An intelligent home security system based on the Internet of Things

Abstract

The application belongs to the technical field of smart home security technology, and specifically discloses a smart home security system based on the Internet of Things, which comprises an image recognition module, a dynamic connection module, a device signal acquisition module and an identity binding module. The image recognition module is used for collecting a face image and extracting face features, comparing the face features with a pre-stored white list, and judging whether a person belongs to a white list person. The dynamic connection module is used for setting a communication connection range as a first range, or expanding the communication connection range to a second range when detecting that there is a non-white list person. The device signal acquisition module is used for scanning and acquiring wireless identification signals of surrounding devices within the communication connection range, extracting multi-dimensional information of the devices, and constructing device soft fingerprints. The identity binding module is used for performing space-time synchronous matching of the face features and the device soft fingerprints, and generating a binding relationship between the face features and the device soft fingerprints. The application has the following advantages: improving the recognition accuracy and tracking capability of the smart door lock, and giving consideration to the endurance and security efficiency.

Description

Technical Field

[0001] This invention relates to the field of smart home security technology, and more specifically, to a smart home security system based on the Internet of Things. Background Technology

[0002] Existing smart locks typically employ two independent authentication chains: facial recognition and fingerprint recognition. A low-resolution fixed-focus camera captures close-up facial images and performs lightweight neural network comparison locally, while a capacitive or semiconductor fingerprint sensor achieves touch-based confirmation through ridge hash matching. Each sensor completes a single verification to trigger unlocking. Due to limitations in battery life and cost, this approach ensures seamless unlocking for whitelisted users within one meter while minimizing battery consumption. However, in this architecture, the camera resolution is insufficient to capture enough facial details beyond one meter, and the fingerprint module relies on user touch. When visitors only briefly stop or observe at the door, the system cannot accurately retain their high-quality facial features, nor does it actively extend its Bluetooth listening range to collect their feature signals. This results in the lock log containing only scattered or blurry images lacking recognizable feature identifiers.

[0003] To address these issues, an IoT-based smart home security system is proposed. Summary of the Invention

[0004] The present invention aims to provide an Internet of Things-based smart home security system to solve or improve the problem that existing smart door locks are unable to obtain relatively rich identifiable features when outsiders do not actively cooperate, resulting in the inability to effectively mark and track the identity of strangers.

[0005] In view of this, a first aspect of the present invention is to provide an Internet of Things-based smart home security system.

[0006] A second aspect of the invention is to provide a method.

[0007] A first aspect of the present invention provides an Internet of Things-based smart home security system, comprising: an image recognition module for acquiring facial images of people in front of a door lock and extracting facial features, comparing the facial features with a pre-stored whitelist to determine whether the person belongs to the whitelist; a dynamic connection module connected to the image recognition module for setting the communication connection range to a first range when the image recognition module detects that only whitelisted people are in front of the door lock, or expanding the communication connection range to a second range when the image recognition module detects that non-whitelisted people are present; a device signal acquisition module connected to the dynamic connection module for scanning and acquiring wireless identification signals of surrounding devices within the communication connection range, extracting multi-dimensional information of the devices to construct a device soft fingerprint; and an identity binding module connected to both the image recognition module and the device signal acquisition module for spatiotemporally synchronized matching of the facial features and the device soft fingerprint, generating a binding relationship between the facial features and the device soft fingerprint, and storing it for personnel identification purposes.

[0008] In any of the above technical solutions, the image recognition module includes: a face detection unit, used to detect faces in the area in front of the door lock in real time when the face is awake, and to determine the position and number of faces; a feature extraction unit, connected to the face detection unit, used to perform feature embedding on the detected faces and extract face feature vectors; and an identity comparison unit, connected to the feature extraction unit, used to perform similarity calculation between the face feature vectors and a pre-stored whitelist, and to determine whether the person corresponding to the face is a person on the whitelist.

[0009] In any of the above technical solutions, the communication connection range of the dynamic connection module also includes a third range. When a device appears in the third range, the face detection unit is activated and the face detection unit is put into a wake-up state.

[0010] In any of the above technical solutions, the spatial size of the third range is smaller than that of the first range, and when the face detection unit is in a wake-up state, the communication connection range is changed from the third range to the first range or the second range.

[0011] In any of the above technical solutions, the dynamic connection module includes: a range decision unit, used to select the first range or the second range based on the result determined by the image recognition module; and a communication parameter adjustment unit, connected to the range decision unit, used to set the power parameters and broadcast period of the communication connection under the first range, the second range, and the third range, respectively.

[0012] In any of the above technical solutions, the device signal acquisition module includes: a signal scanning unit, used to continuously scan signals broadcast by nearby devices within the communication connection range and acquire multi-dimensional information of the device; and a soft fingerprint generation unit, connected to the signal scanning unit, used to construct and store the device soft fingerprint based on the multi-dimensional information including device identity information.

[0013] In any of the above technical solutions, the device signal acquisition module further includes: an identity injection interaction unit, used to actively broadcast an identity handshake request signal and monitor whether surrounding devices issue a response; and an identity authentication response unit, connected to the identity injection interaction unit, used to receive binding identification information sent by the responding device and send the binding identification information as the device identity information to the soft fingerprint generation unit.

[0014] In any of the above technical solutions, the identity binding module includes: a spatiotemporal synchronization unit, used to perform spatiotemporal matching of the facial features and the device soft fingerprint to determine the correlation between the facial features and the device soft fingerprint; and a binding confirmation unit, connected to the spatiotemporal synchronization unit, used to confirm the binding relationship based on the correlation determination result, and to form a binding record by binding information of the facial features and the device soft fingerprint.

[0015] In any of the above technical solutions, the identity binding module further includes: a risk marking unit, connected to the binding confirmation unit, used to mark the corresponding person as a potential risk person and generate alarm information when the binding confirmation unit determines that the binding confidence of the facial feature and the device soft fingerprint is lower than a preset threshold; and a binding database unit, connected to the binding confirmation unit, used to store and manage the binding records.

