Intelligent security and protection system for airport boundaries

Abstract

The invention discloses an intelligent security and protection system for an airport boundary, and the system comprises a front-end sensing system, a data transmission system and a central control system which are sequentially connected through an electric signal and a communication signal. Any partition front-end sensing system is provided with a nine-axis sensor capable of collecting dynamic signals and static signals of the surrounding environment of the fence in real time, an edge calculation all-in-one machine connected with the nine-axis sensor, and a switch connected with the edge calculation all-in-one machine; the edge calculation all-in-one machine receives environment data sensed by the nine-axis sensor and attitude data generated by attitude changes of the nine-axis sensor, and the edge calculation all-in-one machine processes the signal data to extract features. The nine-axis sensor selected by the invention performs machine learning and modeling based on a time-frequency characteristic time sequence spectrogram, combines information of a wind speed and rainfall sensor, optimizes weather interference except extreme meteorological conditions by processing data information, and is linked with a video to reduce a false alarm rate.

Description

Technical Field

[0001] The present invention relates to the field of security technology, in particular to an airport perimeter intelligent security system. Background Art

[0002] Regarding the airport perimeter, the surrounding environment is relatively complex, with villages, roads, grass, and trees. There is a strong interference to the operation of the perimeter security system, and there are security risks. Once an intrusion occurs, it is likely to cause huge losses. Equipped with a perimeter security prevention system, it can effectively alarm for intrusions.

[0003] The current conventional vibration optical fiber / cable intrusion alarm system uses a single detection of perimeter vibration events to alarm intrusion behaviors, and cannot accurately filter out interference behaviors such as non-intrusive vibrations, cars, trees, and small animals. Video intelligent analysis-based intrusion alarms are often not unaffected in harsh environments such as no light at night, strong light, and backlight. The system needs to reduce false alarms and missed alarms caused by the above situations.

[0004] Traditional intrusion alarm systems have a short effective detection distance and insufficient coverage.

[0005] The requirements for intrusion alarm systems are not only to serve as warnings and protection, but also to review and collect evidence. The traditional intrusion alarm system has a weak linkage mechanism and cannot realize image review when an alarm is triggered, nor can it record the alarm in real time for storage and evidence collection.

[0006] In view of this, it is necessary to improve the traditional intrusion alarm system. Summary of the invention

[0007] In view of the deficiencies in the prior art, the technical problem to be solved by the present invention is to provide an airport perimeter intelligent security system, the purpose of designing the airport perimeter intelligent security system is to prevent false alarms and missed alarms.

[0008] To solve the above technical problems, the present invention is implemented through the following scheme: an airport fence intelligent security system of the present invention includes a front-end perception system, a data transmission system and a central control system connected in sequence through electrical signals and communication signals, the front-end perception system is arranged in a partitioned network, and any partitioned front-end perception system is provided with a nine-axis sensor capable of collecting dynamic signals and static signals of the surrounding environment of the fence in real time, an edge computing all-in-one connected to the nine-axis sensor, and a switch connected to the edge computing all-in-one, the edge computing all-in-one receives environmental data perceived by the nine-axis sensor and posture data generated by the posture change of the nine-axis sensor itself, and the edge computing all-in-one processes the signal data to extract features;

[0009] The edge computing integrated machine identifies the intrusion mode and the non-intrusion mode based on the extracted feature data. The edge computing integrated machine uses the cloud big data model and combines the big data model of the central control system to conduct in-depth analysis and prediction of the feature data, classify the feature data, identify the intrusion behavior category, and make a decision on whether to issue an alarm;

[0010] The airport perimeter intelligent security system further includes an artificial intelligence learning module, which includes an artificial intelligence learning method, and the artificial intelligence learning method includes the following steps:

[0011] Step 1: The nine-axis sensor senses and measures the mechanical vibration within a certain range around itself, and sends the sensed and measured feature data to the edge computing all-in-one machine;

[0012] Step 2: The edge computing integrated machine analyzes the feature data sent back by each nine-axis sensor through a big data model;

[0013] Step 3: Compare the feature data sent back by adjacent or all nine-axis sensors through an artificial intelligence learning algorithm, and then combine it with the current weather information of the airport area to identify whether there is any abnormal feature data;

[0014] Step 4: Accurately determine the intrusion behavior category based on the judged feature data.

