Electricity larceny prevention analysis method and device

Abstract

The invention discloses an electricity larceny prevention analysis method and device, and relates to the technical field of electricity larceny prevention analysis, and the method comprises the steps: deploying a broadband electromagnetic sensor network, collecting the magnetic field intensity in real time, calculating the energy density, dynamically triggering a self-adaptive notch filter to suppress an interference signal, and combining fundamental wave frequency matching to achieve precise filtering; a strain sensor is used for monitoring deformation of the box body, deformation quantity calculation is optimized through Kalman filtering, and an electromagnetic locking device is activated to protect metering equipment; an RFID electronic seal is combined with image feature point Euclidean distance matching, the seal integrity is judged, and a high-sensitivity mode is triggered when the seal is abnormal; in intrusion detection, a confidence threshold of the YOLOv5 model is adjusted, infrared and millimeter wave radar data are fused to extract a motion track, a target path is predicted in combination with Kalman filtering, and a protection log of block chain evidence storage is generated. According to the invention, active defense and quick response of electricity larceny behaviors are realized, and the safety and anti-electricity larceny efficiency of metering equipment are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of anti-electricity theft analysis, and in particular to an anti-electricity theft analysis method and device. Background Art

[0002] With the development of intelligent power systems, electricity theft has become more technical and covert. In existing technologies, electricity thieves may use electromagnetic interference devices to interfere with the signals of metering equipment, resulting in distortion of electricity metering data or failure of equipment functions. However, traditional anti-electricity theft devices mostly rely on single sensor monitoring, lack the ability to dynamically suppress wide-band electromagnetic interference, and are difficult to adapt to the needs of accurate detection in complex electromagnetic environments.

[0003] A Chinese patent application with publication number CN104933272B discloses a method and device for anti-electricity theft analysis, including: obtaining abnormal electricity consumption events and corresponding event times of target users based on records in an electricity consumption information collection system, wherein abnormal electricity consumption events include electric energy meter operation events and electric energy meter abnormal events; when an electric energy meter abnormal event occurs after an electric energy meter operation event, determining that the target user is suspected of electricity theft. This method analyzes a large amount of historical data in the electricity consumption information collection system to find out certain specific sequence patterns related to electricity theft behavior, and then determines whether the user is suspected of electricity theft. It can improve the effectiveness and pertinence of using the electricity consumption information collection system to prevent and investigate electricity theft, give full play to the role of the electricity consumption information collection system in preventing electricity theft, and improve data utilization.

[0004] In the above-mentioned prior art, a large amount of historical data in the electricity consumption information collection system is analyzed to find out certain specific sequence patterns related to electricity theft behavior, and then judge whether the user is suspected of electricity theft. However, it lacks comprehensive consideration of electromagnetic interference and signal interference. The electricity thief can use electromagnetic interference equipment to generate a strong electromagnetic field to disrupt the anti-electricity theft detection device, making it unable to work normally or causing metering errors. At the same time, the safety and sealing performance of the anti-electricity theft analysis device is limited, and it is easy to be damaged, and it cannot effectively protect the metering equipment. Therefore, it is necessary to provide an anti-electricity theft analysis method and device to solve the above-mentioned problems. Summary of the invention

[0005] In order to solve the above technical problems, a method and device for anti-electricity theft analysis are provided. The technical solution solves the problem that the prior art proposed in the above background technology lacks comprehensive consideration of electromagnetic interference and signal interference. Electricity thieves can use electromagnetic interference equipment to generate strong electromagnetic fields to disrupt the anti-electricity theft detection device, causing it to be unable to work normally or cause metering errors. At the same time, the anti-electricity theft analysis device has limited safety and sealing performance, is more easily damaged, and cannot effectively protect the metering equipment.