[0016] A second aspect of the present invention provides a method comprising the following steps: acquiring a facial image of a person in front of a door lock, extracting facial features and comparing them with a whitelist;

[0017] Based on the comparison results, the communication connection range is adjusted to either the first range or the second range; within the adjusted communication connection range, wireless communication signals sent by surrounding devices are scanned, the multi-dimensional information is extracted, and the device soft fingerprint is generated; the device soft fingerprint and the corresponding facial features are spatially matched within the same time window to obtain the binding relationship between the facial features and the device soft fingerprint, and stored as a basis for personnel identification.

[0018] The beneficial effects of this invention compared to the prior art are as follows:

[0019] By integrating image recognition, dynamic communication range adjustment, wireless device signal acquisition, and spatiotemporal binding of face and device soft fingerprints into smart door locks, multi-dimensional enhancements to home security are achieved: The image recognition module pre-identifies individuals, significantly improving the authentication speed and accuracy of whitelisted personnel; the dynamic connection module flexibly adjusts the communication range based on personnel category, reducing power consumption by narrowing the range when only whitelisted members are present, and expanding the range to promptly detect risks when unknown individuals are present, balancing the lock's battery life with security monitoring effectiveness; the device signal acquisition module actively collects multi-dimensional signals from unknown devices within the extended communication range and constructs soft fingerprints, greatly improving the ability to identify and track unfamiliar individuals and their personal electronic devices; the identity binding module accurately matches facial features with device soft fingerprints in spatiotemporal and stores binding records, effectively addressing the limitations of traditional face recognition in capturing device information and low-resolution cameras in accurately identifying unfamiliar individuals. This enables efficient marking of non-whitelisted personnel, timely warnings, and reliable tracing of subsequent events, comprehensively improving the intelligence, accuracy, and security of the home smart security system.

[0020] Additional aspects and advantages of embodiments of the invention will become apparent in the following description or may be learned by practice of embodiments of the invention. Attached Figure Description

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0022] Figure 1 This is a system logic block diagram of the present invention;

[0023] Figure 2 This is the main flowchart of the system of the present invention;

[0024] Figure 3 This is a flowchart of the dual-state dynamic connection module of the present invention;

[0025] Figure 4 This is a flowchart of the device signal acquisition and binding process of the present invention;

[0026] Figure 5 This is a flowchart illustrating the risk management process of the present invention.

[0027] Figure 6 This is a timing diagram of the multi-module linkage of the present invention;

[0028] Figure 7 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0029] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0030] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0031] Please see Figures 1-7 The following describes an Internet of Things-based smart home security system according to some embodiments of the present invention.

[0032] An embodiment of the first aspect of the present invention provides a smart home security system based on the Internet of Things. In some embodiments of the present invention, such as... Figures 1-6 As shown, the smart home security system includes:

[0033] The image recognition module is used to capture facial images of people in front of the door lock and extract facial features. The facial features are then compared with a pre-stored whitelist to determine whether the person belongs to the whitelist.

[0034] The dynamic connection module connects to the image recognition module and is used to set the communication connection range to the first range when the image recognition module detects that only whitelisted personnel are present in front of the door lock, or to expand the communication connection range to the second range when the image recognition module detects that non-whitelisted personnel are present.

[0035] The device signal acquisition module, connected to the dynamic connection module, is used to scan and acquire the wireless identification signals of surrounding devices within the communication connection range, and extract multi-dimensional information of the devices to construct a device soft fingerprint.

[0036] The identity binding module is connected to the image recognition module and the device signal acquisition module respectively. It is used to perform spatiotemporal synchronous matching of facial features and device soft fingerprints, generate the binding relationship between facial features and device soft fingerprints, and store it for personnel identification.

[0037] This invention provides an IoT-based smart home security system. The image recognition module is typically installed on or near a smart door lock and consists of a camera, an image processing chip, and image algorithm components. It automatically activates the camera to capture a facial image when it detects someone approaching in front of the door lock. The camera can be an embedded infrared camera or a wide-angle visible light camera, supporting dual-mode recognition for both day and night. After capturing the image, the system quickly locates the facial region in the image using an integrated face detection algorithm and extracts key facial features using a lightweight embedding algorithm to generate a corresponding high-dimensional facial feature vector.

[0038] Subsequently, the facial feature vector will be sent to the local identity verification process for similarity comparison with the pre-stored whitelist data in the system. The whitelist can be entered by the user in advance through a mobile app or management backend, and includes the facial feature vectors of authorized family members, frequent visitors, etc. The comparison process uses Euclidean distance, cosine similarity, or a learning-based classification network for measurement. When the similarity is higher than a set threshold, the system determines that the person is a whitelist user; otherwise, it marks them as a non-whitelist or unknown person.

[0039] Specifically, when a homeowner approaches the front door with their mobile phone, their facial features are captured and processed by the camera, identifying them as a member of the whitelist. The image recognition module then returns an authenticated person status, providing a basis for subsequent communication strategy adjustments and unlocking authorization. Conversely, if the camera captures an unregistered person (such as a delivery person or a stranger), the system records their facial features and reports them as a non-whitelisted person, triggering the next step of dynamic connection adjustment and risk recording logic.

[0040] The dynamic connection module connects to the image recognition module and dynamically adjusts the parameters of the wireless communication connection based on the recognition results. This ensures the door lock's battery life while enhancing its ability to monitor potentially suspicious individuals. Specifically, the dynamic connection module controls the switching of the communication connection range, enabling real-time adjustment of parameters such as transmission power, scanning frequency, and broadcast period for short-range communication technologies like Bluetooth, Wi-Fi, UWB, or NFC, thereby altering the device's perceptible spatial coverage.

[0041] When the system detects that only whitelisted individuals are present before the door lock, the dynamic connectivity module restricts the communication range to a primary range, i.e., a preset communication parameter configuration with a small coverage distance and low power consumption. For example, the Bluetooth transmission power can be set to -20dBm, and the broadcast period can be set to more than 1 second, maintaining only a short-range, intermittent device listening state. In this mode, the system primarily focuses on the unlocking intent of whitelisted individuals and performs low-frequency identification and interaction with their personal devices such as smartphones and wearable devices. This significantly reduces the power consumption of the communication module and effectively extends battery life without affecting the user experience.

