front-end recognition device is used to identify the appearance of strangers in front of the door, and collect target data, wherein the target data includes video stream and displacement trajectory data;

[0028] An edge computing module is used to extract behavior summary data according to the target data, calculate the abnormal frequency of the appearance of the stranger, and mark the stranger as an attention object if the abnormal frequency exceeds the abnormal behavior threshold;

[0029] A security linkage unit is used to refuse the remote unlocking instruction when the stranger is marked as an attention object, start voice warning, synchronize the risk event to the user APP interface, prompt the risk score, start the high-frequency recording mode and encrypt the behavior segment;

[0030] A user feedback terminal is used to prompt the attention object to the user, and receive the feedback result of the user based on the user feedback mechanism;

[0031] A community sharing database is used to receive the feedback result of the user, and participate in the formation of the community risk portrait according to the feedback result;

[0032] A subscription identification module is configured to automatically match when the concerned object appears in front of other households again, and trigger a local high alert mode if the matching is successful.

[0033] A blacklist marking module is configured to include the concerned object in a community blacklist candidate pool if the cumulative risk value of the concerned object exceeds a preset threshold.

[0034] In addition, to achieve the above-mentioned purpose, the application also provides an intelligent door lock, wherein the intelligent door lock comprises a memory, a processor, and an intelligent door lock joint defense program based on repeated behavior identification and community-level anonymous sharing stored on the memory and capable of running on the processor, and the intelligent door lock joint defense program based on repeated behavior identification and community-level anonymous sharing, when executed by the processor, implements the steps of the intelligent door lock joint defense method based on repeated behavior identification and community-level anonymous sharing.

[0035] In addition, to achieve the above-mentioned purpose, the application also provides a computer readable storage medium, wherein the computer readable storage medium stores an intelligent door lock joint defense program based on repeated behavior identification and community-level anonymous sharing, and the intelligent door lock joint defense program based on repeated behavior identification and community-level anonymous sharing, when executed by a processor, implements the steps of the intelligent door lock joint defense method based on repeated behavior identification and community-level anonymous sharing.

[0036] In the application, when a stranger appears in front of a door, target data is collected, behavior summary data is extracted according to the target data, the abnormal frequency of the appearance of the stranger is calculated, and the stranger is marked as a concerned object if the abnormal frequency exceeds the abnormal behavior threshold value; after the stranger is marked as a concerned object, a remote unlocking instruction is rejected, a voice warning is started, a risk event is synchronized to a user APP interface, a risk score is prompted, a high-frequency video mode is started and a behavior segment is encrypted and stored; after the concerned object is prompted to the user, a feedback result of the user is received based on a user feedback mechanism, the concerned object is marked with a risk level, and the feedback result is synchronized and uploaded to a community sharing pool to participate in community risk portrait formation; when the concerned object appears in front of other households again, automatic matching is performed, and a local high alert mode is triggered if the matching is successful; and if the cumulative risk value of the concerned object exceeds a preset threshold, the concerned object is included in a community blacklist candidate pool. The application establishes an anonymized community sharing mechanism, realizes multi-household collaborative identification of risk personnel, and improves the detection accuracy of potential risks. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is a flowchart of a preferred embodiment of the intelligent door lock joint defense method based on repeated behavior identification and community-level anonymous sharing of the application;

[0038] Figure 2 is the schematic diagram of the entire implementation process in the preferred embodiment of the intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing of the application;

[0039] Figure 3 is the composition schematic diagram of the intelligent door lock system in the preferred embodiment of the intelligent door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing of the application;

[0040] Figure 4 is the structure diagram of the preferred embodiment of the intelligent door lock of the application. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical scheme and advantages of the application more clear and explicit, the application will be further described in detail below with reference to the drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the application and not to limit the application.

