Unmanned aerial vehicle intrusion detection system based on personnel intrusion alarm

Abstract

The invention relates to the technical field of unmanned aerial vehicle intrusion detection, in particular to an unmanned aerial vehicle intrusion detection system based on personnel intrusion alarm, which comprises a user authentication module, an area setting module, an unmanned aerial vehicle control module, a shooting control module, an intrusion trend judgment module, an intrusion identification module, a display terminal, a server and a cloud database. According to the invention, a user can flexibly set a monitoring area after authentication, generates an unmanned aerial vehicle control signal, dispatches an unmanned aerial vehicle in a target aircraft nest base station, accurately determines a dispatchable unmanned aerial vehicle, inputs task data information to the dispatchable unmanned aerial vehicle, photographs personnel in the monitoring area through the dispatchable unmanned aerial vehicle, generates an image set, and sends the image set to the target aircraft nest base station. And then an intrusion trend value is calculated according to the image sets to determine whether a face recognition instruction is triggered or not, and identity recognition judgment is performed according to the face recognition instruction to determine whether an alarm signal is generated or not, so that omnibearing and high-efficiency monitoring of a monitoring area is realized, and the accuracy and reliability of safety monitoring are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle intrusion detection, and in particular to a unmanned aerial vehicle intrusion detection system based on human intrusion alarm. Background Art

[0002] At present, most regional monitoring systems rely on fixed cameras for monitoring. Although this traditional security monitoring method can meet static security needs to a certain extent, its limitations are also very obvious. It is difficult to meet large-scale and efficient security requirements. In recent years, drone technology has provided a new idea for solving this problem due to its advantages such as strong mobility, wide coverage and real-time video transmission.

[0003] However, despite the significant advantages of drones in security monitoring, their application still faces the following shortcomings. First, the efficient dispatch of drones is still a problem. Traditional drone control methods usually rely on manually set flight paths and manual control operations, lacking an intelligent task allocation mechanism, resulting in insufficient optimization of drone resource utilization. Second, how to accurately identify intruders and conduct accurate identity authentication, especially in complex environments, still poses great technical difficulties. Existing monitoring systems are still difficult to achieve accuracy and efficiency in personnel identification, behavior analysis, and target positioning.

[0004] Therefore, although drone technology has brought new opportunities for security monitoring, how to efficiently dispatch drones and accurately identify intruders remains a key technical problem that needs to be solved in current drone security. Summary of the invention

[0005] The purpose of the present invention is to solve the limitations of traditional monitoring methods, such as the inability to intelligently and efficiently dispatch drones and the inability to accurately identify intruders, and to propose a drone intrusion detection system based on human intrusion alarms.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] The drone intrusion detection system based on human intrusion alarm includes:

[0008] User authentication module, used to authenticate the user's identity information;

[0009] The area setting module is used to set the monitoring area to generate the drone control signal;

[0010] The drone control module is used to receive drone control signals and perform control analysis on the drone in the target nest base station. The specific analysis is: retrieve the relevant monitoring parameter setting table of the monitoring area, and extract the monitoring period value, monitoring frequency value, monitoring duration value and continuous cycle value from it for comprehensive analysis to calculate the monitoring degree value of the monitoring area;

[0011] Matching the monitoring degree value with the monitoring degree determination table to obtain the monitoring degree level, matching the corresponding type of drone according to the monitoring degree level, extracting the key state parameters of each drone matching the drone type for comprehensive analysis to calculate the dispatch value, determining the dispatchable drone based on the dispatch value, and inputting the mission data information into the dispatchable drone, wherein the mission data information includes the monitoring flight path and the relevant monitoring parameter setting table;

[0012] A shooting control module is used to control the dispatchable drone to shoot the faces of people in the shooting area and people with intrusion tendency, and to generate and transmit a set of person images of each person and a set of face images of each person with intrusion tendency;

[0013] An intrusion trend determination module is used to calculate an intrusion trend value based on a set of images of each person, thereby determining and analyzing the intrusion trend of each person to determine whether to trigger a face recognition instruction;

[0014] The intrusion identification module is used to perform identity recognition and analysis based on the face image set of each intrusion trend person to determine whether to generate an alarm signal.

