Remote millimeter wave distance measurement and AI identification system and method suitable for power transmission line

Abstract

The invention discloses a remote millimeter wave distance measurement and AI identification system and method suitable for a power transmission line. The system comprises an acquisition module for acquiring an external damage target position around a power transmission line and relative distance data between the external damage target position and a wire in real time; the AI intelligent identification module is used for carrying out feature identification on the acquired radar echo data by utilizing a built-in AI algorithm, and judging the hidden danger type and the minimum gap relation between the hidden danger and a wire; and the communication transmission module is used for uploading the identification result and the positioning data of the AI intelligent identification module to a monitoring platform so as to realize remote hidden danger early warning, image display and operation and maintenance task distribution. According to the invention, accurate identification and positioning of external damage hidden dangers can be realized, and the online monitoring efficiency and the operation and maintenance response capability of the high-voltage transmission line are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of power system safety operation and maintenance and intelligent sensing technology, and in particular to a remote millimeter-wave ranging and AI recognition system and method applicable to transmission lines. Background Technology

[0002] With the acceleration of urbanization and the continuous expansion of the power grid, overhead transmission lines are increasingly traversing mountainous areas, urban-rural fringe areas, and densely populated construction zones—regions with complex terrain and open environments. These areas are constantly occupied by various construction projects, material transport, and machinery operations, which can easily lead to external objects approaching the transmission lines. If not detected and dealt with in a timely manner, this could result in insufficient electrical clearance, tripping faults, or even power accidents, posing a serious threat to power transmission safety.

[0003] Currently, the main methods for monitoring potential external damage to power transmission lines include manual inspection, video surveillance, and drone inspection. Manual inspection has significant drawbacks such as slow response, high labor intensity, and low efficiency, and its difficulty increases significantly in extreme weather conditions such as high temperatures, heavy rain, and cold. While video surveillance can provide real-time images, its accuracy is greatly reduced at night and in foggy or hazy environments due to limitations such as installation angle, obstructions, and lighting conditions. Drone inspection offers flexibility and a degree of automation, but its application in high-frequency, continuous, and remote areas remains challenging due to limitations in flight time, weather conditions, and management regulations.

[0004] Furthermore, while some millimeter-wave radars used in special scenarios possess excellent target ranging capabilities, they mostly only acquire distance information and cannot be integrated with spatial positioning technologies, nor do they have intelligent identification or remote linkage capabilities. In power field environments with multi-target interference, complex obstacles, and strong background noise, traditional radar systems struggle to effectively distinguish and analyze the acquired echo signals. Especially in complex tasks requiring accurate identification of hazard types and real-time labeling of hazard locations and conductor gaps, single ranging or monitoring methods are no longer sufficient to meet the current technical demands of intelligent power grid operation and maintenance.

[0005] Therefore, in view of the above-mentioned problems and limitations, this invention proposes a remote millimeter-wave ranging and AI identification system suitable for power transmission lines. By integrating millimeter-wave radar, BeiDou high-precision positioning, AI intelligent identification, remote communication transmission, and a low-power power supply module, an integrated remote hazard monitoring terminal is constructed, solving the problems of low identification accuracy, lack of spatial positioning, complex deployment, and weak real-time response capabilities in existing technologies. This system can be deployed along power transmission lines, acquiring distance information between the external damage target and the conductor through millimeter-wave radar, and simultaneously acquiring the spatial position of the target using a BeiDou module to achieve three-dimensional spatial annotation. The system has a built-in AI identification processing unit that can analyze the echo signal in real time, automatically identify the type of external damage and determine its minimum gap with the conductor. The identification results are uploaded to a remote monitoring platform through a communication module, thereby achieving intelligent identification and remote early warning of external damage hazards to power transmission lines. Summary of the Invention

[0006] The purpose of this invention is to provide a remote millimeter-wave ranging and AI recognition system suitable for power transmission lines, which can not only accurately identify and locate potential external damage hazards, but also significantly improve the online monitoring efficiency and operation and maintenance response capability of high-voltage power transmission lines.

[0007] Another objective of this invention is to provide a remote millimeter-wave ranging and AI recognition method suitable for power transmission lines.

