Human Behavior Recognition Device and Method for Ecological Reserves Based on Gait Feature Recognition

Abstract

A method and device for human behavior recognition in ecological protection zones based on gait feature recognition are proposed. The method involves laying a ring-shaped first optical fiber detection layer around the perimeter of the protected area's isolation zone; deploying a radial second optical fiber detection layer along the isolation zone towards the center of the protected area; and deploying a tri-band WiFi probe array within the isolation zone. An intrusion location recognition module uses vibration signals detected by the first optical fiber detection layer to calculate the intrusion location coordinates using a time difference positioning formula, and activates the monitoring channels of the corresponding radial second optical fiber detection layer based on the location coordinates. A gait feature fusion and extraction module extracts multi-scale energy features and generates comprehensive gait features. An identity discrimination module inputs the comprehensive gait features into a pre-trained classification model and determines the identity of the intruder based on the comprehensive gait features, thereby constructing an intelligent, dynamic, interconnected, and low-cost ecological protection technology system for ecological protection zones.

Description

Technical Field

[0001] This invention relates to the field of behavioral feature recognition technology in biometric identification, and more particularly to human feature recognition and identification, specifically a human behavior recognition device and method based on gait features. Background Technology

[0002] Establishing hazard isolation zones and key protection isolation zones within ecological reserves is a crucial measure for balancing human activities and nature conservation. Isolation zones effectively block direct human interference with core ecological areas through physical barriers, creating continuous and stable living or preservation spaces for endangered species and geological relics. Hazard isolation zones can effectively prevent accidents involving tourists entering the reserve; for example, setting up isolation zones in areas prone to rockfalls, cliffs, and landslides can improve the safety of reserve construction.

[0003] However, existing technological isolation systems still have significant shortcomings: traditional monitoring methods rely on physical barriers and manual inspections, making it difficult to track dynamic intrusion behavior in real time; isolation facilities lack sufficient intelligence and adaptive adjustment capabilities, failing to cope with complex environmental changes; most isolation measures rely solely on physical barriers and promotional slogans, which are ineffective in isolating children, and physical barriers such as fences significantly impact the ecological aesthetics. Furthermore, when isolation facilities are breached, it is impossible to promptly identify the individuals entering the isolation area—whether they are legitimate maintenance personnel or intruders.

[0004] To identify individuals, additional identification equipment is typically installed on top of existing isolation facilities. Most existing identification equipment uses methods such as footprint detection and facial recognition. Conventional footprint detection and facial recognition usually employ image recognition methods to analyze gait or facial features to determine an individual's identity, which further increases the construction cost of isolation zones.

[0005] To overcome these bottlenecks, it is necessary to build a technological system that enables intelligent sensing, dynamic response, full-domain linkage, and low cost, and promote the transformation of ecological protection from passive defense to proactive protection. Summary of the Invention

[0006] The purpose of this invention is to provide a human behavior recognition device and method for ecological protected areas based on gait feature recognition, so as to solve the technical problems in the prior art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A method for identifying human behavior in ecological protected areas based on gait feature recognition includes the following steps:

[0009] Step 1: Deployment of the dual-modal sensor:

[0010] (1) Lay a ring-shaped first fiber optic vibration sensor around the isolation zone of the protected area to form the first fiber optic detection layer;

[0011] (2) Deploy radial second fiber optic vibration sensors in the isolation zone toward the center of the isolation area. Each radial fiber extends 5-50 meters, with a spacing of 1-5 meters and an angle of 0°-15° between adjacent radial fibers, forming a second fiber optic detection layer.

[0012] (3) Deploy a tri-band WiFi probe array within the isolation zone, with a probe spacing of ≤100 meters, supporting the 802.11ax protocol and MAC address sniffing function;

[0013] Step 2: Intrusion Location Identification:

[0014] (1) When the first fiber optic detection layer detects a vibration signal, the coordinates of the intrusion location are calculated using the time difference positioning formula;

[0015] (2) Activate the monitoring channel of the radial second optical fiber detection layer in the corresponding direction according to the position coordinates;

[0016] Step 3: Gait feature fusion and extraction in the second fiber detection layer:

[0017] (1) Perform wavelet packet decomposition on the activated k radiating fiber signals to extract multi-scale energy features;

[0018] (2) Generate gait comprehensive features through multi-scale energy features;

[0019] Step 4: Intruder identification and alarm linkage:

[0020] (1) Input the gait comprehensive features into the pre-trained classification model and judge the identity of the intruder based on the gait comprehensive features;

[0021] (2) Trigger the minor alert mode when the person is identified as a minor (confidence level > 85%);

[0022] (3) When the person is judged to be an adult (confidence level > 85%), the adult alarm mode is triggered.