[0015] Furthermore, the feature extraction method adopts Fourier transform $$F(\omega)=\sum_{-\infty}^{\infty}f(t)e^{-j\omega t}dt$$ and wavelet transform method to extract key features.

[0016] Furthermore, the edge computing integrated machine (12) classifies feature data using an XGBoost algorithm.

[0017] Furthermore, the classification formula of the XGBoost algorithm is expressed as:

[0018] $$F_m(x)=F_{m-1}(x)+\lambda\cdot h(x;\gamma_m,\nu_m)$$;

[0019] in:

[0020] $$F_m(x)$$ is the prediction result of the mth tree for sample x;

[0021] $$F_{m-1}(x$$ is the prediction result of the first m-1 trees for sample x;

[0022] $$\lambd$$ is a regularization parameter used to control model complexity;

[0023] $$h(x;\gamma_m,\nu_m$$ is the prediction function of the mth tree, $$\gamma_m and \nu_m$$

[0024] are the parameters of the tree respectively;

[0025] The optimization goal of the XGBoost algorithm is to minimize the following objective function:

[0026] $$Obj(F)=\sum_{i=1}^{N}l(y_i,F(x_i))+\sum_{m=1}^{M}[\Omega(f_m)$$

[0027] in:

[0028] $$$$ is the sample size;

[0029] $$l(y_i,F(x_i)$$ is the loss function, which measures the error between the model prediction and the actual value;

[0030] $$$$ is the number of trees;

[0031] · $$\Omega (f_m$$ is a regularization term, including the complexity of the tree). Further, the nine-axis sensor is a vibration sensor integrating a three-axis gyroscope, a three-axis accelerometer, and a three-axis magnetometer.

[0032] Furthermore, the edge computing all-in-one machine is a distributed cloud processing computer having an open third-party interface, which can at least be connected to one or more of an alarm system, a lighting system, a warning light and sirens system, an Internet platform, and a geographic information system.

[0033] Furthermore, the characteristic data at least includes the vibration amplitude and frequency of the nine-axis sensor and the tilt angle of the nine-axis sensor itself.

[0034] Furthermore, the airport perimeter intelligent security system also includes a third-party linkage system networked with the front-end sensing system, and the third-party linkage system at least includes a video monitoring system, an audible and visual alarm system, a night auxiliary lighting system, a voice broadcast system, and a wind speed and rainfall monitoring system;

[0035] The video surveillance system is connected to the switch, and when the perimeter is invaded and an alarm is triggered, the generated alarm signal drives the video surveillance system to remotely review;

[0036] The sound and light alarm system includes a control room end sound and light alarm system and an on-site sound and light alarm system, wherein the control room end sound and light alarm system is driven by a sensor processor capable of generating a vibration signal in the front-end sensing system, and the on-site sound and light alarm system and the night auxiliary lighting system are driven by an I / O module in the front-end sensing system;

[0037] The voice broadcast system is provided with IP audio terminals, loudspeakers, sound and light alarms distributed in the perimeter intrusion identification area and connected to the switch, and a control center broadcasting device located in the control room and connected to the switch;

[0038] The wind speed and rainfall monitoring system is a meteorological unit installed in the enclosure intrusion identification area for real-time monitoring of the meteorological conditions in the enclosure intrusion identification area. The meteorological unit will add the obtained real data to the sound and light alarm system algorithm in a timely and reliable manner, and eliminate the interference vibration caused by severe weather conditions through analysis and processing.

[0039] Furthermore, the nine-axis sensors are connected and networked by using aviation connectors, and the nine-axis sensors are fixed on the fence by using a buckle mesh structure.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] 1. The alarm accuracy of the present invention is high. The intrusion alarm system of the present invention needs to ensure that when an intrusion occurs, the system can automatically alarm with a high alarm accuracy. A variety of detection technologies are used to review the alarm environment. When an alarm occurs, the video surveillance image is linked for review, an expulsion sound alarm is set at the front end of the fence, and a sound alarm is set at the system monitoring terminal to prompt the monitoring personnel to handle it.

[0042] 2. The nine-axis sensor selected in the present invention performs machine learning and modeling based on the time-frequency feature time series spectrogram, and combines the wind speed and rainfall sensor information to optimize the weather interference except for extreme meteorological conditions through the processing of data information, and links the video to reduce the false alarm rate.