[0008] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0009] This invention provides a remote millimeter-wave ranging and AI recognition system suitable for power transmission lines, comprising:

[0010] The data acquisition module acquires real-time data on the location of external damage targets around the transmission line and their relative distance from the conductor;

[0011] The AI ​​intelligent recognition module uses built-in AI algorithms to identify features in the collected radar echo data and determine the type of hidden danger and its relationship with the minimum gap between the conductor and the conductor.

[0012] The communication transmission module uploads the recognition results and positioning data of the AI ​​intelligent recognition module to the monitoring platform to realize remote hazard warning, image display and operation and maintenance task dispatch;

[0013] The acquisition module, communication transmission module, solar panel, and lithium battery are integrated and packaged into a single unit to form a lightweight intelligent monitoring terminal that can be remotely deployed.

[0014] Alternatively, the acquisition module is an integrated structure of a millimeter-wave radar module and a BeiDou positioning module, with the millimeter-wave radar module and the BeiDou positioning module using a unified control unit for data acquisition and time synchronization.

[0015] Optionally, in the AI ​​intelligent identification module, the AI ​​algorithm adopts a convolutional neural network (CNN) structure and combines a multi-channel feature fusion mechanism to perform time-frequency domain analysis on radar echo signals, achieving low-latency hazard identification at the edge terminal; the identification content includes typical external damage hazard types, and at the same time, it combines distance to determine whether it constitutes an interference risk.

[0016] Optionally, the communication transmission module supports 4G / 5G cellular networks or LoRa low-power wide-area communication protocols, and has the functions of data breakpoint resume and intelligent compression. After the identification results are uploaded to the monitoring platform, they are marked on the map through the GIS module, and the background early warning mechanism is linked to generate risk reports or push operation and maintenance scheduling tasks.

[0017] A remote millimeter-wave ranging and AI recognition method suitable for power transmission lines includes the following steps:

[0018] S1: Deploy data acquisition devices integrating millimeter-wave radar and BeiDou positioning around the power transmission line to acquire real-time data on the location of external damage targets and their relative distance to the conductor.

[0019] S2: Use AI algorithms to identify features in the collected radar echo data to determine the type of hidden danger and its relationship with the minimum gap between the conductor and the target.

[0020] S3: Upload the identification results and positioning data after feature recognition to the monitoring platform to realize remote hazard early warning, image display and operation and maintenance task dispatch.

[0021] Optionally, in step S1, the millimeter-wave radar adopts the FMCW (Frequency Modulated Continuous Wave) system, operates at a frequency of 77 GHz, has a horizontal scanning angle of ±60° and a vertical elevation angle of ±20°, a field of view coverage radius of 100 meters, and scans periodically with a default frame rate of 5 Hz; when an abnormally strong echo target is detected, it automatically upgrades to a 20 Hz high-frequency scanning mode.

[0022] The BeiDou positioning system uses a dual-frequency high-precision receiver and supports real-time differential positioning (RTK) function. The polar coordinate information of the target point collected by the millimeter-wave radar will be combined with the radar installation point coordinates obtained by BeiDou to calculate the three-dimensional coordinates of the target.

[0023] The specific spatial coordinates of the target are calculated using the following formula:

[0024]

[0025] Where: X t ,Y t Z tThe target's three-dimensional spatial position in the geographic coordinate system; r is the target distance measured by the millimeter-wave radar; θ is the horizontal deflection angle; φ is the vertical elevation angle; X0, Y0, Z0 are the radar installation point coordinates, provided by BeiDou.

[0026] Alternatively, step S2 may specifically include the following steps:

[0027] (1) Echo signal preprocessing: The original signal output by the millimeter-wave radar is a time-varying echo sequence, and its mathematical expression is as follows:

[0028] S(t)=A(t)·cos(2πf0t+φ(t)) (2)

[0029] Where S(t) is the echo signal strength; A(t) is the envelope function, representing the change of echo amplitude over time; f0 is the radar carrier frequency; and φ(t) is the signal phase, which includes range and Doppler information.

[0030] The original signal is analyzed in time and frequency using Short Time Fourier Transform (STFT) to obtain its time-frequency spectrum, which serves as the input for AI recognition.

[0031]

[0032] Where X(t,ω) is the complex amplitude of the echo signal at time t and frequency ω; S(τ) is the original time series signal, i.e., the echo signal to be analyzed; ω(τ-t) is the window function, usually a Hamming window or a Gaussian window; e -jwτ τ is the Fourier kernel function, which projects the local signal into the frequency domain; τ is the integral variable, representing the sliding variable over time.