[0006] In order to achieve the above purpose, the technical solution adopted by the present invention is: An anti-electricity theft analysis method, comprising: S1. Obtain the target user's metering equipment environment information, and deploy the sensor network based on the target user's metering equipment environment information; S2. The wide-band electromagnetic sensor in the sensor network collects the magnetic field strength of the power supply line in real time, and synchronously obtains the fundamental frequency of the metering equipment when it is running, so as to determine the magnetic field energy density. Then, according to the magnetic field energy density, it is determined whether to trigger the adaptive notch filter. If the adaptive notch filter is triggered, the adaptive notch filter dynamically generates a frequency suppression mask and outputs a signal X. clean (t); S3. Receive the signal X output by the adaptive notch filter clean (t), and extract its effective value and phase difference, and then determine whether to activate the box protection system based on the effective value and phase difference; S4. Obtain the switch status signal in the box protection system, and read the RSSI value and image feature points of the RFID electronic seal, and then determine whether the seal is abnormal based on the RSSI value and image feature points. If the seal is abnormal, the intrusion detection module is triggered; S5. After the intrusion detection module is triggered, it enters the high-sensitivity mode, adjusts the confidence threshold of the YOLOv5 model, and extracts the motion trajectory of the intrusion target through the fusion perception of the infrared camera and millimeter-wave radar in the sensor network. According to the motion trajectory of the intrusion target, a protection log is generated and stored in the blockchain.

[0007] In an optional embodiment, step S2 specifically includes: The magnetic field strength B(t) of the power supply line is collected in real time through a wide-band electromagnetic sensor array, and the fundamental frequency f0 of the metering equipment during operation is simultaneously obtained; The magnetic field strength of the power supply line is collected in real time using a wide-band electromagnetic sensor to obtain the magnetic field energy density E(t); When the magnetic field energy density E(t)>E threshold When (f0), the adaptive notch filter is triggered to synchronously acquire the original electrical signal, and then the original electrical signal is input into the adaptive notch filter for filtering; Get the frequency suppression mask M(f) dynamically generated by the adaptive notch filter and synchronously determine the output signal X of the adaptive notch filter clean (t), where E threshold (f0) is the threshold function; Wherein, the calculation formula of the magnetic field energy density E(t) is: Where E(t) is the magnetic field energy density of the power supply line at time t, is the magnetic field strength of the power supply line at time t, is the integral variable, representing the arrive Every moment in the time period, is the preset time window length; The threshold function E threshold The expression formula of (f0) is: In the formula, and are frequency band coefficients, It is the fundamental frequency of the measuring equipment when it is running; The output signal X of the adaptive notch filter clean The expression formula of (t) is: ; In the formula, is the signal after interference suppression output by the adaptive notch filter at time t, is the original electrical signal input into the adaptive notch filter at time t, is the number of interference frequency bands, is the frequency band index, For the The filtered value of the original electrical signal that needs to be filtered, For the A binary function of the filtered value of the original electrical signal that needs to be filtered.

[0008] In an optional embodiment, step S3 specifically includes: Receive the adaptive notch filter output signal X clean (t), through , extract X clean (t) Corresponding effective value ,in, is the DC voltage value equivalent to the original electrical signal on the resistor, is the period for calculating the effective value; pass , extract X clean (t) The corresponding phase difference ,in, is the active power corresponding to the original electrical signal at time t, is the reactive power corresponding to the original electrical signal at time t; Using the anti-electricity theft analysis device to set the effective value limit threshold , Phase difference limit threshold and box shape variable threshold , when the anti-theft analysis device detects The fluctuation range exceeds or phase difference Greater than When the cabinet protection system is activated; Obtaining box shape variables through strain sensors in the sensor network , if the box shape variable Greater than the box shape variable threshold , then the electromagnetic locking device is started and an encrypted control instruction is sent to the actuator; Among them, the box shape The calculation formula is: In the formula, is the deformation length, is the original length.

[0009] In an optional embodiment, step S4 specifically includes: Get the switch status signal S in the box protection system lock , and read the RSSI value and image feature points of the RFID electronic seal; The RSSI limit threshold and the seal abnormality judgment threshold are set by the anti-electricity theft analysis device. When the switch state signal is equal to 1 and the RSSI value of the RFID electronic seal is less than the RSSI limit threshold set by the anti-electricity theft analysis device, the Euclidean distance of the image feature point is obtained; If the Euclidean distance corresponding to the image feature point is greater than the seal abnormality judgment threshold, the seal is judged to be abnormal, and a seal abnormality signal is sent synchronously. When the anti-electricity theft analysis device receives the seal abnormality signal, the intrusion detection module is activated and the high-sensitivity mode is performed; The calculation formula of the Euclidean distance of the image feature points is: ; In the formula, d is the geometric distance between two feature points in space, is the i-th dimension coordinate of the seal feature point to be tested, is the i-th dimension coordinate of the feature point of the database template, and n is the number of dimensions of the feature point.