[0015] Furthermore, the specific analysis process of setting the monitoring area is as follows:

[0016] Acquire the address information entered by the user in the search box of the operation interface, thereby generating a corresponding search map, and generate a detailed map of the monitoring area based on the monitoring area preliminarily marked by the user on the search map. On the detailed map of the monitoring area, the user manually delineates each target monitoring area on the map by selecting a boundary box of different colors. The colors of the boundary box include red, yellow and blue;

[0017] After completing the manual demarcation, a dialog box will pop up on the operation interface, displaying the relevant monitoring parameter setting table for the monitoring area. The relevant monitoring parameter setting table includes the monitoring time period, monitoring frequency, monitoring duration and duration. The user needs to select specific monitoring parameters in the relevant monitoring parameter setting table and submit after confirmation. Once the specific monitoring parameter settings are completed and confirmed, the drone control signal will be automatically generated.

[0018] Furthermore, the specific analysis process for selecting the target machine nest base station is as follows:

[0019] Based on the captured drone control signal, the location information of the monitoring area and the location information of each machine nest base station around the monitoring area are retrieved, the distance value between the monitoring area and each machine nest base station around the monitoring area is extracted, and the machine nest base station corresponding to the minimum distance value is marked as the target machine nest base station.

[0020] Furthermore, the specific solution process for the dispatch value is as follows:

[0021] Extract the key state parameters of each drone that matches the drone type from the target nest base station. The key state parameters include the execution time difference value Gzxin r , operating efficiency value Gkanz r and connection stability value Glian r , where r represents the number of each drone matching the drone type in the target nest base station, and r = 1, 2, 3…m, and m represents the total number of drone numbers matching the drone type in the target nest base station. The three values ​​are extracted for normalization, according to the formula: Calculate the dispatch value PQZ of each drone r , where η1, η2 and η3 represent the set value coefficients respectively.

[0022] Furthermore, the specific analysis process of monitoring the flight path is as follows:

[0023] Obtain the location coordinates of the target nest base station and the monitoring area, and then calculate the initial flight path of the dispatchable drone from the target nest base station to the monitoring area through the shortest path algorithm;

[0024] Then, the position coordinates of each target monitoring area in the monitoring area are obtained, and the initial and final flight paths are constructed with the position coordinates between the target monitoring areas in the monitoring area, and the initial flight path and the initial and final flight paths are integrated to form a complete monitoring flight path;

[0025] The following steps are followed when constructing the final flight path:

[0026] The order of priority is determined by identifying the color of the bounding box corresponding to each target monitoring area. The target monitoring area with a red bounding box is regarded as a priority target, and the drone can be dispatched to monitor it in order of color.

[0027] The starting and ending flight paths are constructed based on the position coordinates of the target monitoring area corresponding to the red bounding box.

[0028] Furthermore, the specific process of acquiring the personnel image set of each person is as follows:

[0029] By obtaining the outer boundaries of each target monitoring area in the monitoring area, and expanding q units proportionally outward along the center point of the corresponding monitoring area according to the p-unit magnitude, the set outer boundaries of each target monitoring area in the monitoring area are obtained, and the area between the outer boundary and the set outer boundary is marked as the shooting area. The dispatchable drone is used to identify and shoot each person in the shooting area, thereby obtaining a set of personnel images of each person.

[0030] Furthermore, the specific solution process for the intrusion trend value is as follows:

[0031] Arrange the image sets of each person in the order of shooting time to obtain the image sequence of each person, mark each image as a detection point, and thus obtain the detection points corresponding to each person;

[0032] By extracting the row distance value RQ1 of each detection point corresponding to each person s yi ;

[0033] Wherein, s represents the number of the bounding box color, and s=1, 2, 3, y represents the number of each person, and y=1, 2, 3...n1, n1 represents the total number of each person number, i represents the number of each detection point corresponding to each person, and i=1, 2, 3...n2, n2 represents the total number of each detection point number corresponding to each person;

[0034] By extracting the line spacing threshold corresponding to each color boundary box as the reference line spacing value RQ1 s* ;

[0035] Compare and analyze the line spacing value of each detection point corresponding to each person with the reference line spacing value. When the line spacing value of a detection point corresponding to each person is less than or equal to the reference line spacing value, the detection point corresponding to each person is marked as an intrusion point. The total number of intrusion points marked for each person is counted, thereby obtaining the intrusion number value RY for each person. s y ;