[0033] (2) Target feature recognition and classification: A lightweight convolutional neural network structure is used to extract features from the echo spectrum and classify targets, outputting the probability distribution of target categories:

[0034]

[0035] Where P is the probability vector for each category; p i Let be the probability value for identifying it as the i-th type of target; n is the total number of categories;

[0036] Final output target category:

[0037]

[0038] (3) Conductor spatial modeling and minimum gap calculation: Based on the actual transmission line structure, the conductor's shape between towers conforms to a parabolic model. Under no-wind static load conditions, its spatial sag is expressed as:

[0039]

[0040] Where Z(s) is the height of the conductor at a lateral distance s; H is the height of the conductor suspension point; w is the weight per unit length of conductor; T is the conductor tension; and s is the lateral distance from the tower.

[0041] By fitting a three-dimensional spatial curve, the spatial trajectory of the conductor is obtained as C(s) = (X(s), Y(s), Z(s)); for a certain target to be identified, its coordinates are (X... t ,Y t Z t If ), then the minimum spatial distance between it and the conductor is:

[0042]

[0043] Among them, D min To identify the minimum gap distance between the target and the traverse; s1, s2 correspond to the spatial parameter range of the start and end points of this traverse segment; (X(s), Y(s), Z(s)) are the spatial function points of the traverse; (X t ,Y t Z t () represents the three-dimensional coordinates of the target identified by AI.

[0044] The present invention has the following beneficial effects:

[0045] This invention overcomes the technical bottlenecks of traditional methods relying on manual inspections or video recognition, which are easily limited in complex environments, suffer from slow response times, and have low accuracy in identifying potential hazards. It innovatively integrates high-precision ranging with millimeter-wave radar, BeiDou positioning, and AI intelligent recognition algorithms to construct a lightweight, integrated, and remote intelligent monitoring terminal. This system can achieve all-weather, fully automated monitoring of external damage targets and minimum clearance measurement around transmission line corridors. Through a communication module, it uploads the identification results and spatial information to the monitoring platform in real time, enabling coordinated hazard alarms, layer display, and maintenance task assignment. The system is highly adaptable and easy to deploy, particularly suitable for high-risk sections of high-voltage overhead lines in mountainous, forested, and farmland areas. It effectively improves the accuracy, timeliness, and intelligent level of external damage hazard identification for transmission lines, demonstrating significant engineering practical value and promising prospects for widespread application. Attached Figure Description

[0046] Figure 1 This is a system structure block diagram of the present invention;

[0047] Figure 2 This is a flowchart of the method of the present invention;

[0048] Figure 3 This is a schematic diagram of the dual-trigger fusion mechanism of radar and AI recognition in this invention. Detailed Implementation

[0049] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0050] Example 1

[0051] like Figure 1 As shown, this invention provides a remote millimeter-wave ranging and AI recognition system suitable for power transmission lines, comprising:

[0052] The acquisition module acquires real-time data on the location of external damage targets around the transmission line and their relative distance to the conductor. The acquisition module is an integrated structure of a millimeter-wave radar module and a Beidou positioning module. The millimeter-wave radar module and the Beidou positioning module use a unified control unit for data acquisition and time synchronization.

[0053] The AI ​​intelligent identification module uses a built-in AI algorithm to perform feature identification on the collected radar echo data, determine the type of hidden danger and its minimum gap relationship with the conductor; in the AI ​​intelligent identification module, the AI ​​algorithm adopts a convolutional neural network (CNN) structure and combines a multi-channel feature fusion mechanism to perform time-frequency domain analysis on the radar echo signal, realizing low-latency hidden danger identification at the edge terminal; the identification content includes typical external damage hidden danger types such as construction machinery, stacked objects, and hoisted objects, and at the same time, it combines distance to determine whether it constitutes an interference risk.

[0054] The communication transmission module uploads the recognition results and location data of the AI ​​intelligent recognition module to the monitoring platform to realize remote hazard warning, image display and operation and maintenance task dispatch. The communication transmission module supports 4G / 5G cellular network or LoRa low power wide area communication protocol, and has the functions of data breakpoint resume and intelligent compression. After the recognition results are uploaded to the monitoring platform, they are marked on the map through the GIS module, and linked with the background early warning mechanism to generate risk reports or push operation and maintenance scheduling tasks.