[0010] In an optional embodiment, step S5 specifically includes: After triggering the intrusion detection module, enter the high-sensitivity mode and adjust the confidence threshold of the YOLOv5 model: ,in, is the Euclidean distance of the image feature points, is the adjustment coefficient, is the default threshold, is the adjusted threshold; Through the fusion perception of infrared camera and millimeter wave radar, the motion trajectory (x(t), y(t)) of the intrusion target is extracted; According to the motion trajectory (x(t), y(t)) of the intrusion target, the moving speed v(t) of the intrusion target is obtained; The number of reference frames N for intrusion judgment and the moving speed limit threshold are set through the anti-electricity theft analysis device. If N consecutive frames satisfy that v(t) is greater than the moving speed limit threshold, the following is executed: S5.1. Send an alarm message to the target user's mobile terminal, the alarm message includes the intrusion target coordinates (x curr ,y curr ) and predicted path ; S5.2. Activate the sound and light deterrent device and remotely lock the meter communication port, and generate a protection log ID event =Hash(S lock , RSSI, d), and store the protection log in the blockchain; Wherein, the calculation formula of the moving speed v(t) is: ; In the formula, is the x-axis coordinate value of the intrusion target at time t in the image perceived by the infrared camera and millimeter-wave radar fusion, is the y-axis coordinate value of the intrusion target at time t in the image perceived by the infrared camera and millimeter-wave radar.

[0011] In an optional embodiment, the frequency band coefficients of the threshold function are determined by offline training and satisfy: in, is the peak energy of electromagnetic interference during the i-th training, is the total number of trainings.

[0012] Furthermore, an anti-electricity theft analysis device is proposed, which is used to implement the analysis method as described in any one of the above items, including: An electromagnetic interference processing unit, the electromagnetic interference processing unit comprising a wide-band sensor array and an adaptive notch filter, supporting dynamic generation of frequency suppression masks; A physical protection controller, wherein a strain sensor and an electromagnetic locking device are integrated therein, and is used to receive the signal output by the adaptive notch filter, and extract its effective value and phase difference, and then determine whether to activate the box protection system according to the effective value and phase difference; A seal verification module, which is equipped with an RFID reader and an image processor, is used to obtain the switch status signal in the box protection system, and read the RSSI value and image feature points of the RFID electronic seal, and then determine whether the seal is abnormal based on the RSSI value and image feature points; An intelligent disposal terminal equipped with a millimeter-wave radar and an edge computing chip for extracting the movement trajectory of the intrusion target; The intrusion detection module is used to generate a protection log according to the movement trajectory of the intrusion target and store the protection log in the blockchain.

[0013] In an optional embodiment, the electromagnetic interference processing unit communicates with the physical protection controller via a PCIe bus, with a transmission delay of <1ms, satisfying the real-time requirement of the signal output by the adaptive notch filter in step S2.

[0014] In an optional embodiment, the seal verification module has a built-in security chip, and the protection log generated in step S5 is encrypted by SM4, and the key update cycle is 24 hours.