[0036] According to the formula: The intrusion distance assessment value RQP1 is calculated, where x1 and x2 represent correction factors respectively;

[0037] According to the formula: Calculate the intrusion speed assessment value RQP2, where T (i，i-1) represents the time between the i-th detection point and the i-1-th detection point corresponding to each person, and x3 represents the correction factor;

[0038] By extracting the row angle value RQ2 of each detection point corresponding to each person s yi ;

[0039] According to the formula: The intrusion angle evaluation value RQP3 is calculated, where x4 represents the correction factor;

[0040] Extract the intrusion distance assessment value RQP1 of each person s y , Intrusion speed assessment value RQP2 s y And the intrusion angle evaluation value RQP3s y The value of is normalized according to the formula: Calculate the intrusion trend value Rqinsg of each person s y , where μ1, μ2 and μ3 represent the weight coefficients of the intrusion distance evaluation value, the intrusion speed evaluation value and the intrusion angle evaluation value, respectively, and μ1>μ3>μ2.

[0041] Furthermore, the specific analysis process for determining whether to trigger the face recognition instruction is as follows:

[0042] The intrusion trend value of each person is compared and analyzed with the preset intrusion trend threshold. When the intrusion trend value of a person is greater than or equal to the preset intrusion trend threshold, the person is determined to be an intrusion trend person, and an intrusion trend signal is generated. The number of intrusion trend signals generated is counted and compared and analyzed with the preset number threshold. When the number of intrusion trend signals is greater than the preset number threshold, the face recognition instruction is triggered.

[0043] Furthermore, the specific process of obtaining the face image set of each intrusion tendency person is as follows:

[0044] Based on the triggered face recognition command, the dispatchable drone is controlled to automatically adjust the flight altitude and angle so that the camera is aimed at the face of the intruder to take pictures, thereby obtaining a collection of face images of each intruder.

[0045] Furthermore, the specific analysis process for determining whether to generate an alarm signal is as follows:

[0046] Through face recognition technology, the face image collection of each intrusion tendency person is matched with the face image of the authorized person stored in the cloud database to confirm the specific identity of the intrusion tendency person and extract his identity authentication status;

[0047] If the identity authentication status of the intrusion trend person is not a passer-by in the target monitoring area, an alarm signal is generated.

[0048] Compared with the known prior art, the technical solution provided by the present invention has the following beneficial effects:

[0049] 1. The present invention allows the user to flexibly define the monitoring area and set relevant monitoring parameters (such as monitoring period, frequency, duration and cycle) in the operation interface, calculates the monitoring degree value of the monitoring area according to the relevant monitoring parameters, thereby dynamically matching the required drone type, extracting the key state parameters of each drone matching the drone type for comprehensive analysis to calculate the dispatch value, determining the dispatchable drone based on the dispatch value, and inputting the mission data information into the dispatchable drone, so that the dispatchable drone goes to the monitoring area to perform the monitoring task, thereby effectively improving the dispatch efficiency of the drone and ensuring a rapid response to abnormal situations in the monitoring area.

[0050] 2. The present invention generates a shooting area by expanding the outer boundary of the monitoring area, and identifies and shoots each person in the shooting area to obtain a set of person images of each person, and extracts the intrusion distance assessment value, intrusion speed assessment value and intrusion angle assessment value to calculate the intrusion trend value, so as to quickly screen out people with intrusion tendencies. When the number of people with suspected intrusion trends reaches a preset threshold, the face recognition instruction is triggered, and their identities are identified through face matching. If the person has no authority to enter the monitoring area, an alarm signal is generated to notify the management personnel in real time, thereby achieving accurate and timely response to intrusion incidents. This multi-level intrusion detection and identification mechanism effectively improves the security monitoring capability and provides important protection for regional security management. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 It is the overall module block diagram of the present invention;

[0053] Figure 2 A monitoring area setting diagram of the present invention;

[0054] Figure 3 This is a shooting area diagram of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] like Figure 1 As shown, the drone intrusion detection system based on human intrusion alarm includes: a drone control platform, a server is set in the platform, and the server is connected to a user authentication module, an area setting module, a drone control module, a shooting control module, an intrusion trend determination module, an intrusion identification module, a display terminal and a cloud database;