[0055] The acquisition module, communication transmission module, solar panel, and lithium battery are integrated and packaged into a single unit to form a lightweight intelligent monitoring terminal that can be remotely deployed.

[0056] To ensure long-term stable operation of the equipment in the field environment of power transmission lines, the millimeter-wave radar module, Beidou positioning module, AI intelligent recognition module and communication transmission module are integrated and packaged in a unified terminal housing, and equipped with a dual-mode power supply system consisting of solar panels and lithium batteries, forming a remote monitoring terminal device with independent operation capabilities.

[0057] The terminal casing is made of high-strength aluminum alloy with a fluorocarbon anti-corrosion coating, meeting outdoor environmental requirements for UV resistance, waterproofing, dustproofing, and bird protection, with a protection rating of no less than IP65. The overall dimensions are controlled within 280mm × 200mm × 130mm, and the total weight does not exceed 3kg, facilitating single-person on-site installation. The casing features quick-connect fittings and replaceable protective covers for easy debugging and maintenance.

[0058] The integrated terminal shell adopts a high temperature resistant, UV resistant, waterproof and dustproof structural design, which meets the environmental adaptation standards for long-term operation of outdoor power equipment; the solar power supply module has maximum power point tracking (MPPT) function, which can continuously provide stable power output in low light or partially shaded environments, and the lithium battery capacity is not less than 30Wh, ensuring independent operation time of no less than 7 days in continuous cloudy and rainy environments.

[0059] The dual-mode power supply system consists of a 20W monocrystalline silicon solar panel and a 40Wh lithium battery, connected via an energy management module. The energy management module embeds a maximum power point tracking (MPPT) algorithm chip to improve the charging efficiency of the solar panel in low-light conditions. The system supports the following three power supply modes:

[0060] (1) Solar priority power supply mode: Direct power supply and battery charging during periods of strong sunlight;

[0061] (2) Lithium battery power supply mode: Automatically switches when there is insufficient sunlight or at night;

[0062] (3) Power-limited energy-saving mode: During continuous rainy days, it enters a low-frequency scanning and low-speed uploading strategy to extend the battery life.

[0063] The communication transmission module adopts a general serial communication architecture (RS-485+UART), supports pluggable communication modules, and comes standard with a 4G wireless data transmission card (supporting all network types). A LoRa communication module can also be selected for low-bandwidth, long-distance scenarios. The system supports breakpoint resume and caching mechanisms, with a maximum caching capacity of no less than 500 original data frames.

[0064] The integrated terminal offers flexible installation options, allowing it to be mounted on transmission tower crossarms, guy wires, and side railings of passageways using the included clamp components. Anchoring methods include U-bolt fixing, ground anchor brackets, or magnetic bases. Installation takes no more than 15 minutes, requires no special tools, and facilitates rapid deployment and replacement.

[0065] After power-on, the system will automatically perform a self-test and start up, and display the operating status of modules such as power supply, communication, and AI recognition through status indicator lights. The equipment has multiple electrical safety strategies, including reverse connection protection, abnormal voltage protection, and battery overcharge and over-discharge protection, to ensure reliable operation in long-term unattended environments.

[0066] The following table lists examples of key terminal parameter configurations:

[0067]

[0068] Through the above structural design and power supply strategy, the monitoring terminal is ensured to have the ability to be easily deployed, self-sustaining in terms of energy, and capable of long-term unattended operation, providing a stable operating foundation for subsequent AI identification and data feedback.

[0069] Example 2

[0070] like Figure 2 As shown, this invention provides a remote millimeter-wave ranging and AI recognition method suitable for power transmission lines, comprising:

[0071] S1: Deploy data acquisition devices integrating millimeter-wave radar and BeiDou positioning around the power transmission line to obtain the location of external damage targets and their relative distance to the conductor in real time.

[0072] The data acquisition device consists of a millimeter-wave radar module and a BeiDou high-precision positioning module, which work together using a unified embedded control chip. It is fixedly installed below transmission towers, at the edge of conductor corridors, or under crossing channels. The installation location is optimized through terrain analysis to ensure that the maximum field of view covers the sensitive areas under the transmission conductors.

