All-weather specific target monitoring and alarm device, server, storage medium

Abstract

Disclosed is an all-weather specific target monitoring and alarm device, comprising: an alarm output module; a microwave radar and a PIR sensor, which are respectively configured to detect whether there is a target intrusion; a controller, which is configured to implement: obtaining the detection result of the microwave radar; obtaining the detection result of the PIR sensor; inputting the detection result of the PIR sensor and the detection result of the microwave radar into a trained ANN network to fuse and decide whether there is a target intrusion, and the weighting coefficients of the ANN network for fusing the detection results of the microwave radar and the PIR sensor consider the influence of geographical information and meteorological information on the detection accuracy of the radar and the PIR sensor; when it is judged that there is a target intrusion based on at least one of the detection result of the microwave radar, the detection result of the PIR sensor, and the fusion decision result of the ANN network, the alarm output module generates an alarm message, and sends the detection result of the microwave radar, the detection result of the PIR sensor, and the fusion decision result of the ANN model to the server through the communication module.

Description

Technical Field

[0001] The present invention belongs to the field of security systems, and particularly relates to an all-weather specific target monitoring and alarming device, a server, and a storage medium. Background Art

[0002] In large substations, military secrecy sites, forest fire prevention areas, field weather stations, large museums and other places, a monitoring system is needed to prevent specific targets such as people and large animals from abnormally invading the warning area. How to accurately monitor and alarm specific targets in all-weather and multi-terrain scenarios is a difficult problem to be solved in this field. Specific targets refer to people illegally invading the warning area or large animals such as dogs, wolves, foxes, and cows. The following respectively elaborates on the corresponding research status at home and abroad for three key technologies related to this difficult problem. These three key technologies are intrusion monitoring technology, multi-sensor fusion technology in complex scenarios, and networking technology for unattended warning lines.

[0003] Currently, the commonly used mainstream specific target intrusion monitoring technologies include active infrared pair, electronic pulse fence, vibrating optical fiber, passive infrared, and microwave radar, etc. The active infrared pair has a transmitter and a receiver, and the emitted and received light beams will form an invisible infrared fence. This detection method is only suitable for places with regular boundaries and less occlusion, and is not suitable for outdoor scenarios such as forests, shrubs, or grasslands; the electronic pulse fence alarm system adopts a pulsed voltage. When someone touches the electronic fence, the fence will be short-circuited or open-circuited, and the alarm system will send an alarm signal. Although the false alarm rate is very low, the pre-installation is complex, the energy consumption is large, and it does not have concealment and is easy to be discovered and cracked; the vibrating optical fiber needs to be installed on an existing fence and has a high cost; passive infrared detection is a sensor technology that monitors the infrared rays emitted by the human body's radiation, and is a non-contact sensor technology for detecting human intrusion, but the false alarm rate is relatively high in high-temperature weather; microwave radar refers to a radar operating in the microwave frequency band, which is a sensor technology for monitoring whether there are objects moving in the warning area, but the false alarm rate is relatively high in strong wind weather.

[0004] Currently, intrusion detection devices mainly target indoor scenarios and small outdoor scenarios in family courtyards, and are not suitable for large-scale outdoor scenarios. The monitoring accuracy in outdoor conditions is relatively low. The reason is that most devices usually have only one sensor, an infrared or microwave sensor, and cannot support the monitoring of specific targets in complex scenarios; even devices containing two or more sensors such as infrared and microwave only perform simple "AND" logic or "OR" logic operations on the monitoring results of each sensor. Since the outdoor environment is more complex, with diverse terrains and changing weather, when existing intrusion monitoring devices are applied to outdoor environments, due to the failure to fully consider the performance changes of infrared and microwave sensors in different scenarios, the false alarm rate will increase.