[0057] The user authentication module is used to authenticate the user's identity information. The specific authentication is: the user logs in to the system interface through the display terminal. The system interface provides a variety of login methods for users to choose, including user name + password, USB KEY + PIN code, biometric authentication, etc. If the user authentication is successful, it will jump to the user-specified operation interface. At the same time, the user's operation status information on the operation interface is collected, and the collected operation status information is transmitted to the area setting module through the server. If the user fails to authenticate multiple times in a row, the account will be automatically locked and allowed to log in again after the set time period;

[0058] It should be noted that the specific operation process of username + password is: the user enters the pre-registered username in the login box, enters the corresponding password in the password box, clicks the login button, and compares the username and password entered by the user with the information stored in the system; the specific operation process of USB KEY + PIN code is: the user inserts the USB KEY containing the private key and digital certificate into the USB port of the computer, enters the PIN code of the USB KEY on the login interface, and verifies the user's identity through the public key algorithm built into the USB KEY; the specific operation process of biometric authentication is: the user puts his finger on the fingerprint reader or puts his eyes on the iris reader, and compares the scanned biometric features with the features stored in the system;

[0059] The area setting module is used to monitor the user's operating status information on the operation interface, and thus set and analyze the monitoring area. The specific analysis process is as follows:

[0060] like Figure 2As shown, the address information entered by the user in the search box of the operation interface is obtained, thereby generating a corresponding search map, and based on the monitoring area preliminarily marked by the user on the search map, a detailed map of the monitoring area is generated. On the detailed map of the monitoring area, the user manually delineates each target monitoring area on the map by selecting boundary boxes of different colors. The colors of the boundary boxes include red, yellow and blue. Among them, these boundary boxes of different colors are not only used to distinguish different target monitoring areas, but also represent the importance level of each target monitoring area, and the importance level of the red boundary box>the importance level of the yellow boundary box>the importance level of the blue boundary box. After completing the manual demarcation, a dialog box will pop up on the operation interface to display the relevant monitoring parameter setting table of the monitoring area, wherein the relevant monitoring parameter setting table includes the monitoring time period, monitoring frequency, monitoring duration and duration period. The user needs to select specific monitoring parameters in the relevant monitoring parameter setting table and submit it after confirming that it is correct. Once the specific monitoring parameter setting is completed and confirmed, the drone control signal will be automatically generated, and the drone control signal will be transmitted to the drone control module through the server;

[0061] It should be noted that the monitoring time period refers to the specific start and end time interval of each monitoring task, which is set based on one day (24 hours). For example, if the monitoring time period is set to 8 am to 6 pm every day, then the monitoring task will be started and executed during this time period;

[0062] The monitoring frequency refers to the number of times the monitoring task is repeated within the set monitoring time period. For example, if the monitoring time period is from 8 am to 6 pm every day, and the monitoring frequency is once an hour, then within the 10 hours from 8 am to 6 pm, the monitoring task will be executed once an hour, for a total of 10 times;

[0063] Monitoring duration refers to the length of time each monitoring task lasts from the start to the end. For example, if the monitoring duration of each monitoring task is set to 30 minutes, then no matter when the monitoring task starts, it will continue for 30 minutes and then end.

[0064] The duration period refers to the total time span from the start to the end of the entire monitoring task. For example, if the duration period is set to one week, the monitoring task will be executed for seven consecutive days from the first day of execution, and then end. The duration period can also be set to a longer period such as one month or one year.

[0065] The drone control module is used to receive drone control signals, thereby performing control analysis on the drone in the target nest base station. The specific analysis process is as follows:

[0066] According to the captured drone control signal, the location information of the monitoring area and the location information of each machine nest base station around the monitoring area are retrieved, the distance value between the monitoring area and each machine nest base station around the monitoring area is extracted, and the machine nest base station corresponding to the minimum distance value is marked as the target machine nest base station;

[0067] At the same time, the relevant monitoring parameter setting table of the monitoring area is retrieved, and the values ​​of the monitoring time period, monitoring frequency, monitoring duration and continuous cycle are extracted from the relevant monitoring parameter setting table to obtain the monitoring time period value, monitoring frequency value, monitoring duration value and continuous cycle value, and they are marked as K respectively. 时间段 , K 频率 , K 时长 and K 周期 , according to the formula: Calculate the monitoring degree value Kspszg of the monitoring area, where e represents a natural constant, and a1, a2, a3 and a4 represent the set weight coefficients respectively;