[0023] Preferably, the first and second fiber optic vibration sensors in step 1 are buried underground.

[0024] Preferably, the first fiber optic vibration sensor in step 1 is a closed loop or is positioned along the length of the isolation strip.

[0025] Preferably, in step 2, the second fiber optic vibration sensor that first senses the vibration signal of human footsteps is first determined based on the coordinates of the intrusion location. After determining the first activated monitoring channel, the adjacent detection channels are further activated based on the vibration waveform change of the first activated monitoring channel.

[0026] Preferably, the intruder's intrusion route is determined by analyzing the amplitude changes of the vibration waveforms of multiple optical fibers.

[0027] Preferably, the multi-scale energy characteristics in step 3 include the impact frequency when the foot lands, the periodic step frequency, the time-domain amplitude, the zero-crossing rate, the peak value, and the amplitude variation trend.

[0028] Preferably, the pre-trained classification model in step 4 includes a fitting function for the magnitude of foot impact at different ages and genders, and a gait frequency feature model for different ages and genders; the identity of the intruder includes age and gender.

[0029] Preferably, the minor alarm mode in step 4 specifically involves capturing MAC addresses near the isolation area using the Wi-Fi probe in step 1, comparing them with the ticketing system database using a matching algorithm, detecting whether the identified MAC addresses near the isolation area correspond to information about an adult accompanying a minor, and whether the age, gender, and other identity information match. If so, the phone number of the adult guardian is obtained, and an alarm SMS is sent through the operator interface to remind the accompanying minor not to enter the isolation zone of the protected area, or to remind the minor to be aware of minors within the isolation zone of the protected area.

[0030] Preferably, in step 4, the adult alarm mode is as follows: if the Wi-Fi probe directly detects the MAC address of someone entering the isolation area, it directly matches the MAC address with the tourist's mobile phone number in the ticketing system database and sends an alarm SMS through the operator interface.

[0031] A human behavior recognition device for ecological protection areas based on gait feature recognition is used to implement the aforementioned human behavior recognition method for ecological protection areas based on gait feature recognition.

[0032] The device includes a dual-modal sensing module, an intrusion location recognition module, a gait feature fusion and extraction module, an identity determination module, and a linkage alarm module.

[0033] The dual-mode sensing module includes a ring-shaped first fiber optic vibration sensor laid around the periphery of the isolation zone of the protected area to form a first fiber optic detection layer; a radial second fiber optic vibration sensor deployed in the isolation zone toward the center of the isolation area to form a second fiber optic detection layer; and a tri-band WiFi probe array deployed within the isolation zone.

[0034] The intrusion location identification module is used to detect vibration signals using the first optical fiber detection layer, calculate the intrusion location coordinates using the time difference positioning formula, and activate the monitoring channel of the radial second optical fiber detection layer in the corresponding direction according to the location coordinates.

[0035] The gait feature fusion and extraction module performs wavelet packet decomposition on the activated radiating fiber signal to extract multi-scale energy features, and generates gait comprehensive features through the multi-scale energy features;

[0036] The identity determination module is used to input gait comprehensive features into a pre-trained classification model and determine the identity of the intruder based on the gait comprehensive features;

[0037] The linkage alarm module is used to trigger the minor alarm mode when the person is identified as a minor, and the adult alarm mode when the person is identified as an adult.

[0038] The beneficial effects of this invention are:

[0039] 1. This invention achieves a technical system of intelligent perception, dynamic response, full-domain linkage, and low cost by setting up an optical fiber vibration sensor in conjunction with a Wi-Fi probe. It promotes the transformation of ecological protection from passive defense to active protection. In particular, the optical fiber vibration sensor accurately identifies the intrusion location, intrusion route, and intruder's identity (age, gender). Combined with the Wi-Fi probe, it achieves low-cost alarm and protection of the isolation zone for minors intruding into the protected area.

[0040] 2. By setting up a ring-shaped first fiber optic detection layer and a radial second fiber optic detection layer composed of fiber optic sensors, rapid and low-cost detection of intrusion locations and routes in the isolated areas of the protected zone is achieved. Attached Figure Description

[0041] Figure 1 This is a schematic diagram illustrating the application scenario of the present invention;

[0042] Figure 2 This is a flowchart illustrating an embodiment of an exemplary method;

[0043] Figure 3 This is a schematic diagram of an arrangement mode for deploying the first and second fiber detection layers.