[0015] Furthermore, an anti-electricity theft system is proposed, which is used to implement any of the above devices, and specifically includes: An interference prediction unit, wherein the interference prediction unit trains an interference prediction model based on a cloud platform using historical data of magnetic field energy density and outputs a threshold function for a future period; A secondary verification unit, after the mobile terminal receives the alarm information of step S5, the secondary verification unit can remotely trigger the RFID electronic seal in step S4 to perform a secondary verification instruction.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This scheme proposes an anti-electricity theft analysis method and device, which collects magnetic field strength signals in real time through a wide-band sensor array, synchronously obtains the fundamental frequency parameters of the power grid, calculates the magnetic field energy density and dynamically generates a frequency suppression mask, thereby accurately shielding the strong electromagnetic interference injected by the electricity theft device, and realizing real-time monitoring of the magnetic field strength of the power supply line and efficient filtering of interference signals. Dynamically matching the interference frequency band through the threshold function avoids measurement errors or failures caused by electromagnetic interference; This scheme proposes an anti-electricity theft analysis method and device, which uses a strain sensor to monitor the deformation of the box in real time, combines the Kalman filter algorithm to optimize the deformation calculation, and dynamically triggers the electromagnetic locking device. When the deformation of the box exceeds the preset threshold, the system immediately activates the protection mechanism and controls the actuator to lock the device through encrypted instructions. Through sub-millimeter deformation monitoring and real-time response, the impact resistance of the box is improved, effectively curbing the behavior of electricity thieves who steal electricity by destroying the box; This scheme proposes an anti-electricity theft analysis method and device, which dynamically adjusts the transmission power of the RFID reader and combines the Euclidean distance matching of image feature points to achieve real-time monitoring of the seal status and identification of tampering traces. The hash value of the seal status is stored on the blockchain to ensure that the tampering evidence chain cannot be tampered with, solving the technical pain points of the lag in traditional seal verification and the easy bypass of single-modal detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flow chart of an anti-electricity theft analysis method proposed by the present invention; Figure 2 A flow chart for dynamically generating a frequency suppression mask and an output signal for the adaptive notch filter in the present invention; Figure 3 This is a flow chart for determining whether a seal is abnormal in the present invention; Figure 4 A module framework diagram of an anti-electricity theft analysis device proposed by the present invention; Figure 5 This is a system framework diagram of an anti-electricity theft analysis system proposed in the present invention. DETAILED DESCRIPTION

[0018] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0019] Reference Figure 1 - Figure 5 As shown, an anti-electricity theft analysis method includes: S1. Obtain the environment information of the target user's metering equipment, and deploy the sensor network based on the environment information of the target user's metering equipment; S2. The wide-band electromagnetic sensor in the sensor network collects the magnetic field strength of the power supply line in real time, and synchronously obtains the fundamental frequency of the metering equipment when it is running, so as to determine the magnetic field energy density. Then, according to the magnetic field energy density, it is determined whether to trigger the adaptive notch filter. If the adaptive notch filter is triggered, the adaptive notch filter dynamically generates a frequency suppression mask and outputs a signal X. clean (t); S3. Receive the signal X output by the adaptive notch filter clean (t), and extract its effective value and phase difference, and then determine whether to activate the box protection system based on the effective value and phase difference; S4. Obtain the switch status signal in the box protection system, and read the RSSI value and image feature points of the RFID electronic seal, and then determine whether the seal is abnormal based on the RSSI value and image feature points. If the seal is abnormal, the intrusion detection module is triggered; S5. After the intrusion detection module is triggered, it enters the high-sensitivity mode, adjusts the confidence threshold of the YOLOv5 model, and extracts the motion trajectory of the intrusion target through the fusion perception of the infrared camera and millimeter-wave radar in the sensor network. Based on the motion trajectory of the intrusion target, a protection log is generated and stored in the blockchain.

[0020] Specifically, in step S1, the environmental information of the target user's metering equipment is obtained, and the sensor network is deployed based on the environmental information of the target user's metering equipment. The survey personnel survey the electromagnetic, vibration, temperature and other environmental parameters around the metering equipment through the system, and the relevant personnel deploy broadband electromagnetic sensors, vibration acceleration sensors and infrared thermal imaging modules to form a multimodal perception network, combined with LoRaWAN networking technology to achieve dynamic data collection, and form a real-time monitoring system covering the key areas of the box, so as to accurately identify magnetic field interference, abnormal deformation and intrusion behavior, and improve the environmental adaptability and detection reliability of the anti-theft device. The sensors in the sensor network are installed in the best position and transmit data and signals through transmission lines or wireless communications.

[0021] Furthermore, step S2 specifically includes: The magnetic field strength B(t) of the power supply line is collected in real time through a wide-band electromagnetic sensor array, and the fundamental frequency f0 of the metering equipment during operation is simultaneously obtained; The magnetic field strength of the power supply line is collected in real time using a wide-band electromagnetic sensor to obtain the magnetic field energy density E(t); When the magnetic field energy density E(t)>E threshold When (f0), the adaptive notch filter is triggered to synchronously acquire the original electrical signal, and then the original electrical signal is input into the adaptive notch filter for filtering; Get the frequency suppression mask M(f) dynamically generated by the adaptive notch filter and synchronously determine the output signal X of the adaptive notch filter clean (t), where E threshold (f0) is the threshold function; Among them, the calculation formula of magnetic field energy density E(t) is: Where E(t) is the magnetic field energy density of the power supply line at time t, is the magnetic field strength of the power supply line at time t, is the integral variable, representing the arrive Every moment in the time period, is the preset time window length; Threshold function E threshold The expression formula of (f0) is: In the formula, and are frequency band coefficients, It is the fundamental frequency of the measuring equipment when it is running; The output signal X of the adaptive notch filter clean The expression formula of (t) is: In the formula, is the signal after interference suppression output by the adaptive notch filter at time t, is the original electrical signal input into the adaptive notch filter at time t, is the number of interference frequency bands, is the frequency band index, For the The filtered value of the original electrical signal that needs to be filtered, For the A binary function of the filtered value of the original electrical signal that needs to be filtered.