[0068] It should be noted that if the monitoring period is from 8 am to 6 pm every day, the monitoring period is 10 hours, then the monitoring period value is 10; if the monitoring frequency is once an hour, then the monitoring frequency value is 1, if the monitoring frequency is three times an hour, then the monitoring frequency value is 3; if the monitoring duration is set to 30 minutes, then the monitoring duration value is 30; if the continuous period is one week, then the continuous period value is 7, if the continuous period is one month, then the continuous period value is 30 (or 31 days, depending on the specific month);

[0069] The monitoring degree value of the monitoring area is matched and analyzed with the monitoring degree determination table stored in the cloud database, thereby obtaining the monitoring degree level of the monitoring area, and each monitoring degree value of the monitoring area corresponds to a monitoring degree level, and it is matched with the drone type corresponding to the monitoring degree level, thereby obtaining the drone type;

[0070] It should be noted that the monitoring level of the monitoring area includes level one monitoring level, level two monitoring level and level three monitoring level, and level one monitoring level corresponds to type one drone, level two monitoring level corresponds to type two drones, and level three monitoring level corresponds to type three drones;

[0071] Type 1 drones represent drones with high endurance, high performance, and high storage capabilities; Type 2 drones represent drones with medium endurance, medium performance, and medium storage capabilities; Type 3 drones represent drones with average endurance, average performance, and average storage capabilities;

[0072] Extract each drone that matches the drone type from the target nest base station, and obtain the key state parameters of each drone to obtain the key state parameters of each drone. The key state parameters include execution time difference value, operation efficiency value and connection stability value, and mark them as Gzxin r , Gkanz r and Glian r , where r represents the number of each drone matching the drone type in the target nest base station, and r = 1, 2, 3…m, m represents the total number of drone numbers matching the drone type in the target nest base station, and extracts the execution time difference value Gzxin from the key state parameters of each drone r , operating efficiency value Gkanz r and connection stability value Glian r The value of is normalized according to the formula: Calculate the dispatch value PQZ of each drone r , where η1, η2 and η3 represent the set value coefficients respectively. The larger the dispatch value of each drone is, the greater the probability that the drone is dispatched to the monitoring area. The dispatch value of each drone is compared with the preset dispatch threshold. If the dispatch value of a drone is greater than the preset dispatch threshold, the drone in the target nest base station is determined to be a dispatchable drone.

[0073] It should be noted that the execution time difference refers to the time interval since the drone last executed a task, which can be obtained by calculating the difference between the current time and the last task completion time. The operation efficiency value refers to the efficiency performance of the drone in task execution, which is solved by the following formula: Among them, the average flight speed refers to the average flight speed of the drone under standard conditions, the remaining power refers to the current battery power of the drone, the mission time refers to the time taken to complete the last mission, and the connection stability value refers to the communication connection quality between the drone and the base station of the machine nest, which is solved by the following formula: connection stability value = signal strength-(communication delay+packet loss rate), among which, signal strength refers to the communication signal strength between the base station and the drone, communication delay refers to the response delay time between the drone and the base station, and packet loss rate refers to the proportion of data lost during the communication process;

[0074] Inputting the set mission data information into the dispatchable UAV, so that the dispatchable UAV goes to the monitoring area to perform the monitoring mission; wherein the mission data information includes the monitoring flight path and the relevant monitoring parameter setting table;

[0075] Obtain the location coordinates of the target nest base station and the monitoring area, and then use the shortest path algorithm (A-star) to calculate the initial flight path of the dispatchable drone from the target nest base station to the monitoring area, ensuring that the drone reaches the monitoring area by the shortest path, thereby saving time and energy consumption;

[0076] Then obtain the position coordinates of each target monitoring area in the monitoring area, and construct the initial and final flight paths with the position coordinates between the target monitoring areas in the monitoring area. Integrate the initial flight path and the final flight path to form a complete monitoring flight path. When constructing the final flight path, follow the following steps:

[0077] The order of priority is determined by identifying the color of the corresponding boundary box of each target monitoring area. The target monitoring area with a red boundary box is regarded as a priority target. The drone can be dispatched to monitor in the order of color (red-yellow-blue) to ensure that the high-priority area is covered first;

[0078] The starting and ending flight paths are constructed based on the position coordinates of the target monitoring area corresponding to the red boundary box. The starting and ending flight paths thus determined can effectively shorten the flight time and improve the monitoring efficiency.