[0044] Figure 4 This is a schematic diagram of another arrangement mode for the deployment of the first and second fiber detection layers;

[0045] Figure 5 This is a block diagram of an embodiment of the identification device;

[0046] In the diagram: Ecological Protection Zone 100, Outer Ring Fence 106, Entrance / Exit 107, Landslide / Rockfall Prone Area 101, Protected Geographical and Historical Sites 102, No-Entry Primeval Forest 103, No-Approach Reservoir 104, Isolation Zone 105, First Fiber Optic Vibration Sensor 201, Second Fiber Optic Vibration Sensor 202, Spacing SP, Route 1 P1, Route 2 P2, Dual-Modal Sensing Module 301, Intrusion Location Recognition Module 302, Gait Feature Fusion Extraction Module 303, Identity Determination Module 304, Linkage Alarm Module 305 Detailed Implementation

[0047] To enable those skilled in the art to better understand the technical solutions in the embodiments of this specification, the technical solutions in the embodiments of this specification will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art should fall within the scope of protection.

[0048] Please see Figure 1 This is a schematic diagram illustrating an application scenario of a human behavior recognition method based on gait feature recognition in an ecological reserve, provided as an exemplary embodiment of this specification.

[0049] like Figure 1 This is a schematic plan of the ecological protection zone 100. Like most closed protected areas, it has a closed outer ring fence 106 and several entrances and exits 107. In addition to the usual scenic spots within the ecological protection zone, there are usually areas that are prohibited from entering, such as landslide-prone and rockfall-prone areas 101, protected geographical and historical sites 102, prohibited virgin forests 103, and reservoirs that are prohibited from being approached 104. Due to the danger or the need for a high level of protection, these areas are usually surrounded by protective isolation zones 105, which prohibit tourists from entering the protected area.

[0050] Establishing hazard isolation zones and key protection isolation zones within ecological reserves is a crucial measure for balancing human activities and nature conservation. Isolation zones effectively block direct human interference with core ecological areas through physical barriers, creating continuous and stable living or preservation spaces for endangered species and geological relics. Hazard isolation zones can effectively prevent accidents involving tourists entering the reserve; for example, setting up isolation zones in areas prone to rockfalls, cliffs, and landslides can improve the safety of reserve construction.

[0051] However, existing technological isolation systems still have significant shortcomings: traditional monitoring methods rely on physical barriers and manual inspections, making it difficult to track dynamic intrusion behavior in real time; isolation facilities lack sufficient intelligence and adaptive adjustment capabilities, failing to cope with complex environmental changes; most isolation measures rely solely on physical barriers and promotional slogans, which are ineffective in isolating children, and physical barriers such as fences significantly impact the ecological aesthetics. Furthermore, when isolation facilities are breached, it is impossible to promptly identify the individuals entering the isolation area—whether they are legitimate maintenance personnel or intruders.

[0052] To identify individuals, additional identification equipment is typically installed on top of existing isolation facilities. Most existing identification equipment uses methods such as footprint detection and facial recognition. Conventional footprint detection and facial recognition usually employ image recognition methods to analyze gait or facial features to determine an individual's identity, which further increases the construction cost of isolation zones.

[0053] The embodiments in this specification are designed to overcome these bottlenecks in this scenario by building a technical system that enables intelligent perception, dynamic response, full-domain linkage, and low cost, thereby promoting the transformation of ecological protection from passive defense to proactive protection.

[0054] It should be noted that the embodiments in this specification aim to detect and identify personnel entering the isolation area of ​​the protected area at low cost through multimodal sensors. The human behavior recognition method for ecological protected areas based on gait feature recognition in the embodiments of this specification is applied in the above-mentioned scenario to assist the protected area managers in carrying out appropriate management of the protected area.

[0055] As follows, based on Figure 2 The example application scenario illustrates the following embodiments to provide a detailed description of this human behavior recognition method.

[0056] Please see Figure 2 This is a flowchart illustrating an embodiment of a human behavior recognition method for ecological protected areas based on gait feature recognition, provided as an exemplary embodiment of this specification. The method is described above... Figure 1 Based on the application scenario shown, the following steps are included:

[0057] Step 1: Deployment of the dual-modal sensor:

[0058] (1) A ring-shaped first fiber optic vibration sensor is laid around the isolation zone 105 of the protected area to form the first fiber optic detection layer;

[0059] (2) Deploy radial second fiber optic vibration sensors in the isolation zone 105 toward the center of the isolation area. Each radial fiber extends 5-50 meters, with a spacing of 1-5 meters and an angle of 0°-15° between adjacent radial fibers, forming a second fiber optic detection layer.