[0022] Furthermore, step S3 specifically includes: Receive the adaptive notch filter output signal X clean (t), through , extract X clean (t) Corresponding effective value ,in, is the DC voltage value equivalent to the original electrical signal on the resistor, is the period for calculating the effective value; pass , extract X clean (t) The corresponding phase difference ,in, is the active power corresponding to the original electrical signal at time t, is the reactive power corresponding to the original electrical signal at time t; Using the anti-electricity theft analysis device to set the effective value limit threshold , Phase difference limit threshold and box shape variable threshold , when the anti-theft analysis device detects The fluctuation range exceeds or phase difference Greater than When the cabinet protection system is activated; Obtaining box shape variables through strain sensors in the sensor network , if the box shape variable Greater than the box shape variable threshold , then the electromagnetic locking device is started and an encrypted control instruction is sent to the actuator; Among them, the box shape variable The calculation formula is: In the formula, is the deformation length, is the original length.

[0023] Specifically, the original electrical signal is the source of data for the anti-theft system analysis, covering multi-dimensional information such as current, voltage, phase, frequency, harmonics, etc. Strain sensors (such as resistance strain gauges, optical fiber strain sensors) output electrical signals by measuring small deformations (stretching or compression) on the surface of objects. The core data is the deformation variable. The original data (deformation variable) required by the strain sensor to obtain the box deformation variable needs to be filtered by Kalman filtering, and its state equation is: In the formula, is the process noise when processing the deformation at time k, is the observation noise when processing the deformation at time k, is the final deformation at time k, It is the original deformation directly measured by the strain sensor.

[0024] Furthermore, step S4 specifically includes: Get the switch status signal S in the box protection system lock , and read the RSSI value and image feature points of the RFID electronic seal; The RSSI limit threshold and the seal abnormality judgment threshold are set by the anti-electricity theft analysis device. When the switch state signal is equal to 1 and the RSSI value of the RFID electronic seal is less than the RSSI limit threshold set by the anti-electricity theft analysis device, the Euclidean distance of the image feature point is obtained; If the Euclidean distance corresponding to the image feature point is greater than the seal abnormality judgment threshold, the seal is judged to be abnormal, and a seal abnormality signal is sent synchronously. When the anti-electricity theft analysis device receives the seal abnormality signal, the intrusion detection module is activated and the high-sensitivity mode is performed; Among them, the calculation formula of the Euclidean distance of image feature points is: In the formula, is the geometric distance between two feature points in space, is the i-th dimension coordinate of the seal feature point to be tested, is the i-th dimension coordinate of the feature point of the database template, and n is the number of dimensions of the feature point.

[0025] Specifically, the database template feature points are the original images of the metering equipment and electronic seals (such as QR codes). Different metering equipment and electronic seals will have different templates, which are stored in the database. The RFID reader adopts a dynamic power control strategy, and the transmission power P RF satisfy: in, 30dBm, 20dBm, It is 10DBm.

[0026] Understandably, Indicates the minimum received signal strength at which the RFID reader can maintain reliable communication. When the reader detects that the RSSI returned by the tag is lower than this threshold, the communication quality is considered unacceptable (for example, the signal is too weak or the tag is too far away). is between and the maximum RSSI value When the intermediate intensity threshold When , the reader may maintain or slightly adjust the transmit power (for example, gradually reduce the power to save energy). When the reader is connected to a wireless network, the reader can further reduce the transmission power to reduce interference and power consumption.