[0042] The existing intelligent door lock and home security system have obvious technical defects and deficiencies in the following aspects: single trigger recognition, ignoring behavior patterns: only one approach or stay in front of the door event is recognized, it is difficult to find the behavior characteristics of some people appearing multiple times in a short period of time, with the tendency of stepping on the dot or harassment; Lack of behavior frequency judgment logic: most current systems do not have built-in "appearance frequency / path behavior modeling" mechanism, and cannot establish a behavior portrait through a time axis. Single door lock linkage strategy: the existing door lock only provides video or notification function for "strangers", and cannot adjust the unlocking, warning and other strategies according to different risk levels; Neighborhood data island problem: each smart device only manages the data in front of its own door, cannot realize cross-housing sharing of risk information, and cannot establish a community-level "suspicious object cooperative identification and response" mechanism; User feedback cannot participate in risk model construction: although users can manually mark "strangers" or "no need to respond", these operations will not affect the overall learning of the system or the community judgment; High privacy risk, data cannot be shared: in the traditional way, if the picture or face data is directly shared, it is easy to cause privacy leakage and related risks, hindering the establishment of community joint defense mechanism.

[0043] The intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing of the preferred embodiment of the application, as shown in Figure 1 and Figure 2 The intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing includes the following steps:

[0044] Step S10, when a stranger appears in front of the door, collect target data, extract behavior summary data according to the target data, calculate the abnormal frequency of the appearance of the stranger, and if the abnormal frequency exceeds the abnormal behavior threshold, mark the stranger as an object of interest.

[0045] Specifically, the front-end recognition device can be a camera (which can be embedded in a smart door lock / peephole), a thermal detection sensor, or a human body detection sensor. When the front-end recognition device identifies the appearance of a stranger in front of the door, target data is collected, wherein the target data includes a video stream and displacement trajectory data. The video stream and the displacement trajectory data are preprocessed and analyzed using an edge computing module to extract behavior summary data (such as stay time, action, and walking pattern). Behavior summary data refers to a structured feature extraction process of the collected raw data, and the result includes time-space behavior characteristics such as personnel stay duration, action category, motion trajectory, and walking frequency, and also includes specific action labels such as whether to lower one's head and look around, whether to approach repeatedly, and whether to knock on the door. Behavior summary data not only records "where and what route was walked", but also includes "what action was done" and "how many times it was done" and other dynamic indicators, which are used for subsequent risk modeling and sharing. The abnormal frequency of the appearance of the stranger is calculated based on the behavior summary data, for example, the appearance frequency is calculated based on a set "risk detection window" (such as 2 appearances within 48 hours), and it is judged whether the abnormal frequency exceeds an abnormal behavior threshold. If the abnormal frequency exceeds the abnormal behavior threshold (for example, 3 times), the stranger is marked as an object of attention. If the abnormal frequency does not exceed the abnormal behavior threshold, regular processing can be performed.

[0046] Step S20, when the stranger is marked as an object of attention, the remote unlocking instruction is rejected, a voice warning is started, the risk event is synchronized to the user APP interface, the risk score is prompted, the high-frequency recording mode is started, and the behavior segment is encrypted and stored.

[0047] Specifically, after the local response and linkage strategy system identifies the "object of attention", the local response mechanism is automatically adjusted, the remote unlocking instruction is rejected, the voice warning "You have been identified by the system, please do not approach" is started, the event is automatically synchronized to the user App interface, the risk score is prompted, the high-frequency recording mode is started, and the behavior segment is encrypted and stored. The high-frequency recording mode refers to that after the system identifies the object of attention, the camera frame rate and resolution are improved, for example, from 5 fps to 15 fps, and the recording is continued for at least 1 minute or more to ensure that the complete abnormal behavior process is obtained.

[0048] The calculation of the risk score is as follows:

[0049] RiskScore = w1xF + w2xB + w3xU;

[0050] Wherein, RiskScore represents the risk score, F represents the frequency score (such as 3 times in 24 hours weighted as high), B represents the behavior characteristic score (such as whether to knock the door, observe the camera, etc.), U represents the historical feedback influence factor (such as whether other users in the community mark as black list), w1, w2 and w3 represent different weights, and the weights can be configured according to the scene.

[0051] Step S30, after prompting the user with the object of interest, receiving the feedback result of the user based on the user feedback mechanism, labeling the risk level of the object of interest, and synchronously uploading the feedback result to the community sharing pool to participate in the formation of the community risk portrait.