[0042] Accordingly, when the image recognition module detects a non-whitelisted person in front of the door lock, the dynamic connection module immediately adjusts the communication range to a second range—a scanning mode with a larger area, higher power, and higher frequency. In this mode, the system increases the transmission power of modules such as Bluetooth or Wi-Fi to their maximum value, for example, 0~4dBm, shortens the broadcast period to 100ms, and continuously performs high-frequency scanning of the space around the door lock to obtain as many wireless identification signals as possible emitted by the unfamiliar person's personal devices, such as MAC address, SSID, and device model. This data will be used by subsequent modules to construct a soft fingerprint to assist in identity verification and event tracing.

[0043] Specifically, when an unregistered stranger stands at the door observing or lingering, and the image recognition module does not match their facial features in the whitelist, it immediately notifies the dynamic connection module to enter alert mode, expanding the communication range to a radius of 5-8 meters and continuously acquiring broadcast data from all detectable devices in the vicinity. Simultaneously, directional antennas or TOA (Time of Arrival) estimation methods can be used to determine the location of device signals, thus initially establishing a spatiotemporal correspondence between facial images and wireless devices. If a detected device repeatedly appears in multiple incidents, the system can also record it as a suspicious device and suggest that the resident check it or call the police.

[0044] The device signal acquisition module connects to the dynamic connection module to actively or passively scan for wireless identification signals emitted by surrounding devices within the communication connection range. It then analyzes and extracts features from the acquired signals, ultimately generating a device soft fingerprint that can be used for identification and behavioral tracking. The system is responsible for sensing the presence of devices and establishing person-device associations. Especially when unfamiliar individuals have not directly touched the door lock, remote identification and subsequent tracking can still be achieved using information from their personal electronic devices. Specifically, the device signal acquisition module continuously operates within the communication connection range set by the dynamic connection module, listening for short-range identification signals emitted by various wireless devices with broadcast capabilities. These signals can originate from Bluetooth, Wi-Fi Probe Request, NFC tags, UWB ultra-wideband response signals, or other passively transmitting smart terminals. The scanning unit supports parallel monitoring of multiple protocols, enabling cross-platform identification of broadcast behaviors from devices from different manufacturers and real-time extraction of differentiated and identifiable parameters as feature data sources.

[0045] Specifically, when an unfamiliar visitor appears in front of the door lock without actively touching it, the system captures a Wi-Fi detection request automatically sent by their mobile phone. This request contains multiple dimensions of information, including device type, manufacturer signature, MAC address (which can be randomized), broadcast period, and RSSI signal strength. The device signal acquisition module receives this information and sends it to the soft fingerprint generation process.

[0046] Specifically, the raw data collected above is used to construct a multi-dimensional feature vector. This multi-dimensional information includes, but is not limited to, the following aspects:

[0047] Broadcast signal type: Identifies whether it is a Wi-Fi, Bluetooth, NFC, or UWB signal to help determine the device type;

[0048] Signal transmission frequency band: such as 2.4GHz, 5GHz, 6GHz, etc., used to initially infer the communication chip model or protocol version;

[0049] Broadcast field content: such as the device name, service UUID, and vendor identifier included in Bluetooth broadcasts;

[0050] Signal strength variation curve (RSSI trajectory): A graph of the intensity change over time constructed after receiving multiple consecutive periodic signals;

[0051] Signal frequency and period: reflects the device's broadcast interval, which helps to distinguish system settings (such as differences between Android and iOS).

[0052] Signal response latency and interaction mode: such as the device's response delay to scan requests and whether it supports handshake response;

[0053] Relative position estimation of signal source: Based on multi-antenna RSSI or UWB ranging technology, the relative position of the device and the door lock is initially determined;

[0054] MAC address mode: Whether a random MAC address (RPA) is used, the frequency of address changes, etc., reflects the privacy settings preference;

[0055] Broadcast content change trend: used to determine whether it is multiple spoofing attempts by the same device.

[0056] All the aforementioned features will be combined into a highly recognizable soft fingerprint vector. This vector constitutes a unique signal behavior of a device within a specific time period and environment. After hashing, encrypting, or fuzzy matching, this vector can be used in subsequent modules to spatiotemporally bind with image recognition results, thereby establishing a correspondence between an unfamiliar person's facial image and their personal device. Furthermore, the system supports clustering or trajectory analysis of soft fingerprints generated when the same device appears at different doors during multiple time periods. When a device appears at multiple door locks in the same community at different times, it can be automatically marked as a frequently encountered unknown device, alerting administrators to the potential risks of surveillance or unusual loitering.

[0057] The identity binding module is connected to both the image recognition module and the device signal acquisition module. It performs a spatiotemporal synchronous comparison between the facial feature information extracted by the former and the device soft fingerprint information constructed by the latter. Spatially, it determines whether the two originate from the same target individual; temporally, it determines whether the two types of information belong to the same event cycle. This establishes a reliable person-device binding relationship and generates a corresponding identity binding record, serving as a crucial basis for subsequent identification, tracking, and risk analysis. Specifically, when the system detects someone appearing at the door and successfully extracts their facial image feature vector, the identity binding module acquires the set of device soft fingerprints recorded by the device signal acquisition module at that moment. At this time, the system opens a binding time window, the length of which can be adjusted according to the actual deployment strategy, for example, within 2 or 5 seconds, to match device broadcast behaviors that occur almost simultaneously with the image event. Simultaneously, the system also refers to the horizontal position of the face in the image plane in the current video frame and the relative reception angle or RSSI intensity gradient of the device signal to calculate the spatial consistency index between the direction of the face's appearance and the direction of the signal source.

[0058] Specifically, when the image recognition module detects the facial features of a stranger directly in front of the door lock, it records the facial vector features. The device signal acquisition module captures two device signals simultaneously. and ,in The signal strength is stable within a 1-meter range to the left. The signal is coming from behind the door, but it's intermittent. The identity binding module will... and Prioritize binding because the temporal consistency (Δt < threshold), spatial orientation consistency (Δθ < threshold), and signal behavior feature matching degree, such as broadcast frequency and response delay, all show a high correlation. Once the identity binding module determines that the comprehensive score of a certain facial feature and a certain device soft fingerprint exceeds the set threshold Sthresh, the system will confirm the binding is successful and call the binding confirmation submodule to generate a formal identity binding record, which includes: facial feature encoding, original screenshot of the facial image, device soft fingerprint vector, binding timestamp, relative spatial location, binding confidence score, etc. This record will be stored in the binding database, and may be pushed to the user terminal for notification or marking according to the risk policy settings.