[0027] Furthermore, step S5 specifically includes: After triggering the intrusion detection module, enter the high-sensitivity mode and adjust the confidence threshold of the YOLOv5 model: ,in, is the Euclidean distance of the image feature points, is the adjustment coefficient, is the default threshold, is the adjusted threshold; Through the fusion perception of infrared camera and millimeter wave radar, the motion trajectory (x(t), y(t)) of the intrusion target is extracted; According to the motion trajectory of the intrusion target (x(t), y(t)), obtain the moving speed v(t) of the intrusion target; The number of reference frames N for intrusion judgment and the moving speed limit threshold are set through the anti-electricity theft analysis device. If N consecutive frames satisfy that v(t) is greater than the moving speed limit threshold, the following is executed: S5.1. Send an alarm message to the target user's mobile terminal, including the intrusion target coordinates (x curr ,y curr ) and predicted path ; S5.2. Activate the sound and light deterrent device and remotely lock the meter communication port, and generate a protection log ID event =Hash(S lock , RSSI, d), and store the protection log in the blockchain; Wherein, the calculation formula of the moving speed v(t) is: ; In the formula, is the x-axis coordinate value of the intrusion target at time t in the image perceived by the infrared camera and millimeter-wave radar fusion, is the y-axis coordinate value of the intrusion target at time t in the image perceived by the infrared camera and millimeter-wave radar.

[0028] Specifically, the intrusion target coordinates (x curr ,y curr ) and predicted path Using the Kalman filter recursive formula: ; in, is the predicted coordinate value of the intrusion target at time k, is the state transfer matrix at time k, is the Kalman gain at time k, is the observation matrix at time k, is the compensation signal at time k (such as the temperature sensor reading), is the control input matrix, used to associate external control signals (such as temperature compensation), To dynamically adjust the weights of predicted and measured values, It is the original coordinate directly collected by the sensor.

[0029] It is understandable that the infrared camera and millimeter wave radar fusion perception adopts a weighted fusion strategy:

[0030] in, is the data after the infrared camera and millimeter-wave radar are fused and sensed at time t, is the data collected by the infrared camera at time t, is the data collected by the millimeter-wave radar at time t, is the temperature weight; The temperature weight is dynamically adjusted according to the ambient temperature. When the temperature at time t is less than 25 degrees, the temperature weight is 0.7. When the temperature at time t is greater than or equal to 25 degrees, the temperature weight is 0.4.

[0031] Furthermore, the frequency band coefficients of the threshold function are determined by offline training, satisfying: ;in, is the peak energy of electromagnetic interference in the i-th training, is the total number of trainings.

[0032] Furthermore, an anti-electricity theft analysis device is proposed, which is used to implement any of the above analysis methods, including: Electromagnetic Interference Processing Unit, which includes a wide-band sensor array and an adaptive notch filter to support dynamic generation of frequency suppression masks; Physical protection controller: The physical protection controller integrates a strain sensor and an electromagnetic locking device to receive the signal output by the adaptive notch filter and extract its effective value and phase difference. Then, based on the effective value and phase difference, it determines whether to activate the box protection system. Seal verification module: The seal verification module is equipped with an RFID reader and an image processor, which is used to obtain the switch status signal in the box protection system and read the RSSI value and image feature points of the RFID electronic seal, and then determine whether the seal is abnormal based on the RSSI value and image feature points; Intelligent disposal terminal, which is equipped with millimeter-wave radar and edge computing chip to extract the movement trajectory of the intruder target; Intrusion detection module,The intrusion detection module is used to generate protection logs according to the movement trajectory of the intrusion target, and store the protection logs in the blockchain.

[0033] Furthermore, the electromagnetic interference processing unit communicates with the physical protection controller via the PCIe bus, and the transmission delay is less than 1ms, which meets the real-time requirement of the signal output by the adaptive notch filter in step S2.

[0034] Furthermore, the seal verification module has a built-in security chip, which performs SM4 encryption on the protection log generated in step S5, and the key update cycle is 24 hours.

[0035] Furthermore, an anti-electricity theft system is proposed, which is used to implement any of the above devices, and specifically includes: An interference prediction unit, which trains an interference prediction model based on the cloud platform using historical data of magnetic field energy density and outputs a threshold function for a future period; The secondary verification unit can remotely trigger the RFID electronic seal in step S4 after the mobile terminal receives the alarm information in step S5 to perform a secondary verification instruction.