[0052] Specifically, based on the user feedback mechanism, the user can choose to label the risk level of the object of interest, and the risk level includes non-risk, high risk, and ignore once and continue to observe, for example, white list (such as delivery personnel) → recorded as non-risk; black list (such as sales personnel) → recorded as high risk; ignore once → continue to observe the behavior frequency; the feedback result of the user is synchronously uploaded to participate in the formation of the community risk portrait; based on the community anonymous sharing mechanism, the face feature code and behavior summary data are extracted by using irreversible encryption (irreversible encryption means that the original image is processed by using a hash algorithm (such as SHA256) or feature extraction coding (such as face embedding vector of deep learning) to make it impossible to restore the original image from the encrypted result, thereby protecting privacy, for example: the face image is coded as a 128-dimensional vector as a unique fingerprint for identification) and uploaded to the community sharing pool; wherein, the anonymous features include “face feature code” and “behavior summary data”, wherein the face feature code is used for identity anonymization comparison, and the behavior summary is used to judge the risk behavior mode, and the two together constitute the “anonymous identification features” of the system; the face feature code: the face image is extracted into a 128-dimensional or 256-dimensional embedding vector by a deep neural network model (such as ArcFace), the vector has uniqueness and can be used for comparison and identification, but cannot be inversely deduced from the original image, thereby realizing privacy protection; behavior summary data: structured behavior trajectory and behavior mode coding, for example {stay for 60s, pace 2 times, upper body forward}, such data does not contain identifiable identity information and is suitable for anonymous sharing.

[0053] The shared data of the community sharing pool includes: frequency of occurrence; risk level; number of related households (the number of related households refers to the number of times the same “object of interest” is recorded by multiple household devices in the system, representing how many households the person has appeared in front of, for quantifying the risk propagation range); access code (desensitized).

[0054] Step S40, when the object of interest appears in front of other households again, automatic matching is performed, and if the matching is successful, a local high alert mode is triggered.

[0055] Specifically, the subscription identifies the shared library of the building / floor to which the default subscription of each household device belongs, and the system uploads the shared information, wherein the fields such as the number of related households and the risk level are collectively referred to as risk data in the drawings, which are used for community sharing and linkage response strategy; if the matching is successful, the local high alert mode will be triggered according to the community cumulative risk value, wherein the local high alert mode includes strong warning playing (strong warning playing refers to issuing a high-intensity voice warning (such as "please do not approach, this area has been guarded") through a speaker in a door lock or a camera module, which has an active deterrent effect, and the volume and playing content can be automatically adjusted according to the risk level), rapid shooting and uploading (high-frequency video clips are stored in the local by default, which are used for evidence), and synchronization to the property and community management background.

[0056] Step S50, if the cumulative risk value of the concerned object exceeds the preset threshold, the concerned object is included in the community blacklist candidate pool.

[0057] Specifically, the cumulative risk value is based on weighted accumulation of multiple dimensions, for example, the calculation of the cumulative risk value is as follows:

[0058] CumulativeRisk = Σ (Fi x Wi) ;

[0059] Wherein, CumulativeRisk represents the cumulative risk value, Fi is the risk score of a certain occurrence, and Wi is the weight related to the number of households and the building propagation range.

[0060] Once the cumulative value exceeds the preset threshold (such as 100 points), it enters the community blacklist candidate pool.

[0061] Further, it can also be linked with the property system to provide an API or platform docking interface for the property to receive community high-risk personnel clues, cooperate with offline patrol, the system supports docking with the property management platform, provides a "high-risk object list" and a "time and place distribution map" for security personnel to conduct offline key patrol, and the system can also support automatic video recording in the controlled area to improve the efficiency and accuracy of offline patrol, forming a community patrol closed loop.

[0062] Advantages of the present application:

[0063] (1) Realize repeated behavior recognition: through trajectory modeling and time window detection, automatic recognition of "frequent appearing personnel" is realized, and the detection accuracy of potential risks is improved.

[0064] (2) Enhance local protection response: the object risk level can be used to adjust the door lock control, voice broadcast and other strategies to form a layered response and improve the active deterrent effect.

[0065] (3) Break the information island, realize the community cascade defense: Establish an anonymized community sharing mechanism, realize multi-household collaborative identification of risk personnel, and improve the overall protection level.

[0066] (4) Realize controllable anonymous sharing and protect privacy safety: Use the coded face feature and behavior summary sharing mechanism to avoid direct sharing of original images and avoid privacy problems.

[0067] (5) Build a self-learning feedback loop: The system can optimize the behavior recognition strategy based on user feedback to realize system adaptive evolution.

[0068] (6) Strong expansibility, suitable for multiple types of intelligent devices and scenes: The scheme has good universality and expansion ability, and is suitable for various environments such as families, communities, smart buildings, etc.