[0059] Furthermore, to further enhance the system's recognition capabilities and long-term tracking performance, the identity binding module supports continuous learning of newly bound objects. Specifically, when a device is stably bound to a certain facial feature multiple times, the system automatically marks this binding relationship as a trusted binding, and it can be directly identified as a known object the next time it appears. Conversely, if a device is confused with different faces in front of multiple door locks, the system can mark it as a suspicious device, triggering further image comparison or alarm mechanisms.

[0060] In summary, this invention not only overcomes the three major shortcomings of existing technologies—the difficulty of independently identifying strangers using biometrics, insufficient sampling by low-resolution cameras, and the failure of fixed low-power Bluetooth to detect suspicious devices—but also achieves a closed loop within the same system encompassing identity determination, communication strategy, device perception, and risk confirmation. In whitelist scenarios, the dynamic connection module automatically converges the communication radius, significantly reducing wireless power consumption and extending battery life. When a stranger appears, the communication range instantly expands, allowing the device signal acquisition module to capture their mobile phone / Wi-Fi / UWB wireless identification signals over a larger space and generate multi-dimensional soft fingerprints. The identity binding module then accurately matches these soft fingerprints with corresponding facial features within a microsecond-level time window, forming a traceable human-machine binding record. Thus, the system maintains the seamless and fast unlocking experience for registered users while significantly improving the ability to perceive, track, and alert on suspicious individuals from a distance, achieving simultaneous improvements in battery life optimization, security, and evidence collection reliability.

[0061] In any of the above embodiments, the image recognition module includes:

[0062] The face detection unit is used to detect faces in the area in front of the door lock in real time when the device is awake, and to determine the location and number of faces.

[0063] The feature extraction unit, connected to the face detection unit, is used to perform feature embedding on the detected faces and extract the face feature vector.

[0064] The identity comparison unit, connected to the feature extraction unit, is used to calculate the similarity between the facial feature vector and the pre-stored whitelist to determine whether the person corresponding to the face is a person on the whitelist.

[0065] In this embodiment, when the door lock accelerometer or infrared distance sensor triggers an event indicating someone is approaching, the system first wakes up the camera and starts the face detection unit. Using a lightweight RetinaFace-Lite network, it scans video frames in real-time at approximately 25fps, outputting the bounding box coordinates, confidence score, and number of faces for each image. Next, the feature extraction unit performs 112×112 normalization preprocessing on each cropped face and calls MobileFaceNet-INT8 for embedding inference, obtaining a 128-dimensional face feature vector on average every 30ms. Finally, the identity comparison unit searches for nearest neighbors in a local whitelist feature library of ≤50 people using cosine similarity. If the maximum similarity is ≥0.75, the person is identified as a whitelisted individual; otherwise, they are marked as a non-whitelisted or unknown individual, and the result is sent to the dynamic connection module.

[0066] For example, if a single face is detected with a similarity of 0.88, it is immediately identified as a member of the whitelist. The dynamic connection module maintains low-power Bluetooth (-20dBm) and directly pushes a successful unlock notification, waking up the motor. The next morning, when the courier arrives, the camera recognizes an unfamiliar face with a confidence level of 0.63. The system determines that the courier is not on the whitelist, triggering an upgrade of the communication range from 1m to 8m. Subsequently, the device signal acquisition module detects the courier's BLE broadcast and generates a soft fingerprint. The identity binding module binds this fingerprint to the courier's face within the same time window, recording it as "Unknown Visitor - Soft Fingerprint IDA3BC…", for quick comparison during subsequent review or a second visit.

[0067] In any of the above embodiments, the communication connection range of the dynamic connection module also includes a third range. When a device appears in the third range, the face detection unit is activated and the face detection unit is put into a wake-up state.

[0068] In this embodiment, during long-term standby, the dynamic connection module maintains extremely low-power monitoring within a third range, with a radius of approximately 30–50 cm, Bluetooth transmission power of -30 dBm, and an advertising interval of ≥3 seconds. It passively scans only short-range wireless communication signals close to the door lock. Once any short-range wireless communication signal from any device, including BLE, NFC, or UWB handshake pulses, is captured within the third range, a wake-up command is immediately issued to power on the camera and activate the face detection unit. Simultaneously, the communication parameters are instantly switched to the normal power configuration of the first or second range. Subsequently, the image recognition module performs real-time face detection and whitelist comparison. Based on the comparison results, the dynamic connection module decides whether to maintain low power consumption (whitelist determination) or enter high-alert extended scanning (stranger determination), thereby achieving a closed loop of intelligent wake-up and risk screening with device-first and image-later verification while ensuring battery life.

[0069] For example, when a Bluetooth-enabled phone brings a keychain near the door lock, the third range listens to the BLE extended advertising packet sent by the phone, immediately triggering the camera to wake up. The face detection unit identifies the resident's face as whitelisted within 600ms, and the system maintains low-power communication in the first range and completes the unlocking process, consuming less than 2mAh of power throughout the entire process. Conversely, during the evening rush hour, an unknown phone attempts to pair with the door lock from 40cm away. The third range is again triggered to wake up, but the camera does not detect any whitelisted faces and captures an image of an unfamiliar face. The dynamic connection module then increases the communication power to the second range radius of 5–8m and continuously scans, recording the phone's soft fingerprint and pushing a suspicious approach alarm to the homeowner, achieving rapid identification and security response to potential tailgating or probing behavior.

[0070] In any of the above embodiments, the spatial size of the third range is smaller than that of the first range, and when the face detection unit is in a wake-up state, the communication connection range changes from the third range to the first range or the second range.

[0071] In this embodiment, the third range in the multi-level communication strategy is designed as a minimum, lowest-power close-to-the-listening area radius of approximately 30–50 cm, maintaining ultra-low-power scanning only when the camera is in sleep mode; its physical coverage is significantly smaller than the first range (≈1m) used for normal unlocking and the second range (≈5–8m) used for alerting. When any wireless identification signal is detected within the third range and triggers the camera to wake up, the image recognition module begins to operate. If the face detection unit confirms that the person belongs to the whitelist, the dynamic connection module immediately upgrades the communication parameters from the third range to the first range—maintaining reliable short-range interaction between the resident's mobile phone and the door lock while keeping power consumption low; if the detection result indicates the presence of a non-whitelisted person, the communication connection range is directly upgraded from the third range to the second range, initiating high-power, short-cycle continuous scanning to maximize the capture of stranger device signals for subsequent soft fingerprint and risk analysis processing.