[0060] (3) Deploy a tri-band WiFi probe array within the isolation zone 105, with a probe spacing of ≤100 meters, supporting 802.11ax protocol and MAC address sniffing function;

[0061] Step 2: Intrusion Location Identification:

[0062] (1) When the first fiber optic detection layer detects a vibration signal, the coordinates of the intrusion location are calculated using the time difference positioning formula;

[0063] (2) Activate the monitoring channel of the radial second optical fiber detection layer in the corresponding direction according to the position coordinates;

[0064] Step 3: Gait feature fusion and extraction in the second fiber detection layer:

[0065] (1) Perform wavelet packet decomposition on the activated k radiating fiber signals to extract multi-scale energy features;

[0066] (2) Generate gait comprehensive features through multi-scale energy features;

[0067] Step 4: Intruder identification and alarm linkage:

[0068] (1) Input the gait comprehensive features into the pre-trained classification model and judge the identity of the intruder based on the gait comprehensive features;

[0069] (2) Trigger the minor alert mode when the person is identified as a minor (confidence level > 85%);

[0070] (3) When the person is judged to be an adult (confidence level > 85%), the adult alarm mode is triggered.

[0071] In step 1 above, a dual-mode vibration sensor was formed by setting up an optical fiber vibration sensor and a Wi-Fi probe. The optical fiber vibration sensor was chosen because it has good anti-electromagnetic interference capabilities, adopts an all-fiber passive design, making it suitable for deployment in outdoor locations lacking power supply. It also boasts high sensitivity and accuracy, capable of detecting vibrations at the micrometer level, suitable for early warning and detection of medium, small, and minute vibrations. Furthermore, a single optical fiber can simultaneously monitor vibration changes at any location within the sensing arm's range, offering wide coverage, strong environmental adaptability, high temperature and corrosion resistance, making it suitable for harsh environments such as humidity and nuclear radiation, and supporting monitoring at the tens of kilometers level. During operation, the optical fiber vibration sensor operates based on the principles of optical wave modulation and interference: when vibration acts on the optical fiber, the fiber deformation causes changes in the phase, wavelength, or polarization state of the light wave. The phase difference of the optical signal is detected using a Mach-Zehnder interferometer (MZI) or a Fabry-Perot interferometer (FPI), and after conversion into an electrical signal, parameters such as the vibration frequency and amplitude are analyzed. Therefore, by setting up fiber optic vibration sensors buried underground, the vibration signals generated on the ground by human walking can be detected, and then the gait characteristics of humans can be extracted by analyzing these vibration signals.

[0072] The Wi-Fi probe is based on the IEEE 802.11 protocol. It obtains the MAC address by listening to ProbeRequest frames actively sent by devices. When a user turns on Wi-Fi, the phone periodically broadcasts this frame to search for available vibration sensors. The probe captures this frame and parses the MAC address, signal strength, and other information contained within. Its advantage is that it eliminates the need for the user to connect to a Wi-Fi vibration sensor; detection is only required when the device has Wi-Fi enabled. Therefore, it is suitable for scenarios like counting people in ecological reserves where there are large fluctuations in pedestrian traffic and direct detection is not advisable. Furthermore, the Wi-Fi probe supports multiple terminals such as iOS, Android systems, tablets, and laptops, exhibiting high compatibility across all devices. Therefore, it can detect any device carried by people entering the reserve. The probe can also automatically detect and record the MAC address, signal strength, and other data of devices throughout the area, enabling real-time monitoring. This makes it highly adaptable to ecological reserves requiring real-time monitoring. When Wi-Fi probes are deployed inside the ecological protection zone, they only collect device MAC addresses and signal parameters, without directly obtaining user identity information (such as mobile phone numbers). The data is transmitted via wired connection and stored encrypted on the central server, which can protect the privacy of people entering the protection zone. Furthermore, by combining external databases, such as operator data and scenic area ticket reservation databases, it is possible to analyze key information such as the profiles of people entering the protection zone, their behavioral trajectories, and information on those passing through.

[0073] In step 1 above, the first fiber detection layer is ring-shaped, and the second fiber detection layer consists of several radial lines pointing towards the interior of the isolation zone. See details... Figure 3-4 The diagram shows the first and second fiber detection layers deployed near the isolation zone 105. This specification exemplarily illustrates two arrangement modes.