[0069] Further, as shown in Figure 3 based on the above-mentioned intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, the present application also correspondingly provides an intelligent door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing, wherein the intelligent door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing comprises:

[0070] a front-end recognition device for identifying the appearance of a stranger in front of a door and collecting target data, wherein the target data includes video stream and displacement trajectory data;

[0071] an edge computing module supporting behavior trajectory recognition and frequency analysis, for extracting behavior summary data from the target data, calculating the abnormal frequency of the appearance of the stranger, and marking the stranger as an object of interest if the abnormal frequency exceeds the abnormal behavior threshold;

[0072] a security linkage unit for rejecting a remote unlocking instruction when the stranger is marked as an object of interest, starting a voice warning, synchronizing a risk event to a user APP interface, prompting a risk score, starting a high-frequency recording mode and encrypting behavior segments, for example, including a door lock control module, a voice broadcast module, a lighting / camera linkage module;

[0073] a user feedback terminal for prompting the object of interest to the user and receiving user feedback results based on a user feedback mechanism, including mobile phone APP, applet, voice assistant and other feedback platforms;

[0074] a community sharing database for receiving user feedback results, participating in community risk portrait formation according to the feedback results, and uploading de-identified risk labels and behavior codes;

[0075] The subscription identification module is used to automatically match the subject of interest when it appears at other residents' doors again. If a match is successful, the local high-alert mode is triggered, and the local comparison engine can quickly identify the subscribed risk tags;

[0076] The blacklist marking module is used to include the object of concern into the community blacklist candidate pool if the cumulative risk value of the object of concern exceeds a preset threshold.

[0077] Furthermore, if Figure 4 As shown, based on the above-mentioned smart door lock joint defense method and system based on repeated behavior recognition and community-level anonymous sharing, the present invention also provides a smart door lock, which includes a processor 10, a memory 20 and a display 30. Figure 4 Only some components of the smart door lock are shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components may be implemented instead.

[0078] In some embodiments, the memory 20 may be an internal storage unit of the smart door lock, such as a hard disk or memory of the smart door lock. In other embodiments, the memory 20 may also be an external storage device of the smart door lock, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the smart door lock. Furthermore, the memory 20 may also include both an internal storage unit of the smart door lock and an external storage device. The memory 20 is used to store application software and various types of data installed in the smart door lock, such as the program code for installing the smart door lock. The memory 20 may also be used to temporarily store data that has been output or is about to be output. In one embodiment, the memory 20 stores a smart door lock inter-defense program 40 based on repeated behavior recognition and community-level anonymous sharing. The smart door lock inter-defense program 40 based on repeated behavior recognition and community-level anonymous sharing can be executed by the processor 10, thereby realizing the smart door lock inter-defense method based on repeated behavior recognition and community-level anonymous sharing in this application.

[0079] In some embodiments, the processor 10 can be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing.

[0080] The display 30 can be an LED display, a liquid crystal display, a touch liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, etc. in some embodiments. The display 30 is used to display information of the smart door lock and to display a visualized user interface. The components 10-30 of the smart door lock communicate with each other through a system bus.

[0081] In an embodiment, the following steps are implemented when the processor 10 executes the smart door lock joint defense program 40 based on repeated behavior recognition and community-level anonymous sharing in the memory 20:

[0082] When a stranger appears in front of the door, target data is collected, behavior summary data is extracted according to the target data, the abnormal frequency of the appearance of the stranger is calculated, and if the abnormal frequency exceeds the abnormal behavior threshold, the stranger is marked as an object of attention;

[0083] When the stranger is marked as an object of attention, a remote unlocking instruction is rejected, a voice warning is started, a risk event is synchronized to a user APP interface, a risk score is prompted, a high-frequency recording mode is started, and a behavior segment is stored in an encrypted manner;

[0084] After the object of attention is prompted to the user, feedback results of the user are received based on a user feedback mechanism, the object of attention is labeled with a risk level, and the feedback results are synchronized and uploaded to a community sharing pool to participate in community risk portrait formation;

[0085] When the object of attention appears again in front of the door of other households, automatic matching is performed, and if the matching is successful, a local high alert mode is triggered;

[0086] If the cumulative risk value of the object of attention exceeds a preset threshold, the object of attention is included in a community blacklist candidate pool.