[0072] For example, when the phone establishes a BLE handshake with the door lock within 40cm, the camera is activated after the signal is detected in the third range. Since the camera does not match a whitelisted face, the dynamic connection module immediately skips the first range and directly increases the communication power and scanning frequency to the second range. This allows the door lock to listen at high frequency to the broadcast fields and RSSI trajectory of the delivery person's phone within a 5m radius, which is used to generate a soft fingerprint for a stranger and record the delivery time. Conversely, when the homeowner returns home from get off work on the same evening, the NFC card in the key pouch triggers the third range when it is brought close to the door lock. After the camera is activated, it quickly identifies a whitelisted face. The system then only increases the communication parameters to the first range to complete the unlocking operation. The camera returns to sleep mode, and the communication power also returns to a low level, thus maintaining a seamless and fast unlocking experience while minimizing power consumption.

[0073] In any of the above embodiments, the dynamic connection module includes:

[0074] The range decision unit is used to select either the first range or the second range based on the result determined by the image recognition module.

[0075] The communication parameter adjustment unit, connected to the range decision unit, is used to set the power parameters and broadcast period of the communication connection in the first range, the second range, and the third range, respectively.

[0076] In this embodiment, the range decision unit inside the dynamic connection module receives the identity determination results output by the image recognition module in real time. When the result indicates only whitelisted individuals, it issues an instruction downstream to select a first range covering approximately 1 meter. When the result indicates the presence of non-whitelisted individuals or that face detection has not yet started but an unknown device has been detected in the third range, a second range covering 5–8 meters is selected. Subsequently, the communication parameter adjustment unit dynamically writes to the communication stack register according to the selected range: in the first range, the Bluetooth or Wi-Fi transmission power is set to -20dBm and the broadcast period to 1s; in the second range, the transmission power is increased to +4dBm and the broadcast period is shortened to 100ms; when the camera is in sleep mode and only the close-range listening third range (30–50cm) is retained, the power is reduced to -30dBm and the broadcast period is extended to 3s. The adjustment unit also synchronizes the current power and period parameters with the device signal acquisition module to ensure that the scanning window and broadcast rhythm are strictly matched, achieving a balance between optimal power consumption and maximum risk perception.

[0077] For example, when the camera recognizes a face as a whitelisted user, the range decision unit immediately selects the first range; the communication parameter adjustment unit sets the BLE power to -20dBm and the advertising interval to 1s, consuming only 2mAh to complete the unlocking process. At 3 PM, an unfamiliar salesperson approaches the door, and their phone is detected in the third range. After the camera wakes up, it determines that the salesperson is not on the whitelist, and the range decision unit immediately switches to the second range; the adjustment unit instantly increases the power to +4dBm and reduces the advertising interval to 100ms. The device signal acquisition module continuously acquires multiple frames of broadcast fields from the phone within an 8m radius and generates a soft fingerprint, while simultaneously triggering an alarm push to the resident's phone, achieving highly sensitive monitoring and evidence collection of suspicious visitors.

[0078] Optionally, the face detection unit can also be activated automatically based on infrared detection.

[0079] In any of the above embodiments, the device signal acquisition module includes:

[0080] The signal scanning unit is used to continuously scan for signals broadcast by nearby devices within the communication connection range and collect multi-dimensional information about the devices.

[0081] The soft fingerprint generation unit, connected to the signal scanning unit, is used to construct and store the device soft fingerprint based on multi-dimensional information including device identity information.

[0082] In this embodiment, the signal scanning unit in the device signal acquisition module continuously monitors short-range wireless communication signals broadcast by nearby mobile terminals in a multi-protocol parallel mode within the communication connection range set by the dynamic connection module—whether it is the first range, the second range, or the low-power third range. These signals include Bluetooth, Wi-Fi Probe, UWB pulse, or NFC touch-to-wake signals. The unit analyzes in real time multi-dimensional information such as signal type, carrying frequency band, broadcast field content, RSSI change curve over time, broadcast period and frequency of occurrence, and signal source angle of arrival (AoA) or time difference of arrival (TDoA). These raw features are then fed into the soft fingerprint generation unit: this unit first standardizes and hashes each feature, then concatenates them into a fixed-length vector V_device according to preset weights. If the signal contains a binding identifier actively responded to by the device (such as an App authentication token or UWBSession-ID), it is written into the vector as a strong identity field. The final generated soft fingerprint also includes a timestamp, door lock ID, and range marker, is stored in a local circular buffer, and periodically synchronized to the cloud for use as basic data for subsequent person-device binding, recurring retrieval, and abnormal trajectory analysis.

[0083] For example, when the system is in whitelist low-power mode, the resident's mobile phone sends a BLE broadcast once per second; the signal scanning unit captures the broadcast packet at -20dBm power, parses the Apple vendor field 0x004C, the local random MAC, TX-Power, and a stable RSSI ≈ -55dBm, and records the broadcast period as 960ms; the soft fingerprint generation unit outputs the fingerprint vector V_A based on this, marking it as a trusted device. Then at 5 pm, an unfamiliar Android phone lingers in front of the door, its Wi-FiProbeRequest appears every 100ms and its RSSI rises sharply from -70dBm to -48dBm; the signal scanning unit simultaneously captures its BLE extended advertisement, parses the vendor field 0x00E0, the random MAC change period of 15min, the RSSI fluctuation characteristic σ=6dB, and the movement direction angle of -22°. The soft fingerprint generation unit integrates multi-protocol features to generate a V_B and tag it with an unknown device. If the device is recorded by different door locks again within three days, the system can automatically determine it as a suspicious repeat visitor based on the high similarity of the V_B, push risk alerts to residents, and provide one-click access to the time, location, and associated facial screenshots of its previous visits, thus achieving accurate tracking and evidence collection.

[0084] In any of the above embodiments, the device signal acquisition module further includes:

[0085] The identity injection interaction unit is used to actively broadcast identity handshake request signals and monitor whether surrounding devices respond.