[0005] At present, the alarm information output interfaces of detection devices are mostly relays, and additional communication modules need to be configured to form a network. The monitoring range of a single device is limited. Only by networking and jointly monitoring can the monitoring range be expanded to form an electronic warning line. However, relay interfaces are mostly used for direct connection with alarm devices. When there are many monitoring devices, the number of interfaces of the alarm host is insufficient, which is not convenient for multi-device networking, not easy to use, and has a high cost. Moreover, it is difficult to play a positioning role and it is difficult to know the operating status of the device. Summary of the Invention

[0006] The present invention aims to solve the technical problem of monitoring and early warning of specific targets in outdoor all-weather multi-terrain scenarios. The present invention combines the advantages of two sensors, namely microwave radar and passive infrared detection, and uses these two sensors in combination to improve the accuracy and stability of monitoring. By comprehensively analyzing the monitoring results of these two sensors, specific targets such as people and large animals can be effectively monitored, and abnormal intrusion situations can be warned in real time.

[0007] In a first aspect, there is provided an all-weather specific target monitoring and alarm device, including: an alarm output module; a microwave radar configured to detect whether a target has intruded; a PIR sensor configured to detect whether a target has intruded; a communication module responsible for information transmission between the device and a server; a controller connected to the alarm output module, the microwave radar, and the PIR sensor, and the controller is configured to implement:

[0008] Obtain the detection result of the microwave radar; obtain the detection result of the PIR sensor; input the detection result of the PIR sensor and the detection result of the microwave radar into a trained ANN network to fuse and decide whether a target has intruded; when it is determined that a target has intruded based on at least one of the detection result of the microwave radar, the detection result of the PIR sensor, and the fusion decision result of the ANN network, cause the alarm output module to generate an alarm message, and send the detection result of the microwave radar, the detection result of the PIR sensor, and the fusion decision result of the ANN model to the server via the communication module.

[0009] In some examples, the weighting coefficients for the ANN network to fuse the detection results of the microwave radar and the PIR sensor consider geographical information and meteorological information.

[0010] In some examples, there are multiple PIR sensors, which respectively detect target intrusions in different directions. The detection results of all the PIR sensors are fused according to an OR logic, and the fused detection result of the PIR sensors and the detection result of the microwave radar are input into the ANN network to fuse and decide whether a target has intruded.

[0011] In some examples, the vector representation of the detection result of the PIR sensor input to the ANN network and the detection result of the microwave radar is Z k =[Z pir (k), Z w (k)] T , where k represents the time series, and the vector Z k is weighted by the weighting coefficient ω k =[ω pir (k), ω w (k)] T to obtain an estimated value This estimated value is sent to a decision maker to obtain the fusion decision result. The mathematical expression of the fusion decision result is Z k ∈{0, 1}, where the input signals of the MLP neural network include: the estimated value S k , the training data set and the d k corresponding to S k , geographical information, meteorological information, and the detection result Z pir (k) of the PIR sensor and the detection result Z w (k) of the microwave radar.

[0012] In some examples, the controller is configured to obtain the geographical information and the meteorological information from a server.

[0013] In a second aspect, a server is provided, including: one or more processors; and a memory for storing one or more computer programs, which when executed by the one or more processors implement: obtaining the detection result of the microwave radar, the detection result of the PIR sensor, and the fusion decision result of the ANN model sent by the all-weather specific target monitoring and alarm device; sending to the all-weather specific target monitoring and alarm device an update of the weighting coefficient of the ANN model calculated from the geographical information and the meteorological information.

[0014] In a third aspect, a computer-readable storage medium is provided, on which a computer program is stored, which when executed by one or more processors implements: obtaining the detection result of the microwave radar, the detection result of the PIR sensor, and the fusion decision result of the ANN model sent by the all-weather specific target monitoring and alarm device; sending to the all-weather specific target monitoring and alarm device an update of the weighting coefficient of the ANN model calculated from the geographical information and the meteorological information. Description of the Drawings

[0015] Figure 1It is a block diagram of an all-weather specific target monitoring and alarm device provided by an embodiment of the present invention.

[0016] Figure 2a It is a schematic structural diagram of an all-weather specific target monitoring and alarm device provided by an embodiment of the present invention.

[0017] Figure 2b It is an exploded view of an all-weather specific target monitoring and alarm device provided by an embodiment of the present invention.

[0018] Figure 3 It is an installation schematic diagram of an all-weather specific target monitoring and alarm device provided by an embodiment of the present invention.