[0073] The millimeter-wave radar module and the Beidou positioning module are integrated into a single structure. They use a unified control unit for data acquisition and time synchronization, which can bind the target's distance information and spatial coordinates into a structured measurement result. It also has high-frequency scanning capability and can cover a range of no less than 100 meters on both sides of the power transmission line channel.

[0074] The millimeter-wave radar employs the FMCW (Frequency Modulated Continuous Wave) system, operates in the 77GHz frequency band, and features a horizontal scan angle of ±60° and a vertical elevation angle of ±20°. Its field-of-view coverage radius reaches 100 meters, and it possesses strong anti-interference and fog-penetration capabilities. The radar scans periodically with a default frame rate of 5Hz; upon detecting an abnormally strong echo target, it automatically upgrades to a 20Hz high-frequency scanning mode.

[0075] The BeiDou module uses a dual-frequency high-precision receiver and supports real-time differential positioning (RTK) with a positioning accuracy better than ±0.2 meters. The polar coordinate information (distance r, horizontal angle θ, elevation angle φ) of the target point acquired by the millimeter-wave radar will be combined with the radar installation point coordinates (X0, Y0, Z0) obtained by BeiDou to calculate the target's three-dimensional coordinates.

[0076] The specific spatial coordinates of the target are calculated using the following formula:

[0077]

[0078] Where: X t ,Y t Z t The target's three-dimensional spatial position in the geographic coordinate system; r is the target distance measured by the millimeter-wave radar; θ is the horizontal deflection angle (relative to the radar's front); φ is the vertical elevation angle; X0, Y0, Z0 are the radar installation point coordinates, provided by BeiDou.

[0079] To ensure consistency of spatiotemporal data, the system appends a timestamp (UTC) and a synchronization frame number during data acquisition. The data record structure formed for each scan is shown in the table below:

[0080]

[0081]

[0082] Through the above mechanism, the system can complete the ranging, positioning and three-dimensional coordinate calculation of a target point within milliseconds, and provide accurate data support for subsequent AI recognition and spatial gap analysis.

[0083] S2: Use AI algorithms to identify features in the collected radar echo data to determine the type of hidden danger and its relationship with the minimum gap between the conductor and the target.

[0084] The AI ​​intelligent recognition module of this invention is used to classify and identify target echo signals acquired by millimeter-wave radar, and calculates the minimum safe gap between the externally damaged target and the conductor by combining spatial modeling of the transmission line conductor, so as to realize intelligent identification and dynamic early warning of hidden dangers in high-voltage transmission lines. This step includes the following key processes:

[0085] (1) Echo signal preprocessing: The original signal output by the millimeter-wave radar module is a time-varying echo sequence, and its mathematical expression is as follows:

[0086] S(t)=A(t)·cos(2πf0t+φ(t)) (2)

[0087] Where S(t) is the echo signal strength; A(t) is the envelope function, representing the change of echo amplitude over time; f0 is the radar carrier frequency (77GHz in this invention); and φ(t) is the signal phase, which includes range and Doppler information.

[0088] The original signal is analyzed in time and frequency using Short Time Fourier Transform (STFT) to obtain its time-frequency spectrum, which serves as the input for AI recognition.

[0089]

[0090] Where X(t,ω) is the complex amplitude of the echo signal at time t and frequency ω; S(τ) is the original time series signal, i.e., the echo signal to be analyzed; ω(τ-t) is the window function, usually a Hamming window or a Gaussian window; e -jwτ τ is the Fourier kernel function, which projects the local signal into the frequency domain; τ is the integral variable, representing the sliding variable over time (a certain moment on the time axis).

[0091] (2) Target feature recognition and classification: The AI ​​intelligent recognition module adopts a lightweight convolutional neural network structure to extract features from the echo spectrum and classify targets, outputting the probability distribution of target categories:

[0092]

[0093] Where P is the probability vector for each category; p i This represents the probability value for identifying a target as the i-th type; n is the total number of categories, for example: i = 0: empty field / no target; i = 1: construction machinery; i = 2: soil pile; i = 3: hoisting equipment; i = 4: personnel activities.