[0086] The identity authentication response unit, connected to the identity injection interaction unit, is used to receive the binding identification information sent by the responding device and send the binding identification information as the device identity information to the soft fingerprint generation unit.

[0087] In this embodiment, in addition to passive scanning, the device signal acquisition module also integrates an identity injection interaction unit and an identity authentication response unit, jointly constructing an active identification link of device active handshake—binding identifier injection—trusted identity reinforcement. Specifically, the identity injection interaction unit takes the door lock as the central node and periodically broadcasts a handshake request packet via BLEGATT, a custom Wi-FiProbe, or an optional UWBBlink frame every 3 seconds. This packet carries the door lock's unique ID, a one-time random number, and a timestamp. If the surrounding mobile terminals have a companion app installed or an authorized SDK embedded, upon receiving a handshake request, they will automatically generate binding identification information with a dynamic signature and send it back via the same protocol, such as a device unique token, user account hash, or security chip signature. The identity authentication response unit is responsible for listening to and parsing the above-mentioned returned data, verifying the validity of the signature, the consistency of the random number, and the legality of the time window. Then, the extracted binding identification field is added as a strong identity feature to inject into the soft fingerprint generation process, so that the soft fingerprint vector, which was originally based on the broadcast field and RSSI, is reinforced with an encrypted level of identity. Thus, without relying on the public disclosure of system-level hardware identifiers, it can achieve trusted identification of its own app users or authorized guest devices and reduce the risk of uniqueness loss caused by MAC randomization.

[0088] For example, a new mobile phone is registered through a home door lock app. The phone then runs a background service. When the phone approaches the door lock again and receives a BLE handshake request from the identity injection interaction unit, the app immediately reads the random number `-timeNonce` from the request, performs a SHA-256 signature on the device's private key in the local security module, encapsulates it into a binding identifier `Token_0x9A17…`, and sends it back within 200ms. After the identity authentication response unit verifies the signature, it concatenates `Token_0x9A17…` with simultaneously captured broadcast fields, RSSI curves, and other features to generate a strengthened soft fingerprint `V_owner`, and writes a high-trust whitelist device label for the phone in the database. Two hours later, a food delivery rider visits. Their phone does not have the accompanying app installed and does not respond to the handshake request. The system can still passively scan and extract their soft fingerprint `V_rider`, but due to the lack of a binding identifier field, it is marked as an unknown device. If the rider delivers food multiple times and obtains temporary visitor authorization, they can be guided to install a lightweight mini-program to complete a one-time binding, allowing subsequent visits to be quickly identified and recorded as a low-risk visitor device via the token.

[0089] In any of the above embodiments, the identity binding module includes:

[0090] The spatiotemporal synchronization unit is used to perform spatiotemporal matching of facial features and device soft fingerprints to determine the correlation between facial features and device soft fingerprints.

[0091] The binding confirmation unit, connected to the spatiotemporal synchronization unit, is used to confirm the binding relationship based on the correlation determination result and form a binding record by binding information of facial features and device soft fingerprint.

[0092] In this embodiment, the identity binding module performs a dual comparison of the facial feature vector output by the image recognition module and the soft fingerprint vector generated by the device signal acquisition module at millisecond-level time axis and sub-meter-level spatial resolution through the pipeline of spatiotemporal synchronization unit → binding confirmation unit. The spatiotemporal synchronization unit first establishes a Δtwindow (typically 2 seconds) for each face and queries the soft fingerprint set within the window. Then, it uses the face center angle θ_f provided by the door lock camera, the angle of arrival carried in the soft fingerprint, or the RSSI gradient to estimate the angle θd, calculates the temporal consistency Ct=exp(-|Δt| / τ) and spatial consistency Cs=exp(-|θf–θd|² / σ²), and superimposes the historical confidence weight Wh of the soft fingerprint to obtain the comprehensive score S=α·Ct+β·Cs+γ·Wh. The binding confirmation unit selects the soft fingerprint vector with the highest comprehensive score for each face and S≥thresh as the pairing object. Once the binding relationship is confirmed, a binding record containing "face vector ID, soft fingerprint ID, timestamp, spatial coordinates, and comprehensive score" is immediately written and synchronized to the local circular buffer and cloud log chain. If all soft fingerprint scores are lower than the threshold, an unbound flag is output to provide potential alarm basis for subsequent risk modules.

[0093] For example, the door lock camera captures the face of an unfamiliar visitor and generates a vector F_x; the spatiotemporal synchronization unit opens a 2-second matching window and receives the device's soft fingerprint V_p (RSSI rises, angle -5°) after 0.3 seconds. Δt = 0.3s, θf = -7°, θd = -5° are calculated, resulting in Ct ≈ 0.86 and Cs ≈ 0.94. Since Vp is being captured for the first time, Wh = 0.5. The overall score S = 0.86 is obtained by weighting α = 0.4, β = 0.4, and γ = 0.2. The system sets a threshold S_thresh = 0.80, which the binding confirmation unit uses to determine the binding accuracy. Mapping is valid; records of unfamiliar visitors to the bound entries. Device SoftID#9A17… Three days later, the same soft fingerprint reappeared in front of another resident's door, and the camera captured a similar face. The spatiotemporal synchronization unit improved the Wh to 0.9 through historical correlation self-learning, and the comprehensive score rose to 0.92. The system automatically associated the soft fingerprint with the repeatedly appearing stranger tag and pushed it to the community security platform so that the administrator could pay close attention or intervene later.