[0074] See Figure 3 The outermost dashed line represents the edge of the isolation zone 105. Inside the isolation zone 105, dashed lines exemplarily represent prohibited areas prone to landslides and rockfalls, 101. A ring-shaped first fiber optic vibration sensor 201 is laid around the perimeter of the isolation zone 105 to form a first fiber optic detection layer. Radial second fiber optic vibration sensors 202 are deployed on the isolation zone 105 toward the center of the isolation area 101 to form a second fiber optic detection layer. Figure 3 In the embodiment shown, the first fiber optic vibration sensor 201 is a closed ring, but it can also be deployed in other closed ring configurations, such as rectangles, ellipses, or irregular geometric closed shapes.

[0075] Figure 4 The diagram shows another variation of the fiber optic vibration sensor deployment. The isolation strip 105 is set along the length of the prohibited area 101. The first fiber optic vibration sensor 201 is also set along the length of the isolation strip 105. The second fiber optic vibration sensor 202 is arranged radially toward the isolation area 101 to form a second fiber optic detection layer.

[0076] Below, in conjunction with Figure 4 The deployment example illustrates the identification and extraction process in steps 2-3. Figure 4 The vibration waveforms, paths, and fiber optic deployments shown are all schematic illustrations.

[0077] like Figure 4 As shown, when the first fiber optic vibration sensor 201 of the first fiber optic detection layer detects the vibration signal W-201, the intrusion location coordinates X are calculated using the time difference positioning formula. The time difference positioning method, based on the time difference between the arrival times of vibration signals from the same vibration source at different sensors and the spatial arrangement of the sensors, establishes and solves equations using their geometric relationships to obtain the precise location of the vibration source. This method is a conventional method for determining vibration location based on sensor signals and belongs to the standard means of vibration signal acquisition and positioning; therefore, it will not be elaborated upon here.

[0078] After calculating the coordinates X of the intrusion location, as follows: Figure 4 The diagram shows nine second fiber optic vibration sensors 202, numbered L1-L9, within this range. Each second fiber optic vibration sensor 202 has the same spacing SP. Therefore, based on the intrusion location coordinate X, the second fiber optic vibration sensor 202 that first senses the vibration signal of a human footstep intrusion can be determined. Specifically, it can be determined by using the nearest fiber optic cable within one SP range to the left and right of coordinate X. Figure 4In the illustrated embodiment, the second fiber optic detection layer that first senses the intrusion vibration is the fiber optic cable numbered L6 and L7. Therefore, the monitoring channels of fiber optic cables L6 and L7 are activated first.

[0079] During the intrusion of human footsteps, the path has two forms: one is route one P1 along the arrangement direction of the second optical fiber vibration sensor 202, and the other is route two P2 at an angle to the arrangement direction of the second optical fiber vibration sensor 202. When it is route one P1, as... Figure 4 In the illustrated embodiment, the vibration waveforms W-L6 and W-L7 of optical fibers L6 and L7 are approximately the same. When an intruder enters from route one P1 to route two P2, due to the movement of the vibration source, such as... Figure 4 In the illustrated embodiment, the vibration waveform W-L7 of fiber L7 gradually attenuates, while the vibration waveform of fiber L6 gradually intensifies. Furthermore, as route two P2 crosses fiber L6 and enters the monitoring range of fiber L5, it excites the vibration waveform W-L5 of fiber L5. Therefore, in step two, after determining the first activated monitoring channel, based on the changes in the vibration waveform of the first activated monitoring channel, such as the attenuation or enhancement of the waveform amplitude, adjacent detection channels are further activated. This saves monitoring computation costs and increases the data range of gait detection, improving the accuracy of gait feature extraction. Moreover, by analyzing the amplitude changes of the vibration waveforms of multiple fibers, the intrusion route of the intruder can be determined. Figure 4 As shown, after the vibration waveform of fiber L7 attenuates, the vibration amplitude of the adjacent fiber L6 increases, and a vibration signal is also detected along with fiber L5. This indicates that after personnel enter the isolation area through fiber L7, they move in the upper left direction of the figure, which is conducive to quickly assisting security personnel in locating intruders at low cost.

[0080] In step 3, the gait feature fusion and extraction of the second fiber detection layer first involves wavelet packet decomposition of the activated k radiating fiber signals to extract multi-scale energy features. Figure 4 In the embodiment shown, wavelet packet decomposition is performed on the signals of optical fibers L6 / L7 / L5 to extract multi-scale energy features.