[0019] Figure 4 It is a software design flow chart of an all-weather specific target monitoring and alarm device provided by an embodiment of the present invention.

[0020] Figure 5 It is a schematic diagram of the detection and fusion of infrared and microwave sensors provided by an embodiment of the present invention.

[0021] Figure 6 It is a schematic diagram of ANN fusion provided by an embodiment of the present invention. Detailed implementation manners

[0022] The microwave and PIR (Passive Infrared) sensors are small in size, good in concealment, low in power consumption, and strong in stability. Moreover, the two sensors have strong complementarity. The present invention combines these two sensors to invent an all-weather multi-terrain specific target monitoring device. The present invention also proposes a fusion method of infrared and microwave based on weather and terrain to improve the accuracy of specific target monitoring.

[0023] Figure 1 The block diagram of the all-weather specific target monitoring and alarm device is shown. As Figure 1 shown, the device includes a power supply module, a sensor module, a control module, an alarm output module, and a communication module.

[0024] The power supply module powers other modules of the device to ensure the normal operation of the device. The sensor module includes a microwave radar and a PIR (Passive Infrared). One microwave radar can be set to monitor the entire warning area. Five PIRs can be set, one downward-looking PIR and four side-looking PIRs. The downward-looking PIR is used to monitor the area directly below the device, and the four side-looking PIRs are respectively used to monitor the areas around the device. The present invention does not limit the number of the microwave radar and the PIR. The communication module can transmit information bidirectionally with the server. The communication module can adopt any known communication protocol, such as the 485 communication protocol. The control module is a single-chip microcomputer, which is used to complete the acquisition of sensor data, the calculation of alarm information, and the sending and receiving of information, etc. The single-chip microcomputer reads data from the sensors, respectively detects whether there is intrusion information according to the monitoring data of each sensor, and then uses the trained ANN model to fuse and decide whether there is a specific target intrusion, and encodes and saves the decision result. At the same time, the device also needs to receive polling information from the server through the communication module. When the device receives the polling information, it sends the encoded decision result through the communication module; at the same time, the polling information also includes weather information, and the device updates the fusion decision model parameters according to the weather information.

[0025] The meteorological information includes temperature, wind speed, light, etc. The geographical information is the geographical environment where the monitoring and alarm device is located, such as desert, grassland, forest, etc.

[0026] The structure of the outer shell 1 of the device is as Figure 2a , 2b shown. The overall shape is a bucket shape with the big head facing up, and there is a bracket 2 for fixing at the upper part. The prior art detection devices are usually wall-mounted, lacking a structure convenient for installation under outdoor conditions. The detection range is fan-shaped and can be directly installed on the wall in scenarios such as families and small courtyards. However, in larger outdoor scenarios, such as forest fire prevention, airports, etc., attention needs to be paid to the installation angle when multiple devices are installed at the same time, and it is difficult to find a wall for installation under field conditions. Four side surfaces 3 and the lower bottom surface 4 of the device of the present invention are inlaid with Fresnel lenses 5 to gather infrared rays in the environment, and there is a hole 6 on one side surface for wire routing; the interior of the device is divided into two layers. The lower layer is a downward-looking infrared sensor 7 and a microwave radar 8 for monitoring the area directly below, and the upper layer is a circuit board, including a controller, a signal processing circuit, an alarm output module, and 4 side-looking infrared sensors 7. There are also several interfaces on the circuit board for power supply and communication; there is also an insulating board 9 above the circuit board main board to protect the circuit board main board, and a skylight is opened for wire routing and maintenance. The structure of the device is not limited to Figure 2a , Figure 2b, those skilled in the art can design according to the actual situation, but the premise is that a microwave radar and a PIR sensor are integrated on the device. In addition, the number of the microwave radar and the PIR sensor is not limited in the present invention, as long as the monitoring ranges of the two completely cover the area where the device is located without dead angles.

[0027] As Figure 3 shown in the schematic diagram of the networking installation of the device, the device is suspended and installed on a cable about 6m above the ground, and the distance between each device is about 10 - 15m. Of course, the height of the device from the ground and the interval between devices are not limited to this. The devices are connected in sequence to form an invisible electronic fence, which can monitor people and animals passing through this electronic fence and will not hinder the activities of people and animals like a physical fence.