[0094] Final output target category:

[0095]

[0096] (3) Conductor spatial modeling and minimum gap calculation: Based on the actual transmission line structure, the shape of the conductor between towers conforms to a parabolic model. Under no-wind static load conditions, its spatial sag can be expressed as:

[0097]

[0098] Where Z(s) is the height of the conductor at a lateral distance s; H is the height of the conductor suspension point (tower); w is the weight per unit length of conductor (approximately 1.2 N / m); T is the conductor tension (determined according to the voltage level); and s is the lateral distance from the tower.

[0099] The spatial trajectory of the conductor, C(s) = (X(s), Y(s), Z(s)), is obtained through three-dimensional spatial curve fitting. For a given target, its coordinates are (X... t ,Y t Z t If ), then the minimum spatial distance between it and the conductor is:

[0100]

[0101] Among them, D min To identify the minimum gap distance between the target and the traverse; s1, s2 correspond to the spatial parameter range of the start and end points of this traverse segment; (X(s), Y(s), Z(s)) are the spatial function points of the traverse; (Xt ,Y t Z t () represents the three-dimensional coordinates of the target identified by AI.

[0102] To reduce false alarm rates, the AI ​​recognition module and radar module employ a dual-trigger fusion mechanism: if radar ranging detects an intrusion, AI recognition is activated for image verification; conversely, if no intrusion is detected, AI first identifies the intrusion and then activates radar precision measurement. Only when both simultaneously determine a potential hazard will alarm data be uploaded. This strategy effectively improves the system's robustness in complex environments such as forests and mountains.

[0103] like Figure 3 As shown, the equipment is fixed to the tower, and the line channel is comprehensively detected by 4D millimeter-wave radar and AI visualization device. When a foreign object appears in the blue area, the system will sound an alarm. The blue area is defined as follows (default): Longitudinal: 10-500 meters; Lateral: -40 to 40 meters; Pitch: 0 to 20 meters, with the area above the normal being positive.

[0104] S3: Upload the identification results and positioning data after feature recognition to the monitoring platform to realize remote hazard early warning, image display and operation and maintenance task dispatch.

[0105] This step aims to upload the core results, such as the type of external damage target, spatial coordinates, and conductor gaps identified by AI, to the power grid monitoring platform in a structured format via wireless communication, thereby achieving information linkage with the main station, risk visualization, and closed-loop management of tasks.

[0106] This invention adopts a modular communication architecture, with the terminal integrating a 4G / 5G communication module, compatible with low-power LoRa modules for use in areas with poor signal, and supports remote data backhaul, status synchronization and interaction with platform control commands.

[0107] 1. Communication format and data structure: Each time the terminal identifies an event, it forms a data packet, the content of which includes:

[0108] Target category (Class): Output by the AI ​​model, such as "lifted object" or "soil pile"; Minimum clearance (D) min Unit: meter; Three-dimensional coordinates (X) t ,Y t Z t ): The location of the target in geospatial space; traverse number and corresponding span number L id ,S id : Used for platform positioning; Risk level R∈{low, medium, high}: Determined by a combination of gap and confidence level; Timestamp T: Identification time, with an accuracy of not less than 1 second; Terminal number E id : Unique identifier for the device uploading the device.

[0109] The platform is parsed in JSON structure or industrial Modbus / TCP format and stored in the database for historical tracking and display.

[0110] 2. Remote platform architecture and linkage logic. The remote monitoring platform supporting this invention includes the following functional modules:

[0111] (1) Real-time data access module: Receives data uploaded by the terminal, performs legality verification, and distributes and stores it in the database.

[0112] (2) Graphic display module: Based on WebGIS technology, superimposes the hidden danger locations on the transmission line geographical map, and displays the risk level, target category, clearance value, etc.

[0113] (3) Early warning module: Triggers alarms according to the rule engine, such as:

[0114]

[0115] (4) Task dispatch module: Links with the dispatching platform or the operation and maintenance App, automatically generates hidden danger handling work orders, including fields such as positioning navigation, target description, response time limit, etc.

[0116] (5) Event closed-loop module: Supports the feedback of event processing results, forming a complete "discovery - response - handling - archiving" closed-loop logic.

[0117] 3. Communication fault tolerance and data integrity strategy. To adapt to areas with unstable transmission line communication such as mountainous areas and forest areas, this invention designs a local cache and breakpoint resumption mechanism: Each terminal can cache no less than 1000 frames of event data; The upload adopts a "success confirmation + time polling" mechanism to avoid packet loss; When the communication anomaly exceeds 10 minutes, it automatically switches to LoRa narrowband low-frequency transmission; All uploaded data is attached with a CRC check code to ensure integrity.