[0081] Wavelet packet decomposition of different fiber optic signals can progressively decompose the original signal into low-frequency and high-frequency subbands. Noise can be removed from the high-frequency subband, and the impact component of the gait signal can be extracted, representing the magnitude of the impact when the foot strikes the ground. Periodic gait frequency features are extracted from the low-frequency subband. Simultaneously, time-domain features such as amplitude, zero-crossing rate, and peak value are extracted to identify the intensity of the gait. Furthermore, based on the amplitude variation in the time domain, the path direction of the gait is determined by comparing the amplitude variation trends of different signals. The impact frequency, periodic gait frequency, time-domain amplitude, zero-crossing rate, peak value, and amplitude variation trend mentioned above constitute the multi-scale energy features extracted by the second fiber optic detection layer. Subsequently, a comprehensive gait feature is formed based on this multi-scale energy feature. The comprehensive gait feature includes the impact magnitude obtained from the impact frequency and time-domain amplitude, the gait frequency obtained from the periodic gait frequency, and the path information obtained from the amplitude variation trend. The impact magnitude obtained from the impact frequency and time-domain amplitude can be directly expressed as a weighted function corresponding to the specific impact frequency and time-domain amplitude. The weights and function form of this weighted function can be obtained by fitting controlled variable experiments. For example, controlling men and women aged 3-60 to run and walk through fiber optic vibration sensors with the same coverage depth, thereby obtaining the fitting function of foot impact magnitude for different ages and genders.

[0082] Step 4 involves inputting gait characteristics into a pre-trained classification model to determine the identity of the intruder. This pre-trained classification model includes the foot impact magnitude fitting function for different ages and genders obtained in Step 3, as well as the gait frequency characteristic model for different ages and genders, also obtained through controlled variable experiments. The gait frequency characteristic model for different ages and genders includes the gait frequency range during walking and running for people of different ages and genders. Furthermore, the foot impact magnitude fitting function and the gait frequency characteristic model for different ages and genders are used to identify individuals entering the isolation zone, and to roughly determine their age and gender.

[0083] When a minor is identified (confidence > 85%), a minor alert mode is triggered. Specifically, this mode uses WiFi probes to capture MAC addresses near the restricted area and compares them with the ticketing system database using a matching algorithm. Since most visitors to the protected area purchase tickets via mobile phone QR code scanning or through the ticketing system, the database stores the MAC address information of the corresponding mobile phone and the visitor's identity information, such as phone number. Tickets for minors entering the protected area are usually booked by their accompanying adults; therefore, the corresponding MAC address stores information about the minor being accompanied by the adult, such as age and gender. When a minor is identified and the minor alert mode is triggered, the system uses WiFi probes to detect if the identified MAC addresses near the restricted area correspond to information about an adult accompanying a minor, such as age and gender. If a match is found, the system obtains the adult guardian's phone number and sends an alert SMS via the carrier interface, reminding the accompanying minor not to enter the protected area's restricted zone, or alerting the system to be aware of minors already within the restricted zone.

[0084] If the Wi-Fi probe does not detect MAC address information or match the guardianship information of minors near the isolation zone, it will detect the communication device of the nearest security personnel and then notify the nearest security personnel to go to the alarm. At the same time, the route information and identity information of the intruder will be notified.

[0085] When an individual is identified as an adult (confidence > 85%), an adult alarm mode is triggered. Similarly, in adult alarm mode, if the Wi-Fi probe directly detects a MAC address entering the restricted area, it directly matches the MAC address with the tourist's mobile phone number in the ticketing system database and sends an alarm SMS via the carrier interface. If no MAC address is detected entering the restricted area, the nearest detected MAC address tourist or security personnel can be notified to alert the intruder and prevent them from advancing further.

[0086] Corresponding to the above method embodiments, this specification also provides a human behavior recognition device for ecological reserves based on gait feature recognition, see [link to documentation]. Figure 5 The diagram shown is an embodiment block diagram of a human behavior recognition device for an ecological reserve based on gait feature recognition, provided in an exemplary embodiment of this specification. The device may include: a dual-modal sensing module 301, an intrusion location recognition module 302, a gait feature fusion extraction module 303, an identity discrimination module 304, and a linkage alarm module 305.

[0087] The dual-mode sensing module 301 includes a ring-shaped first optical fiber vibration sensor laid around the periphery of the isolation zone 105 in the protected area, forming a first optical fiber detection layer; a radial second optical fiber vibration sensor deployed in the isolation zone 105 toward the center of the isolation area, with each radial optical fiber extending 5-50 meters, spaced 1-5 meters apart, and the angle between adjacent radial optical fibers being 0°-15°, forming a second optical fiber detection layer; and a tri-band WiFi probe array deployed within the isolation zone 105, with a probe spacing ≤100 meters, supporting the 802.11ax protocol and MAC address sniffing function.