[0043] 3. The technology combining vibration sensing and high-definition video surveillance cameras is adopted to realize the two-way complementarity of physical detection and video detection. The intelligent algorithm based on deep learning can perform secondary recognition and tracking of human targets for the front-end perimeter alarm behavior, which can effectively reduce false alarms and improve the efficiency of police handling. At the same time, the airport perimeter intelligent security system of the present invention can link night auxiliary lights, warning lights / sirens, and automatic voice broadcasting equipment to achieve timely, strict, three-dimensional, and seamless perimeter protection.

[0044] 4. The sensor module of the present invention adopts a nine-axis sensor with a 3-axis gyroscope, a 3-axis accelerometer, and a 3-axis magnetometer. The nine-axis sensor collects weak mechanical vibration signals in multiple dimensions and perceives tiny posture changes more accurately. It supports but is not limited to the detection of mechanical physical changes such as vibration, swing, and roll. It can monitor the vibration, sway, and tilt posture of sensor carriers such as fences around the clock, and supports vibration intrusion vibration alarm, disconnection alarm, anti-dismantling alarm, and communication failure alarm. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is the architecture diagram of the airport perimeter intelligent security system of the present invention.

[0046] Figure 2 The block diagram of the video monitoring system of the present invention is networked with a front-end sensing system.

[0047] Figure 3 This is the alarm linkage flow chart of the present invention.

[0048] Marked in the attached figure: front-end perception system 1, central control system 2, IP audio terminal 3, large-screen display 4, alarm keyboard 5, execution terminal 6, speaker equipment 7, camera 8, broadcasting equipment 9, wind and rain sensor 10, switch 11, edge computing all-in-one machine 12, lighting equipment 13, nine-axis sensor 14, application server 21, data server 22. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more explicit definition of the protection scope of the present invention. Obviously, the embodiments described in the present invention are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0050] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0051] The general concept of the present invention is as follows: Figure 1 As shown, Figure 1The diagram is the architecture diagram of the airport fence intelligent security system of the present invention. An airport fence intelligent security system of the present invention includes a front-end perception system 1, a data transmission system and a central control system 2 connected in sequence through electrical signals and communication signals. The front-end perception system 1 is arranged in a partitioned network. Any partitioned front-end perception system 1 is provided with a nine-axis sensor 14 capable of collecting dynamic signals and static signals of the surrounding environment of the fence in real time, an edge computing all-in-one machine 12 connected to the nine-axis sensor 14, and a switch 11 connected to the edge computing all-in-one machine 12. The switch 11 is also connected to a lighting device 13, a wind and rain sensor 10, a broadcasting device 9 and a gun-shaped camera. The switch 11 also exchanges data with an application server 21 and a data server 22. The application server 21 and the data server 22 are connected to an IP audio terminal 3 and a server terminal, and the server terminal is connected to a large-screen display 4. The application server 21 and the data server 22 are also connected to an alarm keyboard 5 and an execution terminal 6, a speaker device 7, and a camera 8.

[0052] The application server 21 and the data server 22 are connected to multiple groups of front-end sensing systems 1.

[0053] The front-end sensing system 1 includes the front-end subsystem, and the front-end subsystem is composed of the front-end sensing equipment such as the intrusion alarm system and the video monitoring system. It includes vibration sensors, intelligent detection cameras, sound and light alarms, lighting, wind speed and rainfall equipment, etc.

[0054] The data transmission system 2 includes a transmission subsystem, which is composed of data exchange equipment such as optical cables, ring network switches and aggregation switches.

[0055] The central control system includes data server, application server, IP broadcasting system, perimeter monitoring and alarm integrated management platform, cloud storage system, decoding and splicing control system, and large-screen display control system.

[0056] Embodiment 1:

[0057] The edge computing integrated machine 12 receives the environmental data sensed by the nine-axis sensor 14 and the posture data generated by the posture change of the nine-axis sensor 14 itself, and processes the signal data to extract features through the edge computing integrated machine 12 .

[0058] The front-end sensing system 1 includes a perimeter intrusion alarm system and a video surveillance system; the transmission system includes optical cables, ring network switches and other data exchange equipment; the central control system includes a cloud storage system, a decoding and splicing control system, a large-screen display control system, and an airport comprehensive security platform.