[0118] This invention constructs an integrated remote hidden danger monitoring terminal by integrating millimeter-wave radar, Beidou high-precision positioning, AI intelligent recognition, remote communication transmission, and low-power energy supply modules, solving the problems of low recognition accuracy, lack of spatial positioning, complex deployment, and weak real-time response ability existing in the prior art. This system can be deployed along the transmission line. The distance information between the external damage target and the conductor is obtained through the millimeter-wave radar, and the spatial position of the target is synchronously obtained in combination with the Beidou module to achieve three-dimensional spatial annotation. The system is built with an AI recognition and processing unit, which can perform real-time analysis on the echo signal, automatically identify the external damage type and judge the minimum clearance between it and the conductor. The recognition result is uploaded to the remote monitoring platform through the communication module, thus realizing the intelligent recognition and remote early warning of external damage hidden dangers on the transmission line.

[0119] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A remote millimeter-wave ranging and AI recognition system suitable for power transmission lines, characterized in that, include: The data acquisition module acquires real-time data on the location of external damage targets around the transmission line and their relative distance from the conductor; The AI ​​intelligent recognition module uses built-in AI algorithms to identify features in the collected radar echo data and determine the type of hidden danger and its relationship with the minimum gap between the conductor and the conductor. The communication transmission module uploads the recognition results and positioning data of the AI ​​intelligent recognition module to the monitoring platform to realize remote hazard warning, image display and operation and maintenance task dispatch; The acquisition module, communication transmission module, solar panel, and lithium battery are integrated and packaged into a single unit to form a lightweight intelligent monitoring terminal that can be remotely deployed.

2. The remote millimeter-wave ranging and AI recognition system for power transmission lines according to claim 1, characterized in that, The acquisition module is an integrated structure of a millimeter-wave radar module and a BeiDou positioning module. The millimeter-wave radar module and the BeiDou positioning module use a unified control unit for data acquisition and time synchronization.

3. The remote millimeter-wave ranging and AI recognition system for power transmission lines according to claim 2, characterized in that, The millimeter-wave radar module adopts the FMCW (Frequency Modulated Continuous Wave) system, operates in the 77GHz frequency band, has a horizontal scanning angle of ±60° and a vertical elevation angle of ±20°, and a field of view coverage radius of 100 meters. The radar scans periodically with a default frame rate of 5Hz. When an abnormally strong echo target is detected, it automatically upgrades to a 20Hz high-frequency scanning mode.

4. The remote millimeter-wave ranging and AI recognition system for power transmission lines according to claim 2, characterized in that, The Beidou positioning module uses a dual-frequency high-precision receiver and supports real-time differential positioning (RTK) function. The polar coordinate information of the target point collected by the millimeter-wave radar will be combined with the radar installation point coordinates obtained by Beidou to calculate the three-dimensional coordinates of the target. The specific spatial coordinates of the target are calculated using the following formula: Where: X t ,Y t Z t The target's three-dimensional spatial position in the geographic coordinate system; r is the target distance measured by the millimeter-wave radar; θ is the horizontal deflection angle; φ is the vertical elevation angle; X0, Y0, Z0 are the radar installation point coordinates, provided by BeiDou.

5. A remote millimeter-wave ranging and AI recognition system for power transmission lines according to claim 1, characterized in that, In the AI ​​intelligent identification module, the AI ​​algorithm adopts a convolutional neural network (CNN) structure and combines a multi-channel feature fusion mechanism to perform time-frequency domain analysis on radar echo signals, achieving low-latency hazard identification at the edge terminal; the identification content includes typical external damage hazard types, and at the same time, it combines distance to determine whether it constitutes an interference risk.

6. The remote millimeter-wave ranging and AI recognition system for power transmission lines according to claim 1, characterized in that, The communication transmission module supports 4G / 5G cellular networks or LoRa low-power wide-area communication protocol, and has the functions of data breakpoint resume and intelligent compression. After the identification results are uploaded to the monitoring platform, they are marked on the map through the GIS module, and the background early warning mechanism is linked to generate risk reports or push operation and maintenance scheduling tasks.