[0094] Specifically, the binding steps for the binding confirmation unit include:

[0095] Step 1: Calculate five differences between each face and each device signal: 1) the time difference between the face appearance time and the device's initial detection time; 2) the azimuth difference between the face orientation angle estimated by the camera and the wireless signal arrival angle; 3) the dynamic difference between the rate of change of the face bounding box size over time and the rate of change of signal strength over time; 4) the cosine difference between the aligned face feature vector and the device fingerprint vector; 5) the divergence between the historical occurrence probability distributions of the face and the device, using the following formulas:

[0096]

[0097] In the formula, Time of appearance of human face Time between the first detection of the device signal Time difference; The standard deviation of the time difference within the recent sliding window; The estimated horizontal orientation angle of the face for the camera Angle of arrival of the signal measured by multiple antennas The azimuth difference; This represents the historical standard deviation of the azimuth difference; This is a sequence of face bounding box height change rates within one second, used to characterize the speed at which a face approaches or moves away. This is a sequence of RSSI intensity change rates within the same time window, used to characterize the speed at which the device approaches or moves away. To scale the RSSI rate of change to a coefficient with the same dimensions as the pixel change; The dynamic time warping distance between the two time series is used to measure the degree of motion synchronization. This is the normalized face embedding vector; This is the normalized device fingerprint vector; The cross-modal projection matrix, obtained through adversarial or maximum mean difference training, maps the device fingerprint to the face embedding space; It is the complement of the cosine similarity between face and device vectors; the smaller the value, the more similar they are. This represents the probability distribution of the appearance of this face over a past period of time. This represents the probability distribution of fingerprints appearing on this device over a past period of time. The Jensen-Shannon divergence measures the similarity between two occurrence patterns. This is the final five-dimensional difference vector, used for subsequent Mahalanobis distance and optimal transport global matching.

[0098] Step two involves performing an exponential moving average on the difference vectors obtained from continuous observations to estimate the covariance matrix online. Then, the Mahalanobis distance formula is used to map the difference vectors from the previous step to a single energy value. The inverse of the covariance matrix automatically adjusts the weights according to the actual dispersion of each dimension, automatically assigning a larger weight to rare but highly discriminative features and a smaller weight to common or noisy features, thus achieving fully data-driven feature fusion. The execution formula includes:

[0099]

[0100] In the formula, Let be the five-dimensional difference vector between the i-th face feature and the j-th device soft fingerprint, with elements being, in order, the normalized time difference, the normalized azimuth difference, the dynamic time warping distance, the cosine difference complement of the face-device vector, and the Jason-Shannon divergence of their historical occurrence distributions. The covariance matrix is ​​obtained by performing an exponential moving average on all difference vectors within a sliding time window. This is the transpose operation for a matrix or vector; This represents the overall energy value.

[0101] Step 3: Organize the energy values ​​between all faces and all devices into a cost matrix, and feed it into the entropy-regularized optimal transport model. The execution formula includes:

[0102]

[0103] In the formula, For the transfer matrix, elements This represents the probability of binding face i to device j; This is the cost matrix, where each element represents the combined energy value of the corresponding face and device; the smaller the value, the better the match. The inner product of two matrices, which is the sum of their element-wise multiplications, is used to characterize the total matching cost. The entropy regularity coefficient; for The negative entropy term; This is the quality relaxation (penalty) coefficient, used to control the degree of relaxation of row and column constraints; It is a column vector consisting entirely of 1s; It is a row and vector used to represent the total probability of each face being assigned to each device; The mass vector on the face side; It consists of columns and vectors, representing the total probability assigned to each face by each device; This is the mass vector on the equipment side; Given a norm of 1, the vectors are summed by taking the absolute value of the vectors.

[0104] Step four: For each face, take the maximum probability of matching with the device, and then calculate the 95th percentile of all maximum probabilities for the day as an adaptive threshold. If the maximum probability of a face is higher than the threshold, the binding is considered successful; otherwise, it is marked as low confidence and a risk alarm is triggered. The formula includes:

[0105]

[0106] In the formula, The optimal matching probability between face i and device j is obtained through the entropy regularized optimal transport algorithm; For a fixed face i, the maximum value among the matching probabilities of all devices j represents the probability of the device with the highest correlation to that face; The highest matching confidence score is for the i-th face. The larger the value, the more confident the system is about the set of highest matching confidence scores for that person. The ninety-fifth percentile of the above set.

[0107] In any of the above embodiments, the identity binding module further includes:

[0108] The risk marking unit, connected to the binding confirmation unit, is used to mark the corresponding person as a potential risk person and generate an alarm message when the binding confirmation unit determines that the binding confidence of the facial features and the device's soft fingerprint is lower than a preset threshold.

[0109] The binding database unit, connected to the binding confirmation unit, is used to store and manage bound records.

[0110] In this embodiment, when the face-device matching confidence S output by the binding confirmation unit is lower than the system's preset threshold Sth, the risk labeling unit immediately intervenes: on the one hand, it assigns a potential risk label to the personnel / device entries corresponding to the face feature vector IDf and the soft fingerprint vector IDd, and generates an alarm data packet containing an event timestamp, matching score, spatial coordinates, and a site thumbnail; on the other hand, it calls the upper-level alarm channel according to the risk level, and sends the push information to the resident's mobile app, home control screen, or property security platform in real time, while triggering the dynamic connection module to enter the second range for continuous high-frequency listening in order to capture more related signals. Subsequently, the binding database unit writes all binding records from the binding confirmation unit—whether successful or failed—into the local circular buffer and cloud log chain according to the "FaceID-DeviceID-Time" three-key index: entries with a confidence level ≥ Sth are marked as trusted bindings, while entries with a confidence level lower than Sth are synchronously stored as risk bindings, and the frequency of risk occurrence for the same FaceID or DeviceID is accumulated and statistically analyzed; the database implements AES-GCM encryption and hash chain verification on all records to ensure that the logs are tamper-proof, and provides APIs for subsequent behavior clustering, blacklist updates, and community big data analysis.

[0111] For example, a stranger wearing a baseball cap stops at the door. The camera captures the face (Fz), and the soft fingerprint is collected by the device (Vm). Because the face is obscured by the cap brim and the lighting is dim, the overall score is only 0.52, and the binding confirmation unit determines that the pairing confidence is insufficient. Based on this, the risk labeling unit generates an alarm for a potential risk person (#2023-07-01-1915) and immediately pushes it to the homeowner's mobile phone. Simultaneously, the door lock's Bluetooth power is increased to +4dBm, and the advertising interval is reduced to 100ms to continue monitoring. The binding database unit writes the record to the "Risk Binding" partition. Two days later, the same soft fingerprint (Vm) reappears at a neighbor's door with a still low matching score. The system detects that Vm's risk count has reached the threshold of 3 times, automatically upgrades its tag to high-risk device, and sends a linkage command to the property security backend, prompting a patrol to be arranged. Conversely, a week later, the courier's mobile phone Vk was first flagged as a risk, but subsequently it was successfully bound to his face multiple times. The database cleared the risk count of Vk to zero and upgraded it to a trusted visitor, avoiding repeated false alarms. This fully demonstrates the security management benefits of the collaborative updating and dynamic evolution of the risk flagging unit and the binding database unit.