[0059] The edge computing all-in-one machine 12 identifies the intrusion mode and the non-intrusion mode based on the extracted feature data. The edge computing all-in-one machine 12 conducts in-depth analysis and prediction of the feature data through the cloud big data model and in combination with the big data model of the central control system, classifies the feature data, identifies the behavior category of the intrusion, and makes a decision on whether to issue an alarm; the feature data at least includes the vibration amplitude and frequency of the nine-axis sensor 14 and the tilt angle of the nine-axis sensor 14 itself.

[0060] The method for extracting features adopts the Fourier transform $$F(\omega)=\sum_{-\infty}^{\infty}f(t)e^{-j\omega t}dt$$ and wavelet transform method to extract key features. This scheme uses nine-axis micro-electromechanical vibration sensing technology. The vibration detector adopts dense partition distributed vibration detection. For various intrusion methods such as climbing and damaging fences (walls), a wall-mounted field vibration detector is installed on the fence. Through the collection of vibration signals by the signal processor, the vibration generated by climbing, pulling, stepping on and cutting the fence can be effectively detected, and alarm signals such as shearing, climbing, tilting and disassembly can be output. The alarm signal analysis is completed by the field controller, and various real intrusion behaviors are judged and identified through the software analysis algorithm, and output to the system to complete the alarm event linkage. Multiple alarm secondary defense zones can be set, and different alarm linkage strategies can be set for different detection ranges. The location of the target can be obtained through the alarm defense zone, the false alarm rate can be reduced, and the intrusion partition can be accurately located. The alarm positioning accuracy can reach a minimum of 3 meters. The use of industrial and military-grade chips and passive components enables the sensor to work stably in the ambient temperature range of -40℃ to +85℃. The full-chip SMTPCB production process, the surface is treated with moisture-proof, salt spray-proof, and mildew-proof, the aluminum alloy shell, and the IP68 protection level improve the reliability of the system.

[0061] The method for classifying feature data by the edge computing integrated machine 12 is XGBoost algorithm classification. The formula of XGBoost algorithm classification is expressed as:

[0062] $$F_m(x)=F_{m-1}(x)+\lambda\cdot h(x;\gamma_m,\nu_m)$$;

[0063] in:

[0064] $$F_m(x)$$ is the prediction result of the mth tree for sample x;

[0065] $$F_{m-1}(x$$ is the prediction result of the first m-1 trees for sample x;

[0066] $$\lambd$$ is a regularization parameter used to control model complexity;

[0067] $$h(x;\gamma_m,\nu_m$$ is the prediction function of the mth tree, $$\gamma_m and \nu_m$$

[0068] are the parameters of the tree respectively;

[0069] The optimization goal of the XGBoost algorithm is to minimize the following objective function:

[0070] $$Obj(F)=\sum_{i=1}^{N}l(y_i,F(x_i))+\sum_{m=1}^{M}[\Omega(f_m)$$

[0071] in:

[0072] $$$$ is the sample size;

[0073] $$l(y_i,F(x_i)$$ is the loss function, which measures the error between the model prediction and the actual value;

[0074] $$$$ is the number of trees;

[0075] $$\Omega(f_m$$ is the regularization term, including the complexity of the tree).

[0076] The technology combining vibration sensing and high-definition video surveillance cameras is used to achieve two-way complementarity between physical detection and video detection. The intelligent algorithm based on deep learning can perform secondary recognition and tracking of human targets for the front-end perimeter alarm (cross-border detection, regional intrusion) behavior, which can effectively reduce false alarms and improve the efficiency of police handling. At the same time, the system can link night auxiliary lights, police lights / sirens, automatic voice broadcasts and other equipment to achieve timely, strict, three-dimensional and seamless perimeter protection.

[0077] Embodiment 2:

[0078] The airport perimeter intelligent security system further includes an artificial intelligence learning module, which includes an artificial intelligence learning method, and the artificial intelligence learning method includes the following steps:

[0079] Step 1: The nine-axis sensor 14 senses and measures the mechanical vibration within a certain range around itself, and sends the sensed and measured feature data to the edge computing integrated machine 12;

[0080] Step 2: The edge computing integrated machine 12 analyzes the feature data sent back by each nine-axis sensor 14 through a big data model;

[0081] Step 3: Compare the feature data sent back by adjacent or all nine-axis sensors 14 through an artificial intelligence learning algorithm, and then combine it with the current weather information of the airport area to identify whether there is abnormal feature data;

[0082] Step 4: Accurately determine the intrusion behavior category based on the judged feature data.