[0088] The intrusion location identification module 302 is used to detect vibration signals using the first optical fiber detection layer, calculate the intrusion location coordinates using the time difference positioning formula, and activate the monitoring channel of the radial second optical fiber detection layer in the corresponding direction according to the location coordinates.

[0089] The gait feature fusion extraction module 303 performs wavelet packet decomposition on the activated k radiating fiber signals to extract multi-scale energy features; and generates gait comprehensive features through the multi-scale energy features.

[0090] The identity determination module 304 is used to input the gait comprehensive features into the pre-trained classification model and determine the identity of the intruder based on the gait comprehensive features.

[0091] The linkage alarm module 305 is used to trigger the minor alarm mode when the person is judged to be a minor (confidence level > 85%), and to trigger the adult alarm mode when the person is judged to be an adult (confidence level > 85%).

[0092] It is understandable that the dual-modal sensing module 301, the intrusion location recognition module 302, the gait feature fusion and extraction module 303, the identity determination module 304, and the linkage alarm module 305 are functionally independent modules, which can, as Figure 5 The shown can be configured simultaneously in the device, or they can be configured separately in the device, therefore Figure 5 The structures shown should not be construed as limiting the embodiments described in this specification.

[0093] Furthermore, the specific implementation process of the functions and roles of each module in the above-mentioned device is detailed in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here. This specification also provides a computer device, which includes at least a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aforementioned people counting method. This method includes at least: intrusion location identification; when the first fiber optic detection layer detects a vibration signal, calculating the intrusion location coordinates using a time difference positioning formula; activating the monitoring channel of the corresponding radial second fiber optic detection layer based on the location coordinates; gait feature fusion extraction of the second fiber optic detection layer: performing wavelet packet decomposition on the activated k radial fiber optic signals to extract multi-scale energy features; generating comprehensive gait features through the multi-scale energy features; intruder identification and linked alarm.

[0094] This specification provides a more specific computer device hardware structure, which may include: a processor, a memory, an input / output interface, a communication interface, and a bus. The processor, memory, input / output interface, and communication interface are interconnected internally via the bus.

[0095] The processor can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0096] The memory can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage devices, dynamic storage devices, etc. The memory can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory and called and executed by the processor.

[0097] Input / output interfaces are used to connect input / output modules to enable information input and output. Input / output modules can be configured as components within a device or connected externally to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0098] The communication interface is used to connect the communication module to enable communication and interaction between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0099] A bus is a pathway that transmits information between various components of a device, such as processors, memory, input / output interfaces, and communication interfaces.

[0100] It should be noted that although the above-described device only shows the processor, memory, input / output interface, communication interface, and bus, in actual implementation, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown.

[0101] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned people counting method. The method includes at least: intrusion location identification; when a vibration signal is detected by the first fiber optic detection layer, calculating the intrusion location coordinates using a time difference positioning formula; activating the monitoring channels of the corresponding radial second fiber optic detection layer based on the location coordinates; gait feature fusion and extraction of the second fiber optic detection layer: performing wavelet packet decomposition on the activated k radial fiber optic signals to extract multi-scale energy features; generating comprehensive gait features through the multi-scale energy features; intruder identification and linked alarm.

[0102] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0103] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that the embodiments of this specification can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of the embodiments of this specification, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this specification.

[0104] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which can take the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email sending and receiving device, game console, tablet computer, wearable device, or any combination of these devices.

[0105] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. When implementing the embodiments of this specification, the functions of each module can be implemented in one or more software and / or hardware. Alternatively, some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0106] The above description is merely a specific implementation of the embodiments of this specification. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles of the embodiments of this specification, and these improvements and modifications should also be considered within the protection scope of the embodiments of this specification.