[0087] Among them, when a stranger appears in front of the door, target data is collected, behavior summary data is extracted according to the target data, the abnormal frequency of the appearance of the stranger is calculated, and if the abnormal frequency exceeds the abnormal behavior threshold, the stranger is marked as an object of attention, which specifically includes:

[0088] When a front-end identification device identifies that a stranger appears in front of the door, target data is collected, wherein the target data includes video stream and displacement trajectory data;

[0089] The video stream and the displacement trajectory data are preprocessed and analyzed by using an edge computing module to extract behavior summary data;

[0090] calculating an abnormal frequency of the stranger appearing according to the behavior summary data, and determining whether the abnormal frequency exceeds an abnormal behavior threshold value;

[0091] if the abnormal frequency exceeds the abnormal behavior threshold value, marking the stranger as an object of attention.

[0092] The behavior summary data comprises a personnel residence duration, an action category, a motion trajectory, and a walking frequency.

[0093] The risk score is calculated as follows:

[0094] RiskScore=w1×F+w2×B+w3×U;

[0095] wherein RiskScore represents a risk score, F represents an appearance frequency score, B represents a behavior feature score, U represents a historical feedback influence factor, and w1, w2, and w3 represent different weights.

[0096] The risk level comprises non-risk, high risk, and this-time ignoring and continuing to observe.

[0097] The local high alert mode comprises strong warning playing, rapid shooting and uploading, and synchronization to a property and community management background.

[0098] The cumulative risk value is calculated as follows:

[0099] CumulativeRisk=Σ(Fi×Wi);

[0100] wherein CumulativeRisk represents a cumulative risk value, Fi is a risk score of a certain appearance, and Wi is a weight related to a number of households and a building spread range.

[0101] The application further provides a computer readable storage medium, wherein the computer readable storage medium stores an intelligent door lock joint defense program based on repeated behavior recognition and community level anonymous sharing, and the intelligent door lock joint defense program based on repeated behavior recognition and community level anonymous sharing is executed by a processor to realize the steps of the intelligent door lock joint defense method based on repeated behavior recognition and community level anonymous sharing.

[0102] In summary, the present application provides an intelligent door lock joint defense method and system based on repeated behavior recognition and community-level anonymous sharing, an intelligent door lock and a storage medium, the method comprising: collecting target data when a stranger appears in front of the door, extracting behavior summary data according to the target data, calculating the abnormal frequency of the appearance of the stranger, and marking the stranger as an object of attention if the abnormal frequency exceeds the abnormal behavior threshold; when the stranger is marked as an object of attention, rejecting a remote unlocking instruction, starting a voice warning, synchronizing a risk event to a user APP interface, prompting a risk score, starting a high-frequency recording mode and encrypting and storing a behavior segment; after prompting the object of attention to the user, receiving a feedback result of the user based on a user feedback mechanism, labeling the object of attention with a risk level, and synchronously uploading the feedback result to a community sharing pool to participate in community risk portrait formation; when the object of attention appears in front of other households again, automatic matching is performed, and if the matching is successful, a local high alert mode is triggered; if the cumulative risk value of the object of attention exceeds a preset threshold, the object of attention is included in a community blacklist candidate pool. The present application establishes an anonymized community sharing mechanism, realizes multi-household collaborative identification of risk personnel, and improves the detection accuracy of potential risks.

[0103] It should be noted that in this paper, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or intelligent door lock including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or intelligent door lock. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of another identical element in the process, method, article or intelligent door lock including the element.

[0104] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware (such as a processor, a controller, etc.), and the program can be stored in a computer-readable computer-readable storage medium, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a magnetic disc, an optical disc, etc.

[0105] It should be understood that the application of the present application is not limited to the above examples, and those skilled in the art can improve or modify the above description, all of which should be within the scope of the appended claims of the present application.

Claims

1. A smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing, characterized in that: The smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing includes: When a stranger appears at the door, target data is collected, behavior summary data is extracted based on the target data, and the abnormal frequency of the stranger's appearance is calculated. If the abnormal frequency exceeds the abnormal behavior threshold, the stranger is marked as an object of attention; When the stranger is marked as a person of interest, the remote unlocking command is rejected, a voice warning is activated, the risk event is synchronized to the user's APP interface, the risk score is prompted, the high-frequency recording mode is turned on, and the behavior clips are encrypted and stored; After notifying the user of the object of concern, the user's feedback is received based on the user feedback mechanism, the risk level of the object of concern is marked, and the feedback results are synchronously uploaded to the community sharing pool to participate in the formation of the community risk profile; When the object of interest appears in front of other residents' doors again, automatic matching is performed. If the match is successful, the local high-alert mode is triggered; If the cumulative risk value of the object of concern exceeds a preset threshold, the object of concern will be included in the community blacklist candidate pool.