[0083] When vibration sensors are used in fence alarm systems as mechanical vibration detection devices, several vibration sensors are connected through communication cables to detect mechanical vibration information of a certain length of fence. Each vibration sensor can sense and measure mechanical vibrations within a certain range around itself, and send characteristic information such as vibration amplitude and frequency to the vibration intrusion controller through communication cables for information processing. The vibration intrusion controller analyzes the mechanical vibration characteristics sent back by each vibration sensor through a big data model, and compares the mechanical vibration characteristics sent back by adjacent or all vibration sensors through an artificial intelligence learning algorithm, and combines the meteorological information at that moment to identify whether there is any abnormality in mechanical vibration. This improves the reliability and accuracy of information processing and greatly reduces the false alarm rate and missed alarm rate of the system.

[0084] Embodiment 3:

[0085] Selection of nine-axis sensor 14:

[0086] The nine-axis sensor 14 selected by the present invention is a nine-axis micro-electromechanical vibration sensor, which is a vibration sensor integrated with a three-axis gyroscope, a three-axis accelerometer, and a three-axis magnetometer. The nine-axis micro-electromechanical vibration sensor is based on time-frequency characteristic time series spectrogram for machine learning and modeling, and combines the wind speed and rainfall sensor information, applies an exclusive research and development algorithm to completely optimize weather interference except extreme meteorological conditions, and links the video to greatly reduce the false alarm rate.

[0087] Installation of the nine-axis sensor 14:

[0088] Along the circumference of the airport fence, each mesh is installed with a nine-axis sensor 14, which can also be called a vibration sensor. The nine-axis sensors 14 are connected and networked by aviation connectors. The nine-axis sensors 14 are fixed on the fence with a buckle mesh structure. All nine-axis sensors 14 form dense partitioned distributed vibration detection, which can accurately identify intrusion behavior, reduce false alarm rate, and accurately locate the intrusion partition with a positioning error of ≤3m. Special cables connecting vibration detectors, signal processors, and field controllers are laid along the fence and 0.3m below the ground through PVC pipes. Protective measures are set at each signal processor and field controller to prevent damage. The fence intrusion detection subsystem consists of a vibration detector, a signal processor, a field controller, a dry contact expansion module, an alarm control software, etc. The field controller receives alarm and warning information through the serial port and the network, and controls the field equipment through the I / O and network interface. Once an intrusion signal occurs, the system generates an alarm, and the LED lights, sound and light alarms, and the video system are switched to the preset position through the pre-set linkage strategy for remote evidence collection and video tracking.

[0089] Working principle of the nine-axis sensor 14:

[0090] The nine-axis sensor 14 is distributed in a nine-axis micro-electromechanical vibration sensor array. Its vibration measurement technology uses a nine-axis micro-electromechanical system MEMS sensor chip to realize a new distributed vibration sensor network. It is a highly accurate and highly sensitive distributed vibration measurement technology. The core of this technology is a nine-axis micro-electromechanical vibration sensor detector, which has a built-in gyroscope, accelerometer, and magnetometer, which can detect mechanical and physical changes such as vibration, swing, and roll of the sensor body. In the field of perimeter security, relying on the high-precision and high-linear restoration ability of the vibration signal of this technology, it is possible to accurately identify the category of the intrusion event by analyzing the signal characteristics to distinguish the intrusion behavior, which not only ensures zero missed reports, but also has a very low false alarm rate. This technology uses a nine-axis micro-electromechanical sensor chip as a detection unit and a high-speed microprocessor as a data acquisition and digital filtering unit. Built-in gyroscopes, addends, and magnetometers can detect mechanical and physical changes such as vibration, swing, and roll of the sensor body. Since the mass of the detection IC unit itself is extremely small, it will neither resonate nor be disturbed by external electric and magnetic fields, and has a strong anti-interference ability. It can detect very weak mechanical vibration signals and small posture changes with high detection sensitivity.

[0091] Advantages of using distributed nine-axis sensors 14: ① Due to the use of distributed processing to reduce the pressure on the central server, an ultra-long fence alarm system can be formed, which can realize long-distance distributed vibration measurement. ② Due to the use of nine-axis sensing element design, the initial installation posture of the sensor is automatically memorized, and the real-time posture data is collected up to 25 times per second, so the system has high vibration measurement sensitivity. Since the vibration sensor unit is fixed on each mesh of the fence, and the distance between each mesh is 3 meters, the vibration measurement accuracy is high.