[0079] The shooting control module is used to identify and shoot each person in the shooting area, thereby obtaining a set of images of each person. The specific shooting analysis process is as follows:

[0080] like Figure 3 As shown, by obtaining the outer boundaries of each target monitoring area in the monitoring area, and expanding the q-unit ratio outward in equal proportion along the center point of the corresponding monitoring area according to the p-unit magnitude, the set outer boundaries of each target monitoring area in the monitoring area are obtained, wherein the area between the outer boundary and the set outer boundary is marked as the shooting area, and the equipment carried by the dispatchable drone is used to identify and shoot each person in the shooting area, thereby obtaining a set of personnel images of each person, and the set of personnel images of each person is transmitted to the intrusion trend determination module through the server;

[0081] The intrusion trend determination module is based on the image set of each person, and thus determines and analyzes the intrusion trend of each person. The specific analysis process is as follows:

[0082] Arrange the image sets of each person in the order of shooting time to obtain the image sequence of each person, mark each image as a detection point, and thus obtain the detection points corresponding to each person;

[0083] By extracting the row distance value of each detection point corresponding to each person and marking it as RQ1 s yi , where the row spacing value refers to the distance between the person and the outer boundary;

[0084] It should be noted that s represents the number of the bounding box color, and s=1, 2, 3. When s=1, it represents a red bounding box, when s=2, it represents a yellow bounding box, and when s=3, it represents a blue bounding box. y represents the number of each person, and y=1, 2, 3...n1, n1 represents the total number of each person number, i represents the number of each detection point corresponding to each person, and i=1, 2, 3...n2, n2 represents the total number of each detection point number corresponding to each person;

[0085] The line spacing threshold corresponding to each color bounding box is extracted from the cloud database as the reference line spacing value and marked as RQ1 s* ;

[0086] Compare and analyze the line spacing value of each detection point corresponding to each person with the reference line spacing value. When the line spacing value of a detection point corresponding to each person is less than or equal to the reference line spacing value, the detection point corresponding to each person is marked as an intrusion point. The total number of intrusion points marked for each person is counted, thereby obtaining the intrusion number value of each person and marking it as RY. s y ;

[0087] According to the formula: The intrusion distance assessment value RQP1 is calculated, where x1 and x2 represent correction factors respectively;

[0088] According to the formula: Calculate the intrusion speed assessment value RQP2, where T (i，i-1) represents the time between the i-th detection point and the i-1-th detection point corresponding to each person, and x3 represents the correction factor;

[0089] By extracting the row angle value of each detection point corresponding to each person and marking it as RQ2 s yi , according to the formula: The intrusion angle evaluation value RQP3 is calculated, where x4 represents the correction factor;

[0090] The line angle value refers to the angle between the midline position of the person's torso and the outer boundary, which is recorded as the line angle value;

[0091] Extract the intrusion distance assessment value RQP1 of each person s y , Intrusion speed assessment value RQP2 s y And the intrusion angle evaluation value RQP3 s y The value of is normalized according to the formula: Calculate the intrusion trend value Rqinsg of each person sy , where μ1, μ2 and μ3 represent the weight coefficients of the intrusion distance assessment value, the intrusion speed assessment value and the intrusion angle assessment value, respectively, and μ1>μ3>μ2;

[0092] The intrusion trend value Rqinsg of each person s y Compared with the preset intrusion trend threshold Rasin s Comparative analysis, when a person's intrusion trend value Rqinsg s y Greater than or equal to the preset intrusion trend threshold Rasin s When a person is identified as an intrusion trend person, an intrusion trend signal is generated, the number of intrusion trend signals generated is counted, and the number is compared and analyzed with a preset number threshold. When the number of intrusion trend signals is greater than the preset number threshold, a face recognition instruction is triggered;

[0093] The triggered face recognition command is transmitted to the shooting control module through the server;

[0094] Based on the face recognition instruction, the shooting control module controls the dispatchable drone to automatically adjust the flight altitude and angle, so that the camera is aimed at the face of the intruder to shoot, thereby obtaining a set of face images of each intruder, where the face of the intruder includes images of the front, side and 45° angles;