[0036] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for analyzing anti-electricity theft, characterized in that: include: S1. Obtain the target user's metering equipment environment information, and deploy the sensor network based on the target user's metering equipment environment information; S2. The wide-band electromagnetic sensor in the sensor network collects the magnetic field strength of the power supply line in real time, and synchronously obtains the fundamental frequency of the metering equipment when it is running, so as to determine the magnetic field energy density. Then, according to the magnetic field energy density, it is determined whether to trigger the adaptive notch filter. If the adaptive notch filter is triggered, the adaptive notch filter dynamically generates a frequency suppression mask and outputs a signal X. clean (t); S3. Receive the signal X output by the adaptive notch filter clean (t), and extract its effective value and phase difference, and then determine whether to activate the box protection system based on the effective value and phase difference; S4. Obtain the switch status signal in the box protection system, and read the RSSI value and image feature points of the RFID electronic seal, and then determine whether the seal is abnormal based on the RSSI value and image feature points. If the seal is abnormal, the intrusion detection module is triggered; S5. After the intrusion detection module is triggered, it enters the high-sensitivity mode, adjusts the confidence threshold of the YOLOv5 model, and extracts the motion trajectory of the intrusion target through the fusion perception of the infrared camera and millimeter-wave radar in the sensor network. According to the motion trajectory of the intrusion target, a protection log is generated and stored in the blockchain.

2. The anti-electricity theft analysis method according to claim 1, characterized in that: Step S2 specifically includes: The magnetic field strength B(t) of the power supply line is collected in real time through a wide-band electromagnetic sensor array, and the fundamental frequency f0 of the metering equipment during operation is simultaneously obtained; The magnetic field strength of the power supply line is collected in real time using a wide-band electromagnetic sensor to obtain the magnetic field energy density E(t); When the magnetic field energy density E(t)>E threshold When (f0), the adaptive notch filter is triggered to synchronously acquire the original electrical signal, and then the original electrical signal is input into the adaptive notch filter for filtering; Get the frequency suppression mask M(f) dynamically generated by the adaptive notch filter and synchronously determine the output signal X of the adaptive notch filter clean (t), where E threshold (f0) is the threshold function; Wherein, the calculation formula of the magnetic field energy density E(t) is: Where E(t) is the magnetic field energy density of the power supply line at time t, is the magnetic field strength of the power supply line at time t, is the integral variable, representing the arrive Every moment in the time period, is the preset time window length; The threshold function E threshold The expression formula of (f0) is: In the formula, and are frequency band coefficients, It is the fundamental frequency of the measuring equipment when it is running; The output signal X of the adaptive notch filter clean The expression formula of (t) is: In the formula, is the signal after interference suppression output by the adaptive notch filter at time t, is the original electrical signal input into the adaptive notch filter at time t, is the number of interference frequency bands, is the frequency band index, For the The filtered value of the original electrical signal that needs to be filtered, For the A binary function of the filtered value of the original electrical signal that needs to be filtered.

3. The anti-electricity theft analysis method according to claim 1, characterized in that: Step S3 specifically includes: Receive the adaptive notch filter output signal X clean (t), through , extract X clean (t) Corresponding effective value ,in, is the DC voltage value equivalent to the original electrical signal on the resistor, is the period for calculating the effective value; pass , extract X clean (t) The corresponding phase difference ,in, is the active power corresponding to the original electrical signal at time t, is the reactive power corresponding to the original electrical signal at time t; Using the anti-electricity theft analysis device to set the effective value limit threshold , Phase difference limit threshold and box shape variable threshold , when the anti-theft analysis device detects The fluctuation range exceeds or phase difference Greater than When the cabinet protection system is activated; Obtaining box shape variables through strain sensors in the sensor network , if the box shape variable Greater than the box shape variable threshold , then the electromagnetic locking device is started and an encrypted control instruction is sent to the actuator; Among them, the box shape The calculation formula is: In the formula, is the deformation length, is the original length.