[0092] Embodiment 4:

[0093] The following is a description of the edge computing all-in-one machine 12:

[0094] The edge computing integrated machine 12 is a distributed cloud processing computer, which reduces the pressure on the central server, so that it can form the characteristics of the ultra-long perimeter alarm system and has an open third-party interface, which can at least access one or more of the alarm system, lighting system, warning light and siren system, Internet platform, and geographic information system. The location information of the alarm point is pushed to the software management platform.

[0095] Embodiment 5:

[0096] The following is the technical solution of the third-party linkage system:

[0097] like Figure 2-3 As shown, the airport perimeter intelligent security system also includes a third-party linkage system networked with the front-end sensing system, and the third-party linkage system at least includes a video monitoring system, an audible and visual alarm system, a night auxiliary lighting system, a voice broadcast system, and a wind speed and rainfall monitoring system;

[0098] The video surveillance system is connected to the switch 11. When the perimeter is invaded and an alarm is triggered, the generated alarm signal drives the video surveillance system to remotely review;

[0099] The sound and light alarm system includes a control room end sound and light alarm system and an on-site sound and light alarm system, wherein the control room end sound and light alarm system is driven by a sensor processor capable of generating a vibration signal in the front-end sensing system 1, and the on-site sound and light alarm system and the night auxiliary lighting system are driven by an I / O module in the front-end sensing system 1;

[0100] The voice broadcast system is provided with an IP audio terminal 3, a loudspeaker, an audible and visual alarm, which are distributed in the perimeter intrusion identification area and connected to the switch, and a control center broadcasting device which is located in the control room and connected to the switch;

[0101] The wind speed and rainfall monitoring system is a meteorological unit installed in the enclosure intrusion identification area for real-time monitoring of the meteorological conditions in the enclosure intrusion identification area. The meteorological unit will add the obtained real data to the sound and light alarm system algorithm in a timely and reliable manner, and eliminate the interference vibration caused by severe weather conditions through analysis and processing.

[0102] Specifically, the video surveillance system is connected to the switch in the nine-axis micro-electromechanical vibration sensor intrusion detection system with a network cable. After the software is set, the alarm signal can drive the video surveillance system to remotely review after the perimeter intrusion alarm system is triggered. The main feature of this method is to use network control signals to control and trigger other security subsystems to respond. The perimeter system server and the network ring switch are connected to form an expandable local area network. Through the expansion, development and integration of the industrial video system, it can be achieved that after the alarm signal of the perimeter intrusion alarm system is triggered, network communication is carried out through network signals and TCP protocol to perform program control on the control object. The perimeter security system software is relatively independent of the video surveillance system software. It is bound to the camera (preset position) through the alarm point, and the Ethernet access method requests video browsing from the video surveillance system platform or NVR, and sends pan / tilt and video matrix control commands. At the same time, the vibration sensor perimeter intrusion alarm management platform can deploy and disarm third-party alarm equipment in real time. The alarm linkage process is as follows: Figure 3 shown.

[0103] 1. Technical solution for sound and light alarm and lighting linkage: Sound and light alarm linkage is divided into two parts: sound and light alarm linkage at the control room end and sound and light alarm linkage at the field end. The sound and light alarm equipment at the control room end can be directly driven by the vibration sensor signal processor. The sound and light alarm and lighting at the field end are driven by the remote I / O module connected to the front-end security network. The perimeter monitoring alarm integrated management software can manually or automatically control the sound and light alarm and lighting equipment. Each defense zone is equipped with 2 sets of 30W horn speakers and 1 set of sound and light alarm, all installed on the camera pole, with an installation height of 3.2 meters.

[0104] 2. Broadcast intercom linkage technical solution: The front-end perimeter broadcast subsystem consists of control center equipment and IP audio terminals, speakers, sound and light alarms, etc. Each defense zone is equipped with a set of IP audio terminals, which are installed in the outdoor equipment control box or outdoor equipment box. When the perimeter alarm system detects intrusion information, the perimeter monitoring alarm integrated management software can manually or automatically control the front-end broadcast equipment and play customized alarm voices; the on-duty personnel can also speak to the scene through the microphone.