[0028] As Figure 4 shown, the controller reads the sensor data from the microwave radar and the PIR (Passive Infrared), and then processes and calculates each sensor data respectively. The microwave radar and the PIR sensor independently detect whether there is a specific target intrusion. Whether or not an intrusion is detected, the data of the two types of sensors need to be fused for decision-making. If a specific target intrusion is detected, the intrusion alarm information detected by each of them also needs to be recorded respectively. Then the data is summarized, including the alarm information of the microwave radar and the PIR respectively, the result of the fusion decision-making, etc. Finally, the summarized data is encoded and saved locally waiting to be sent. At the same time, the controller also needs to receive information from the server. The information of the server includes the address information. If the address information matches the address of this device, the encoded information of this device is sent to the server and the parameters of the fusion decision-making are updated.

[0029] Data structure: The data uploaded from the device end to the server end includes the number and operation status of this device, the alarm information of the microwave radar and the PIR respectively, and the result of the fusion decision-making. The format of one frame of data is:

[0030] · Start bit: One byte, used to represent the start position of the data packet, set to 0xAA.

[0031] · Device address: One byte, used to identify the device address of the lower computer.

[0032] · PIR alarm bit: One byte, representing the alarm status of the PIR sensor, 0x01 for alarm and 0x00 for no alarm.

[0033] · Microwave radar alarm bit: One byte, representing the alarm status of the microwave radar sensor, 0x01 for alarm and 0x00 for no alarm.

[0034] · Fusion decision-making alarm calculation result: Two bytes, representing the result of the fusion decision-making calculation.

[0035] · Fusion decision result: one byte, representing the fusion decision alarm result. 0x01 indicates alarm, and 0x00 indicates no alarm.

[0036] · Parity check: one byte, used to verify the integrity and correctness of data.

[0037] As Figure 4 shown in the flowchart on the right, it is the process of information transmission and reception of the device. To ensure the real-time nature of data interaction, the information transmission and reception are completed in the interrupt and have the highest priority. When the device receives an instruction from the server, it directly enters the interrupt and determines whether the target device of the instruction is this device according to the address bit in the instruction. If it is this device, it updates the parameters of the fusion decision and uploads the encoded alarm information. If it is not this device, it jumps out of the interrupt and enters the monitoring and early warning state.

[0038] As Figure 5 shown, it is the schematic diagram of the detection fusion of infrared and microwave sensors, mainly including three steps: 1. Each infrared and microwave sensor detects a specific target separately; 2. Five infrared sensors detect in five different directions respectively, and their spatial intersection is empty, that is, the intersection of the five detection results is an empty set. Therefore, the detection results of the five infrared sensors are fused according to the OR logic; 3. The fusion detection result of the infrared sensor is fused with the detection result of the microwave sensor according to the linear model of ANN.

[0039] Specifically as follows:

[0040] 1. The target detection models of single sensors (infrared and microwave) are as follows:

[0041] H 0 : y[n] = ω[n]; n ∈ {0, …, N - 1}

[0042] H 1 : y[n] = A + ω[n]

[0043] n is the time series, ω is the noise, A is the signal generated by the infrared sensor when the target appears, H 0 indicates that the sensor does not detect the target, the detection result is 0, and the observed value of the sensor is the noise of the device; H 1 indicates that the sensor detects the target, the detection result is 1, and the observed value of the sensor is the detection result (high level) plus the noise of the device.

[0044] Taking the minimization of the error probability as the objective function:

[0045] P e = P(H 0 / H 1 )P(H 1 ) + P(H 1 / H0 )P(H 0 )

[0046] If i.e., p(y / H 1 )P(H 1 ) > p(y / H 0 )P(H 0 ), then it is determined that the target appears. Th is the detection threshold. According to Bayesian theory, the target detection formula is deduced as: P(y / H 1 ) > P(y / H 0 ).