4. The anti-electricity theft analysis method according to claim 1, characterized in that: Step S4 specifically includes: Get the switch status signal S in the box protection system lock , and read the RSSI value and image feature points of the RFID electronic seal; The RSSI limit threshold and the seal abnormality judgment threshold are set by the anti-electricity theft analysis device. When the switch state signal is equal to 1 and the RSSI value of the RFID electronic seal is less than the RSSI limit threshold set by the anti-electricity theft analysis device, the Euclidean distance of the image feature point is obtained; If the Euclidean distance corresponding to the image feature point is greater than the seal abnormality judgment threshold, the seal is judged to be abnormal, and a seal abnormality signal is sent synchronously. When the anti-electricity theft analysis device receives the seal abnormality signal, the intrusion detection module is activated and the high-sensitivity mode is performed; The calculation formula of the Euclidean distance of the image feature points is: In the formula, is the geometric distance between two feature points in space, is the i-th dimension coordinate of the seal feature point to be tested, is the i-th dimension coordinate of the feature point of the database template, and n is the number of dimensions of the feature point.

5. The anti-electricity theft analysis method according to claim 1, characterized in that: Step S5 specifically includes: After triggering the intrusion detection module, enter the high-sensitivity mode and adjust the confidence threshold of the YOLOv5 model: ,in, is the Euclidean distance of the image feature points, is the adjustment coefficient, is the default threshold, is the adjusted threshold; Through the fusion perception of infrared camera and millimeter wave radar, the motion trajectory (x(t), y(t)) of the intrusion target is extracted; According to the motion trajectory (x(t), y(t)) of the intrusion target, the moving speed v(t) of the intrusion target is obtained; The number of reference frames N for intrusion judgment and the moving speed limit threshold are set through the anti-electricity theft analysis device. If N consecutive frames satisfy that v(t) is greater than the moving speed limit threshold, the following is executed: S5.

1. Send an alarm message to the target user's mobile terminal, the alarm message includes the intrusion target coordinates (x curr ,y curr ) and predicted path ; S5.

2. Activate the sound and light deterrent device and remotely lock the meter communication port, and generate a protection log ID event =Hash(S lock , RSSI, d), and store the protection log in the blockchain; Wherein, the calculation formula of the moving speed v(t) is: ； In the formula, is the x-axis coordinate value of the intrusion target at time t in the image perceived by the infrared camera and millimeter-wave radar fusion, is the y-axis coordinate value of the intrusion target at time t in the image perceived by the infrared camera and millimeter-wave radar.

6. The anti-electricity theft analysis method according to claim 2, characterized in that: The frequency band coefficients of the threshold function are determined by offline training and satisfy: in, is the peak energy of electromagnetic interference during the i-th training, is the total number of trainings.

7. An anti-electricity theft analysis device, used to implement the analysis method according to any one of claims 1 to 6, characterized in that: include: An electromagnetic interference processing unit, the electromagnetic interference processing unit comprising a wide-band sensor array and an adaptive notch filter, supporting dynamic generation of frequency suppression masks; A physical protection controller, wherein a strain sensor and an electromagnetic locking device are integrated therein, and is used to receive the signal output by the adaptive notch filter, and extract its effective value and phase difference, and then determine whether to activate the box protection system according to the effective value and phase difference; A seal verification module, which is equipped with an RFID reader and an image processor, is used to obtain the switch status signal in the box protection system, and read the RSSI value and image feature points of the RFID electronic seal, and then determine whether the seal is abnormal based on the RSSI value and image feature points; An intelligent disposal terminal equipped with a millimeter-wave radar and an edge computing chip for extracting the movement trajectory of the intrusion target; The intrusion detection module is used to generate a protection log according to the movement trajectory of the intrusion target and store the protection log in the blockchain.

8. The anti-electricity theft analysis device according to claim 7, characterized in that: The electromagnetic interference processing unit communicates with the physical protection controller via the PCIe bus, with a transmission delay of <1ms, satisfying the real-time requirement of the signal output by the adaptive notch filter in step S2.

9. The anti-electricity theft analysis device according to claim 7, characterized in that: The seal verification module has a built-in security chip, and performs SM4 encryption on the protection log generated in step S5, and the key update cycle is 24 hours.

10. An anti-electricity theft system, used to implement the device according to any one of claims 7 to 9, characterized in that: Specifically include: An interference prediction unit, wherein the interference prediction unit trains an interference prediction model based on a cloud platform using historical data of magnetic field energy density and outputs a threshold function for a future period; A secondary verification unit, after the mobile terminal receives the alarm information of step S5, the secondary verification unit can remotely trigger the RFID electronic seal in step S4 to perform a secondary verification instruction.

Images