[0105] 3. Wind speed and rainfall monitoring technical solution: An independent meteorological unit is set up at the front end. A meteorological unit specially used for detecting wind, rain and hail is installed in the system to monitor the meteorological conditions of the surrounding area in real time. The real data obtained is added to the alarm system algorithm in a timely and reliable manner, and the interference vibration caused by severe weather conditions is eliminated through analysis and processing.

[0106] In summary, the technical effects achieved by the airport perimeter intelligent security system of the present invention after operation are as follows:

[0107] 1) The system uses large-capacity and high-density vibration sensor detectors as sensors to accurately monitor vibration signals at any location in the entire defense zone in real time. Through rapid analysis and behavior recognition of vibration signals at various locations, it can accurately determine intrusion events and locations, send out alarm signals, and store alarm information.

[0108] 2) The alarm system has no blind spots or dead angles in its deployment area, and the minimum defense area is 3 meters. The defense area is divided mainly by physical means, supplemented by software, and realized by combining the front and back ends. Each group of vibration sensor detectors at the front end is independent of each other.

[0109] 3) In the armed state, if an intrusion occurs in the defense area (climbing the fence, destroying the fence, entering the defense area, etc.), the alarm control device will display the location where the alarm occurred. When multiple channels and multiple points alarm at the same time, the alarm control device will display the area or address where the alarm occurred in turn, and send out sound and light alarm information. The alarm information should be able to be maintained until manual reset, and the alarm signal should not be lost.

[0110] 4) In the defense state, if an intrusion occurs in the defense area, the video surveillance system camera can be linked to the alarm area, and the control room will immediately send out an audible and visual alarm, so that the monitoring personnel can intuitively and in real time grasp the on-site situation and perform targeted related operations according to the actual situation on the scene.

[0111] 5) The on-site equipment has anti-dismantling function and fault alarm function. When the front-end processor is turned on, the transmission line is broken, the control host power fails, etc. in the armed or disarmed state, the alarm control device can send out sound and light alarm information. The alarm information can be maintained until manual reset, and the alarm signal should not be lost.

[0112] 6) After the alarm occurs, the alarm platform can trigger the sound and light alarm to send out sound and light alarm information, and the alarm information can be kept until it is manually reset without loss. In the disarmed state, the system does not respond to intrusion events.

[0113] 7) The system has the function of displaying and recording information such as alarms, faults, and damage (including power on, power off, disarming, and changes). System records should include the time, location, and nature of the event (disconnection, intrusion), alarm record processing information (confirmed, unconfirmed), etc. The recorded information cannot be changed, but alarm record statistical reports can be generated. The system only allows users with system administrator privileges to operate on the data.

[0114] 8) The signal processor checks each defense zone in real time. If any defense zone is damaged, it will send out sound and light alarm information and display the fault location on the electronic map.

[0115] 9) The system response time shall not exceed 3s.

[0116] 10) The detection equipment has strong resistance to adverse weather interference, and the sensitivity can be adjusted according to different site conditions and climate. The sensitivity of each defense zone can be set separately. At the same time, the sensor is not affected by lightning, electromagnetic and radio interference, and can work stably in harsh environments.

[0117] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An airport fence intelligent security system, comprising a front-end sensing system (1), a data transmission system (2) and a central control system connected in sequence through electrical signals and communication signals, wherein the front-end sensing system (1) is arranged in a partitioned network, and any partitioned front-end sensing system (1) is provided with a nine-axis sensor capable of collecting dynamic signals and static signals of the fence surrounding environment in real time, an edge computing integrated machine (12) connected to the nine-axis sensor, and a switch (11) connected to the edge computing integrated machine (12), characterized in that: The edge computing integrated machine (12) receives environmental data sensed by the nine-axis sensor and posture data generated by the posture change of the nine-axis sensor itself, and processes the signal data to extract features through the edge computing integrated machine (12); The edge computing integrated machine (12) identifies the intrusion mode and the non-intrusion mode based on the extracted feature data. The edge computing integrated machine (12) uses a cloud big data model and combines the big data model of the central control system to conduct in-depth analysis and prediction of the feature data, classify the feature data, identify the intrusion behavior category, and make a decision on whether to issue an alarm; The airport perimeter intelligent security system further includes an artificial intelligence learning module, which includes an artificial intelligence learning method, and the artificial intelligence learning method includes the following steps: Step 1: The nine-axis sensor senses and measures mechanical vibrations within a certain range around itself, and sends the sensed and measured feature data to the edge computing all-in-one machine (12); Step 2: the edge computing integrated machine (12) analyzes the feature data sent back by each nine-axis sensor through a big data model; Step 3: Compare the feature data sent back by adjacent or all nine-axis sensors through an artificial intelligence learning algorithm, and then combine it with the current weather information of the airport area to identify whether there is any abnormal feature data; Step 4: Accurately determine the intrusion behavior category based on the judged feature data.