7. A recognition method for a remote millimeter-wave ranging and AI recognition system for power transmission lines according to any one of claims 1-6, characterized in that, Includes the following steps: S1: Deploy data acquisition devices integrating millimeter-wave radar and BeiDou positioning around the power transmission line to acquire real-time data on the location of external damage targets and their relative distance to the conductor. S2: Use AI algorithms to identify features in the collected radar echo data to determine the type of hidden danger and its relationship with the minimum gap between the conductor and the target. S3: Upload the identification results and positioning data after feature recognition to the monitoring platform to realize remote hazard early warning, image display and operation and maintenance task dispatch.

8. A remote millimeter-wave ranging and AI recognition method for power transmission lines according to claim 7, characterized in that, In step S1, the millimeter-wave radar adopts the FMCW (Frequency Modulated Continuous Wave) system, operates at a frequency of 77 GHz, has a horizontal scanning angle of ±60° and a vertical elevation angle of ±20°, and a field of view coverage radius of 100 meters. The radar scans periodically with a default frame rate of 5 Hz. When an abnormally strong echo target is detected, it automatically upgrades to a high-frequency scanning mode of 20 Hz.

9. A method for long-range millimeter-wave ranging and AI recognition applicable to power transmission lines according to claim 7, characterized in that, In step S1, the BeiDou positioning uses a dual-frequency high-precision receiver that supports real-time differential positioning (RTK) function. The polar coordinate information of the target point collected by the millimeter-wave radar will be combined with the radar installation point coordinates obtained by BeiDou to calculate the three-dimensional coordinates of the target. The specific spatial coordinates of the target are calculated using the following formula: Where: X t ,Y t Z t The target's three-dimensional spatial position in the geographic coordinate system; r is the target distance measured by the millimeter-wave radar; θ is the horizontal deflection angle; φ is the vertical elevation angle; X0, Y0, Z0 are the radar installation point coordinates, provided by BeiDou.

10. A method for long-range millimeter-wave ranging and AI recognition applicable to power transmission lines according to claim 7, characterized in that, Step S2 specifically includes the following steps: (1) Echo signal preprocessing: The original signal output by the millimeter-wave radar is a time-varying echo sequence, and its mathematical expression is as follows: S(t)=A(t)·cos(2πf0t+φ(t)) (2) Where S(t) is the echo signal strength; A(t) is the envelope function, representing the change of echo amplitude over time; f0 is the radar carrier frequency; and φ(t) is the signal phase, which includes range and Doppler information. The original signal is analyzed in time and frequency using Short Time Fourier Transform (STFT) to obtain its time-frequency spectrum, which serves as the input for AI recognition. Where X(t,ω) is the complex amplitude of the echo signal at time t and frequency ω; S(τ) is the original time series signal, i.e., the echo signal to be analyzed; ω(τ-t) is the window function, usually a Hamming window or a Gaussian window; e -jwτ τ is the Fourier kernel function, which projects the local signal into the frequency domain; τ is the integral variable, representing the sliding variable over time. (2) Target feature recognition and classification: A lightweight convolutional neural network structure is used to extract features from the echo spectrum and classify targets, outputting the probability distribution of target categories: Where P is the probability vector for each category; p i Let be the probability value for identifying it as the i-th type of target; n is the total number of categories; Final output target category: (3) Conductor spatial modeling and minimum gap calculation: Based on the actual transmission line structure, the conductor's shape between towers conforms to a parabolic model. Under no-wind static load conditions, its spatial sag is expressed as: Where Z(s) is the height of the conductor at a lateral distance s; H is the height of the conductor suspension point; w is the weight per unit length of conductor; T is the conductor tension; and s is the lateral distance from the tower. By fitting a three-dimensional spatial curve, the spatial trajectory of the conductor is obtained as C(s) = (X(s), Y(s), Z(s)); for a certain target to be identified, its coordinates are (X... t ,Y t Z t If ), then the minimum spatial distance between it and the conductor is: Among them, D min To identify the minimum gap distance between the target and the traverse; s1, s2 correspond to the spatial parameter range of the start and end points of this traverse segment; (X(s), Y(s), Z(s)) are the spatial function points of the traverse; (X t ,Y t Z t () represents the three-dimensional coordinates of the target identified by AI.

Images