[0112] The second aspect of the present invention provides a method, such as Figure 7 As shown, it includes the following steps:

[0113] S101: Collect facial images of people in front of the door lock, extract facial features and compare them with a whitelist; adjust the communication connection range to the first range or the second range based on the comparison results.

[0114] S102, within the adjusted communication connection range, scan the wireless communication signals sent by surrounding devices, extract multi-dimensional information and generate device soft fingerprints;

[0115] S103, Spatial matching of the device's soft fingerprint and the corresponding facial features within the same time window is performed to obtain the binding relationship between the facial features and the device's soft fingerprint, and the data is stored as a basis for personnel identification.

[0116] The present invention provides a method that first determines the communication connection range by comparing a face whitelist (S101), then generates a device soft fingerprint within the adaptive range, and completes face-device spatial matching and database entry within the same time window. This method achieves power saving, fast access, and high-risk detection within a closed loop: when a familiar person is detected, a small radius and low power consumption communication are maintained to ensure seamless unlocking and extend battery life; once a stranger appears, the communication range is instantly expanded, enabling the capture of their mobile phone / Wi-Fi / UWB signals from a distance and the construction of a soft fingerprint, which is then synchronously bound to the image to form a traceable dual-feature profile of the person and the device. This not only makes up for the deficiency of low-resolution cameras in independently marking strangers, but also avoids the battery waste caused by fixed high-power monitoring, thus improving the real-time identification accuracy, timely risk alarm, and reliability of post-event evidence collection for home security.

[0117] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program may include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium may include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in a computer-readable medium may be appropriately added to or subtracted according to the requirements of legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0118] The above embodiments are only used to illustrate the technical solutions of this disclosure, and are not intended to limit it. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be included within the protection scope of this disclosure.

Claims

1. A smart home security system based on the Internet of Things, characterized in that, include: The image recognition module is used to capture facial images of people in front of the door lock and extract facial features, compare the facial features with a pre-stored whitelist, and determine whether the person belongs to the whitelist. A dynamic connection module, connected to the image recognition module, is used to set the communication connection range to a first range when the image recognition module detects that only whitelisted personnel exist in front of the door lock, or to expand the communication connection range to a second range when the image recognition module detects that non-whitelisted personnel exist. The device signal acquisition module is connected to the dynamic connection module and is used to scan and acquire the wireless identification signals of surrounding devices within the communication connection range, extract multi-dimensional information of the devices, and construct a device soft fingerprint. The identity binding module is connected to the image recognition module and the device signal acquisition module respectively. It is used to perform spatiotemporal synchronous matching of the facial features and the device soft fingerprint, generate the binding relationship between the facial features and the device soft fingerprint, and store it for personnel identification.

2. The smart home security system according to claim 1, characterized in that, The image recognition module includes: The face detection unit is used to detect faces in the area in front of the door lock in real time when the device is awake, and to determine the location and number of faces. The feature extraction unit, connected to the face detection unit, is used to perform feature embedding on the detected face and extract the face feature vector; An identity comparison unit, connected to the feature extraction unit, is used to calculate the similarity between the facial feature vector and a pre-stored whitelist to determine whether the person corresponding to the face is a person on the whitelist.

3. The smart home security system according to claim 2, characterized in that, The communication connection range of the dynamic connection module also includes a third range. When a device appears in the third range, the face detection unit is activated and the face detection unit is put into a wake-up state.

4. The smart home security system according to claim 3, characterized in that, The third range is smaller than the first range, and when the face detection unit is in a wake-up state, the communication connection range changes from the third range to the first range or the second range.

5. The smart home security system according to claim 3, characterized in that, The dynamic connection module includes: A range decision unit is used to select either the first range or the second range based on the result determined by the image recognition module. A communication parameter adjustment unit, connected to the range decision unit, is used to set the power parameters and broadcast period of the communication connection in the first range, the second range, and the third range, respectively.

6. The smart home security system according to claim 1, characterized in that, The device signal acquisition module includes: The signal scanning unit is used to continuously scan for signals broadcast by nearby devices within the communication connection range and collect multi-dimensional information about the devices. A soft fingerprint generation unit, connected to the signal scanning unit, is used to construct and store the device soft fingerprint based on the multi-dimensional information including device identity information.

7. The smart home security system according to claim 6, characterized in that, The device signal acquisition module also includes: The identity injection interaction unit is used to actively broadcast identity handshake request signals and monitor whether surrounding devices respond. The identity authentication response unit, connected to the identity injection interaction unit, is used to receive binding identification information sent by the response device and send the binding identification information as the device identity information to the soft fingerprint generation unit.

8. The smart home security system according to claim 1, characterized in that, The identity binding module includes: A spatiotemporal synchronization unit is used to perform spatiotemporal matching of the facial features and the device soft fingerprint to determine the correlation between the facial features and the device soft fingerprint. The binding confirmation unit is connected to the spatiotemporal synchronization unit and is used to confirm the binding relationship based on the correlation determination result, and to form a binding record by binding information of facial features and device soft fingerprint.

9. The smart home security system according to claim 8, characterized in that, The identity binding module also includes: A risk marking unit, connected to the binding confirmation unit, is used to mark the corresponding person as a potential risk person and generate an alarm message when the binding confirmation unit determines that the binding confidence of the facial feature and the device soft fingerprint is lower than a preset threshold. A binding database unit is connected to the binding confirmation unit and is used to store and manage binding records.

10. A method implemented by the smart home security system according to any one of claims 1-9, characterized in that, Includes the following steps: Collect facial images of people in front of the door lock, extract facial features and compare them with a whitelist; adjust the communication connection range to the first range or the second range based on the comparison results; Within the adjusted communication connection range, scan the wireless communication signals sent by surrounding devices, extract the multi-dimensional information, and generate the device soft fingerprint; The device's soft fingerprint and the corresponding facial features are spatially matched within the same time window to obtain the binding relationship between the facial features and the device's soft fingerprint, and stored as a basis for personnel identification.

Images