2. The airport perimeter intelligent security system according to claim 1, characterized in that: The method for extracting features adopts Fourier transform $$F(\omega)=\sum_{-\infty}^{\infty}f(t)e^{-j\omega t}dt$$ and wavelet transform method to extract key features.

3. The airport perimeter intelligent security system according to claim 1, characterized in that: The method for classifying feature data by the edge computing integrated machine (12) is XGBoost algorithm classification.

4. The airport perimeter intelligent security system according to claim 3, characterized in that: The XGBoost algorithm classification formula is expressed as: $$F_m(x)=F_{m-1}(x)+\lambda\cdot h(x;\gamma_m,\nu_m)$$; in: $$F_m(x)$$ is the prediction result of the mth tree for sample x; $$F_{m-1}(x$$ is the prediction result of the first m-1 trees for sample x; $$\lambd$$ is a regularization parameter used to control model complexity; $$h(x;\gamma_m,\nu_m$$ is the prediction function of the mth tree, $$\gamma_m and \nu_m$$ are the parameters of the tree respectively; The optimization goal of the XGBoost algorithm is to minimize the following objective function: $$Obj(F)=\sum_{i=1}^{N}l(y_i,F(x_i))+\sum_{m=1}^{M}[\Omega(f_m)$$ in: $$$$ is the sample size; $$l(y_i,F(x_i)$$ is the loss function, which measures the error between the model prediction and the actual value; $$$$ is the number of trees; $$\Omega(f_m$$ is the regularization term, including the complexity of the tree).

5. The airport perimeter intelligent security system according to claim 1, characterized in that: The nine-axis sensor is a vibration sensor that integrates a three-axis gyroscope, a three-axis accelerometer, and a three-axis magnetometer.

6. The airport perimeter intelligent security system according to claim 1, characterized in that: The edge computing integrated machine (12) is a distributed cloud processing computer having an open third-party interface, which can at least be connected to one or more of an alarm system, a lighting system, a warning light and sirens system, an Internet platform, and a geographic information system.

7. The airport perimeter intelligent security system according to claim 1, characterized in that: The characteristic data at least includes the vibration amplitude and frequency of the nine-axis sensor and the tilt angle of the nine-axis sensor itself.

8. The airport perimeter intelligent security system according to claim 1, characterized in that: The airport perimeter intelligent security system also includes a third-party linkage system networked with the front-end sensing system, and the third-party linkage system at least includes a video monitoring system, an audible and visual alarm system, a night auxiliary lighting system, a voice broadcast system, and a wind speed and rainfall monitoring system; The video surveillance system is connected to the switch (11), and when an alarm is triggered by an intrusion into the perimeter, the generated alarm signal drives the video surveillance system to conduct a remote review; The sound and light alarm system comprises a control room end sound and light alarm system and an on-site sound and light alarm system, wherein the control room end sound and light alarm system is driven by a sensor processor capable of generating a vibration signal in the front-end sensing system (1), and the on-site sound and light alarm system and the night auxiliary lighting system are driven by an I / O module in the front-end sensing system (1); The voice broadcast system is provided with IP audio terminals, loudspeakers, sound and light alarms distributed in the perimeter intrusion identification area and connected to the switch, and a control center broadcasting device located in the control room and connected to the switch; The wind speed and rainfall monitoring system is a meteorological unit installed in the enclosure intrusion identification area for real-time monitoring of the meteorological conditions in the enclosure intrusion identification area. The meteorological unit will add the obtained real data to the sound and light alarm system algorithm in a timely and reliable manner, and eliminate the interference vibration caused by severe weather conditions through analysis and processing.

9. The airport perimeter intelligent security system according to claim 1, characterized in that: The nine-axis sensors are connected to form a network by using aviation connectors, and the nine-axis sensors are fixed on the fence by using a buckle net structure.

Images