[0095] The facial image collection of each intrusion tendency person is transmitted to the intrusion identification module through the server;

[0096] The intrusion identification module is based on the face image collection of each intrusion trend person, and thus performs identification and judgment analysis on each intrusion trend person. The specific analysis process is as follows:

[0097] Through face recognition technology, the face image collection of each intrusion tendency person is matched with the face image of the authorized person stored in the cloud database to confirm the specific identity of the intrusion tendency person and extract his identity authentication status;

[0098] If the identity authentication status of the intrusion trend person is a passer-by in the target monitoring area, no alarm signal is generated. If the identity authentication status of the intrusion trend person is not a passer-by in the target monitoring area, an alarm signal is generated and transmitted to the display terminal of the administrator through the server for display notification;

[0099] The cloud database is used to store the monitoring degree determination table, the line spacing thresholds corresponding to each color boundary box, and the stored facial images of authorized personnel.

[0100] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A drone intrusion detection system based on human intrusion alarm, comprising a user authentication module and an area setting module for setting a monitoring area and generating a drone control signal, characterized in that: Also includes: The drone control module is used to receive drone control signals and perform control analysis on the drone in the target nest base station. The specific analysis is: call the relevant monitoring parameter setting table of the monitoring area, and extract the monitoring period value, monitoring frequency value, monitoring duration value and continuous cycle value to calculate the monitoring degree value of the monitoring area; Matching the corresponding type of drone according to the monitoring degree value, extracting the key state parameters of each drone matching the drone type for comprehensive analysis to calculate the dispatch value, determining the dispatchable drone based on the dispatch value, and inputting the mission data information into the dispatchable drone, wherein the mission data information includes the monitoring flight path and the relevant monitoring parameter setting table; A shooting control module is used to control the dispatchable drone to shoot the faces of people in the shooting area and people with intrusion tendency, and to generate and transmit a set of person images of each person and a set of face images of each person with intrusion tendency; An intrusion trend determination module, used to calculate an intrusion trend value based on a set of personnel images of each person, so as to determine whether to trigger a face recognition instruction; The intrusion identification module is used to perform identity recognition and analysis based on the face image set of each intrusion trend person to determine whether to generate an alarm signal.

2. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific analysis process for setting the monitoring area is as follows: Acquire the address information entered by the user in the search box of the operation interface, thereby generating a corresponding search map, and generate a detailed map of the monitoring area based on the monitoring area preliminarily marked by the user on the search map. On the detailed map of the monitoring area, the user manually delineates each target monitoring area on the map by selecting a boundary box of different colors. The colors of the boundary box include red, yellow and blue; After completing the manual demarcation, the relevant monitoring parameter setting table of the monitoring area will be displayed. The relevant monitoring parameter setting table includes the monitoring time period, monitoring frequency, monitoring duration and duration period. The user needs to select specific monitoring parameters in the relevant monitoring parameter setting table and submit it after confirmation. Once the specific monitoring parameter settings are completed and confirmed, the drone control signal will be automatically generated.

3. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific analysis process for selecting the target machine nest base station is as follows: Based on the captured drone control signal, the location information of the monitoring area and the location information of each machine nest base station around the monitoring area are retrieved, the distance value between the monitoring area and each machine nest base station around the monitoring area is extracted, and the machine nest base station corresponding to the minimum distance value is marked as the target machine nest base station.

4. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific solution process for the dispatch value is as follows: Extract the key state parameters of each drone that matches the drone type from the target nest base station. The key state parameters include the execution time difference value Gzxin r , operating efficiency value Gkanz r and connection stability value Glian r , where r represents the number of each drone matching the drone type in the target nest base station, and r = 1, 2, 3…m, and m represents the total number of drone numbers matching the drone type in the target nest base station. The three values ​​are extracted for normalization, according to the formula: Calculate the dispatch value PQZ of each drone r , where η1, η2 and η3 represent the set value coefficients respectively.

5. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific analysis process of monitoring flight paths is as follows: Obtain the location coordinates of the target nest base station and the monitoring area, and then calculate the initial flight path of the dispatchable drone from the target nest base station to the monitoring area through the shortest path algorithm; Then, the position coordinates of each target monitoring area in the monitoring area are obtained, and the initial and final flight paths are constructed with the position coordinates between the target monitoring areas in the monitoring area, and the initial flight path and the initial and final flight paths are integrated to form a complete monitoring flight path; The following steps are followed when constructing the final flight path: The order of priority is determined by identifying the color of the bounding box corresponding to each target monitoring area. The target monitoring area with a red bounding box is regarded as a priority target, and the drone can be dispatched to monitor it in order of color. The starting and ending flight paths are constructed based on the position coordinates of the target monitoring area corresponding to the red bounding box.

6. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific process of acquiring the personnel image set of each person is as follows: By obtaining the outer boundaries of each target monitoring area in the monitoring area, and expanding q units proportionally outward along the center point of the corresponding monitoring area according to the p-unit magnitude, the set outer boundaries of each target monitoring area in the monitoring area are obtained, and the area between the outer boundary and the set outer boundary is marked as the shooting area. The dispatchable drone is used to identify and shoot each person in the shooting area, thereby obtaining a set of personnel images of each person.

7. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific solution process for the intrusion trend value is as follows: Arrange the image sets of each person in the order of shooting time to obtain the image sequence of each person, mark each image as a detection point, and thus obtain the detection points corresponding to each person; By extracting the row distance value RQ1 of each detection point corresponding to each person s yi ; Wherein, s represents the number of the bounding box color, and s=1, 2, 3, y represents the number of each person, and y=1, 2, 3...n1, n1 represents the total number of each person number, i represents the number of each detection point corresponding to each person, and i=1, 2, 3...n2, n2 represents the total number of each detection point number corresponding to each person; By extracting the line spacing threshold corresponding to each color boundary box as the reference line spacing value RQ1 s* ; Compare and analyze the line spacing value of each detection point corresponding to each person with the reference line spacing value. When the line spacing value of a detection point corresponding to each person is less than or equal to the reference line spacing value, the detection point corresponding to each person is marked as an intrusion point. The total number of intrusion points marked for each person is counted, thereby obtaining the intrusion number value RY for each person. s y ; According to the formula: The intrusion distance assessment value RQP1 is calculated, where x1 and x2 represent correction factors respectively; According to the formula: Calculate the intrusion speed assessment value RQP2, where T (i，i-1) represents the time between the i-th detection point and the i-1-th detection point corresponding to each person, and x3 represents the correction factor; By extracting the row angle value RQ2 of each detection point corresponding to each person s yi ; According to the formula: The intrusion angle evaluation value RQP3 is calculated, where x4 represents the correction factor; Extract the intrusion distance assessment value RQP1 of each person s y , Intrusion speed assessment value RQP2 s y And the intrusion angle evaluation value RQP3 s y The value of is normalized according to the formula: Calculate the intrusion trend value Rqinsg of each person s y , where μ1, μ2 and μ3 represent the weight coefficients of the intrusion distance evaluation value, the intrusion speed evaluation value and the intrusion angle evaluation value, respectively, and μ1>μ3>μ2.

8. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific analysis process for determining whether to trigger the face recognition command is as follows: The intrusion trend value of each person is compared and analyzed with the preset intrusion trend threshold. When the intrusion trend value of a person is greater than or equal to the preset intrusion trend threshold, the person is determined to be an intrusion trend person, and an intrusion trend signal is generated. The number of intrusion trend signals generated is counted and compared and analyzed with the preset number threshold. When the number of intrusion trend signals is greater than the preset number threshold, the face recognition instruction is triggered.

9. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 8 is characterized in that: The specific process of obtaining the face image collection of each intrusion trend person is as follows: Based on the triggered face recognition command, the dispatchable drone is controlled to automatically adjust the flight altitude and angle so that the camera is aimed at the face of the intruder to take pictures, thereby obtaining a collection of face images of each intruder.

10. The unmanned aerial vehicle intrusion detection system based on human intrusion alarm according to claim 1 is characterized in that: The specific analysis process for determining whether to generate an alarm signal is as follows: Through face recognition technology, the face image collection of each intrusion tendency person is matched with the face image of the authorized person stored in the cloud database to confirm the specific identity of the intrusion tendency person and extract his identity authentication status; If the identity authentication status of the intrusion trend person is not a passer-by in the target monitoring area, an alarm signal is generated.

Images