[0047] Also assume that the noise ω[n] is normally distributed as ω[n] ~ N(0, σ 2 ), and then the error probability is obtained as:

[0048]

[0049] Q(·) represents the Q function. The detection error probability of a single sensor monotonically decreases with

[0050] 2. Fusion of 5 infrared sensors

[0051] Five infrared sensors respectively monitor targets in five different directions. Assume D 1 , D 2 , D 3 , D 4 , D 5 respectively represent the detection results in five directions. Then We fuse the detection results of the five infrared sensors with "OR" logic to obtain the detection result Z pir = D 1 + D 2 + D 3 + D 4 + D 5 .

[0052] Z pir 's detection error probability is equivalent to that of a single infrared sensor, and the correct probability is P r = 1 - P e .

[0053] 3. Fusion of infrared sensors and microwave sensors

[0054] The microwave sensor and the infrared sensor detect the target independently of each other. Since the target detection mechanisms of the microwave and infrared sensors are different, and the environment and meteorology have different impacts on these two types of sensors, the multi-sensor fusion method based on traditional statistical theory is not applicable to the fusion of these two types of sensors. We use the machine learning model ANN (artificial neural networks) to fuse the target detection results of the microwave and infrared sensors. The specific steps are as Figure 6 shown: Input the decision result vectors Z k =[Z pir (k), Z w (k)] T of the infrared and microwave sensors into the ANN network, where k represents the time series. The vector Z k is weighted by the weighting coefficients ω k =[ω pir (k), ω w (k)] T generated by the MLP neural network to obtain the estimated value This estimated value is sent to the decision maker to obtain the fused decision result Z k ∈{0,1}. The input signals of the MLP neural network include: the estimated value S k , the training data set d k , the geographical information Geo, the meteorological information Atm, and the decision results Z pir (k) and Z w (k) of the infrared and microwave.

[0055] The present invention also provides an embodiment of a server. The server can be a physical machine and cloud service. The server includes a processor and a memory. The memory and the processor can be interconnected through a bus system and / or other forms of connection mechanisms. The memory is used to store non-temporary instructions (such as one or more program modules). The processor is used to run the non-temporary instructions. When the non-temporary instructions are run by the processor, they can execute: sending polling information to the all-weather specific target monitoring and alarm device; obtaining the data sent by the all-weather specific target monitoring and alarm device (the data format has been introduced in detail before); sending the ANN model weighting coefficients calculated based on the current geographical information and meteorological information to the all-weather specific target monitoring and alarm device to update the ANN model weighting coefficients of the device.

[0056] For example, the processor can be a central processing unit (CPU), a graphics processing unit (GPU), or other forms of processing units with data processing capabilities and / or program execution capabilities. For example, the central processing unit (CPU) can be of the X86 or ARM architecture, etc. The processor can be a general-purpose processor or a dedicated processor, and can control other components in the electronic device to perform the desired functions.

[0057] For example, the memory can be volatile memory and / or non-volatile memory. Volatile memory can include, for example, random access memory (RAM) and / or cache memory, etc. Non-volatile memory can include, for example, read-only memory (ROM), hard disk, erasable programmable read-only memory (EPROM), compact disc read-only memory (CD-ROM), USB memory, flash memory, etc. One or more program modules can be stored on the memory, and the processor can run one or more program modules to implement various functions of the electronic device.

[0058] The present invention also provides a storage medium for storing non-temporary instructions that, when executed by a processor, can perform: sending polling information to the all-weather specific target monitoring and alarm device; obtaining data sent by the all-weather specific target monitoring and alarm device (the data format has been introduced in detail above); sending the ANN model weighting coefficients calculated based on the current geographical information and meteorological information to the all-weather specific target monitoring and alarm device to update the ANN model weighting coefficients of the device. For the relevant description of the storage medium, reference can be made to the corresponding description of the memory of the server above, which will not be elaborated here.

[0059] Advantages of the present invention:

[0060] All-weather adaptability: The present invention uses the combination of a microwave radar and a passive infrared detector, making full use of their complementarity. The combination of the two can improve the accuracy and reliability of monitoring and alarming for specific targets. The present invention is applicable to various places with complex meteorological conditions. Whether in harsh weather conditions, such as strong wind weather, or in high-temperature weather, this device can maintain stable monitoring performance and reduce the false alarm rate.