Claims

1. A method for identifying human behavior in ecological protected areas based on gait feature recognition, characterized in that: Includes the following steps: Step 1: Deployment of the dual-modal sensor: (1) Lay a ring-shaped first fiber optic vibration sensor around the isolation zone of the protected area to form the first fiber optic detection layer; (2) Deploy radial second fiber optic vibration sensors in the isolation zone toward the center of the isolation area. Each radial fiber extends 5-50 meters, with a spacing of 1-5 meters and an angle of 0°-15° between adjacent radial fibers, forming a second fiber optic detection layer. (3) Deploy a tri-band WiFi probe array within the isolation zone, with a probe spacing of ≤100 meters, supporting the 802.11ax protocol and MAC address sniffing function; Step 2: Intrusion Location Identification: (1) When the first fiber optic detection layer detects a vibration signal, the coordinates of the intrusion location are calculated using the time difference positioning formula; (2) Activate the monitoring channel of the radial second fiber detection layer in the corresponding direction according to the position coordinates; Step 3: Gait feature fusion and extraction in the second fiber detection layer: (1) Perform wavelet packet decomposition on the activated k radiating fiber signals to extract multi-scale energy features; (2) Generate gait comprehensive features through multi-scale energy features; Step 4: Intruder identification and alarm linkage: (1) Input the gait comprehensive features into the pre-trained classification model and judge the identity of the intruder based on the gait comprehensive features; (2) When it is determined to be a minor, the minor alarm mode is triggered. The minor alarm mode is to capture the MAC address near the isolation area through the wifi probe in step 1, and compare it with the ticketing system database through the matching algorithm to detect whether the MAC address identified near the isolation area corresponds to the information of an adult carrying a minor, and whether the identity information such as age and gender matches. If so, the mobile phone number of the adult guardian is obtained, and an alarm SMS is sent through the operator interface to remind the accompanying minor not to enter the isolation zone of the protected area, or to remind the minor in the isolation zone of the protected area to pay attention. (3) When the person is determined to be an adult, the adult alarm mode is triggered. The adult alarm mode is that if the WiFi probe directly detects the MAC address of the person entering the isolation area, the MAC address is directly matched with the tourist's mobile phone number in the ticketing system database, and an alarm SMS is sent through the operator interface.

2. The method for human behavior recognition in ecological protected areas based on gait feature recognition as described in claim 1, characterized in that: The first and second fiber optic vibration sensors in step 1 are buried underground.

3. The method for human behavior recognition in ecological protected areas based on gait feature recognition as described in claim 1, characterized in that: The first fiber optic vibration sensor in step 1 is a closed loop type, or it is set along the length of the isolation strip.

4. The method for human behavior recognition in ecological protected areas based on gait feature recognition as described in claim 1, characterized in that: In step 2, the second fiber optic vibration sensor that first senses the vibration signal of human footsteps is first determined based on the coordinates of the intrusion location. After determining the first activated monitoring channel, the adjacent detection channels are further activated based on the vibration waveform change of the first activated monitoring channel.

5. The method for human behavior recognition in ecological protected areas based on gait feature recognition as described in claim 4, characterized in that: By analyzing the amplitude changes of vibration waveforms in multiple optical fibers, the intruder's intrusion route can be determined.

6. The method for human behavior recognition in ecological protected areas based on gait feature recognition as described in claim 1, characterized in that: The multi-scale energy characteristics in step 3 include the impact frequency when the foot lands, the periodic step frequency, the time-domain amplitude, the zero-crossing rate, the peak value, and the amplitude variation trend.

7. The method for human behavior recognition in ecological protected areas based on gait feature recognition as described in claim 1, characterized in that: The pre-trained classification model in step 4 includes fitting functions for foot impact magnitude at different ages and genders, as well as step frequency feature models at different ages and genders; the identity of the intruder includes age and gender.

8. A human behavior recognition device for ecological reserves based on gait feature recognition, used to implement the method as described in any one of claims 1-7, characterized in that: The device includes a dual-modal sensing module, an intrusion location recognition module, a gait feature fusion and extraction module, an identity determination module, and a linkage alarm module. The dual-mode sensing module includes a ring-shaped first fiber optic vibration sensor laid around the periphery of the isolation zone of the protected area to form a first fiber optic detection layer; a radial second fiber optic vibration sensor deployed in the isolation zone toward the center of the isolation area to form a second fiber optic detection layer; and a tri-band WiFi probe array deployed within the isolation zone. The intrusion location identification module is used to detect vibration signals using the first optical fiber detection layer, calculate the intrusion location coordinates using the time difference positioning formula, and activate the monitoring channel of the radial second optical fiber detection layer in the corresponding direction according to the location coordinates. The gait feature fusion and extraction module performs wavelet packet decomposition on the activated radiating fiber signal to extract multi-scale energy features, and generates gait comprehensive features through the multi-scale energy features; The identity determination module is used to input gait comprehensive features into a pre-trained classification model and determine the identity of the intruder based on the gait comprehensive features; The linkage alarm module is used to trigger the minor alarm mode when the person is identified as a minor, and the adult alarm mode when the person is identified as an adult.

Images