2. The smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing according to claim 1 is characterized in that: The method of collecting target data when identifying a stranger at the door, extracting behavior summary data based on the target data, calculating the abnormal frequency of the stranger's appearance, and marking the stranger as an object of attention if the abnormal frequency exceeds the abnormal behavior threshold, specifically includes: When the front-end recognition device identifies a stranger in front of the door, it collects target data, wherein the target data includes video stream and displacement trajectory data; Preprocessing and analyzing the video stream and the displacement trajectory data using an edge computing module to extract behavior summary data; Calculating the abnormal frequency of the stranger's appearance based on the behavior summary data, and determining whether the abnormal frequency exceeds an abnormal behavior threshold; If the abnormal frequency exceeds the abnormal behavior threshold, the stranger is marked as an object of concern.

3. The smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing according to claim 1 or 2 is characterized in that: The behavior summary data includes the length of time a person stays, action type, movement trajectory, and movement frequency.

4. The smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing according to claim 1 is characterized in that: The risk score is calculated as: RiskScore=w1×F+w2×B+w3×U; Among them, RiskScore represents the risk score, F represents the frequency score, B represents the behavioral characteristic score, U represents the historical feedback influence factor, and w1, w2 and w3 represent different weights.

5. The intelligent door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing according to claim 1 is characterized in that: The risk levels include non-risk, high risk, and ignore and continue to observe.

6. The smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing according to claim 1 is characterized in that: The local high alert mode includes strong warning playback, rapid shooting and uploading, and synchronization to the property and community management background.

7. The smart door lock joint defense method based on repeated behavior recognition and community-level anonymous sharing according to claim 1 is characterized in that: The cumulative risk value is calculated as: CumulativeRisk=Σ(Fi×Wi); Among them, CumulativeRisk represents the cumulative risk value, Fi is the risk score of a certain occurrence, and Wi is the weight related to the number of residents and the spread range of the building.

8. An intelligent door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing, characterized by: The smart door lock joint defense system based on repeated behavior recognition and community-level anonymous sharing includes: A front-end recognition device is used to identify a stranger in front of the door and collect target data, wherein the target data includes video stream and displacement trajectory data; an edge computing module, configured to extract behavior summary data based on the target data, calculate an abnormal frequency of the stranger appearing, and mark the stranger as an object of interest if the abnormal frequency exceeds the abnormal behavior threshold; The security linkage unit is used to reject remote unlocking commands, start voice warnings, synchronize risk events to the user app interface, indicate risk scores, start high-frequency recording mode, and encrypt and store behavior clips when the stranger is marked as a person of concern; A user feedback terminal is used to prompt the user of the object of interest and receive the user's feedback result based on the user feedback mechanism; A community shared database for receiving user feedback and participating in the formation of community risk profiles based on the feedback; A subscription recognition module is used to automatically match the target of interest when it appears at other residents' doors again. If the match is successful, a local high-alert mode is triggered; The blacklist marking module is used to include the object of concern into the community blacklist candidate pool if the cumulative risk value of the object of concern exceeds a preset threshold.

9. A smart door lock, characterized in that: The smart door lock includes: a memory, a processor, and a smart door lock inter-defense program based on repeated behavior recognition and community-level anonymous sharing, which is stored in the memory and can be run on the processor. When the smart door lock inter-defense program based on repeated behavior recognition and community-level anonymous sharing is executed by the processor, the steps of the smart door lock inter-defense method based on repeated behavior recognition and community-level anonymous sharing are implemented as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a smart door lock inter-defense program based on repeated behavior recognition and community-level anonymous sharing. When the smart door lock inter-defense program based on repeated behavior recognition and community-level anonymous sharing is executed by the processor, the steps of the smart door lock inter-defense method based on repeated behavior recognition and community-level anonymous sharing are implemented as described in any one of claims 1-7.

Images