[0061] Convenient outdoor installation: Considering the complexity of the outdoor environment, the present invention designs an external structure that is convenient for outdoor installation. The device is no longer limited to wall-mounted installation, but can be adapted to various outdoor scenarios through flexible mounting brackets to obtain a better monitoring range and field of view.

[0062] Reliable communication networking interface: The device of the present invention uses an alarm output interface based on the 485 communication protocol. Compared with the traditional relay interface, it can network hundreds of monitoring and warning devices arranged in a long-distance linear pattern, realizing reliable information transmission between the device and the device management service system. Moreover, it is convenient to expand monitoring devices and is conducive to installation and debugging.

Claims

1. An all-weather specific target monitoring and alarm device, characterized in that, comprising: A microwave radar configured to detect whether a target has invaded; A PIR sensor configured to detect whether a target has invaded; An alarm output module; A communication module; A controller connected to the alarm output module, the communication module, the microwave radar and the PIR sensor, and the controller is configured to perform: Obtain the detection result of the microwave radar; Obtain the detection result of the PIR sensor; Among them, the detection models of a single microwave radar and PIR sensor are as follows: H 0 : y[n] = ω[n]; n ∈ {0, …, N - 1} H 1 : y[n] = A + ω[n] n is a time series; ω is noise; A is the signal generated by the sensor when the target appears; H 0 indicates that the sensor does not detect the target, the detection result is 0, and the observed value of the sensor is the noise of the device; H 1 indicates that the sensor detects the target, the detection result is 1, and the observed value of the sensor is the detection result plus the noise of the device; Based on the above detection models, with minimizing the error probability as the objective function, the detection results of the PIR sensor and the microwave radar are obtained respectively; The vector representation Z of the detection result of the PIR sensor and the detection result of the microwave radar k = [Z pir (k), Z w (k)] T is input into the ANN network, where: Z pir (k) and Z w (k) and Z are respectively the detection results of the PIR sensor and the microwave radar optimized with the minimization of the error probability as the objective function, where k represents the time series; Vector Z k The weighted coefficients ωk = [ωpir(k), ωw(k)] generated by the MLP neural network T After weighting, an estimated value is obtained And this estimated value is input into the decision maker to generate a fusion decision result. The mathematical expression of the fusion decision result is Z k ∈ {0, 1}; The input signals for training the MLP neural network include: the estimated value S k , the training data set d k , geographical information, meteorological information, and the detection result Z pir (k) of the PIR sensor and the detection result Z w (k); When it is determined that a target has invaded based on at least one of the detection result of the microwave radar, the detection result of the PIR sensor or the fusion decision result of the ANN network, trigger the alarm output module to generate an alarm message, and upload the detection result of the microwave radar, the detection result of the PIR sensor and the fusion decision result of the ANN network to the server through the communication module.

2. The all-weather specific target monitoring and alarm device according to claim 1, characterized in that, There are multiple PIR sensors, which respectively detect target invasions in different directions, and fuse the detection results of all PIR sensors according to the OR logic.

3. The all-weather specific target monitoring and alarm device according to claim 1, characterized in that, The controller is configured to obtain geographical information and meteorological information from the server.

4. A server, characterized in that, comprising: One or more processors; And a memory for storing one or more computer programs, which when executed by the one or more processors perform: Obtain the detection result of the microwave radar, the detection result of the PIR sensor and the fusion decision result of the ANN network sent by the all-weather specific target monitoring and alarm device according to any one of claims 1 to 3.

5. The server according to claim 4, characterized in that, Send geographical information and meteorological information for updating the weighting coefficients to the all-weather specific target monitoring and alarm device according to any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, A computer program is stored thereon, which when executed by one or more processors performs: Obtain the detection result of the microwave radar, the detection result of the PIR sensor and the fusion decision result of the ANN network sent by the all-weather specific target monitoring and alarm device according to any one of claims 1 to 3.

7. The computer-readable storage medium according to claim 6, characterized in that, When the computer program is executed by one or more processors, it performs: Send the weighting coefficients of the ANN network calculated from geographical information and meteorological information for updating to the all-weather specific target monitoring and alarm device according to any one of claims 1 to 3.

Images