Search

CN-121985312-A - 5G edge AI-based transmission line unmanned aerial vehicle anti-external-damage monitoring device

CN121985312ACN 121985312 ACN121985312 ACN 121985312ACN-121985312-A

Abstract

The invention relates to the technical field of on-line monitoring and safety protection of electric power facilities, in particular to an outer breakage prevention monitoring device of an electric transmission line unmanned aerial vehicle based on a 5G edge AI, which comprises a terminal sensing layer, an edge intelligent layer and a cloud management platform; the terminal perception layer is used for collecting visible light, heat radiation data and video streams of a line corridor and realizing primary identification of a risk target through edge calculation, the edge intelligent layer receives a structured AI identification result, operates an AI decision model to analyze the target, predicts the behavior and evaluates the risk level and issues a control instruction, can judge a motion track and a trend by combining with target historical position data, can simultaneously receive results of a plurality of terminal perception layers and fuse the results for analysis, and the cloud management platform receives early warning information, performs data storage, task planning and scheduling and is responsible for training, optimizing and downward deployment of the AI model. The invention realizes real-time accurate identification and instant early warning of the risk of external damage, and improves response speed and reliability.

Inventors

  • WANG ZEXI
  • CHEN LIMING
  • LOU LE
  • YUAN JIA
  • MAO PENG

Assignees

  • 国网新疆电力有限公司塔城供电公司
  • 塔城精益电力建设有限公司

Dates

Publication Date
20260505
Application Date
20260210

Claims (10)

  1. 1. The utility model provides an electric transmission line unmanned aerial vehicle prevents broken monitoring devices outward based on 5G edge AI which characterized in that includes terminal perception layer, edge intelligent layer and high in the clouds management platform; the terminal perception layer is used for collecting visible light, heat radiation data and video streams of the line corridor and realizing primary identification of a risk target through edge calculation; The edge intelligent layer is used for receiving a structured AI identification result output by the unmanned aerial vehicle from the terminal perception layer, operating an AI decision model to analyze a target, predict behavior and evaluate risk level, and issuing a control instruction to the terminal perception layer; The cloud management platform is used for receiving the early warning information issued by the edge intelligent layer, carrying out data storage, task planning and scheduling, and is responsible for training and optimizing the AI model and deploying the AI model downwards to the edge intelligent layer and the terminal perception layer.
  2. 2. The 5G edge AI-based transmission line unmanned aerial vehicle outward-breakage prevention monitoring device according to claim 1, wherein the terminal perception layer integrates a sensing module, a 5G communication module and an edge AI computing unit, wherein the sensing module is used for collecting visible light and thermal radiation data of a line corridor, the 5G communication module is used for realizing connection of the unmanned aerial vehicle and a 5G network, and the edge AI computing unit is internally provided with a deep learning model for processing collected video streams.
  3. 3. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-breakage prevention monitoring device of claim 2, wherein the sensing module comprises a high-definition visible light camera and an infrared thermal imager, the high-definition visible light camera is used for collecting visible light data of a line corridor, and the infrared thermal imager is used for collecting heat radiation data of the line corridor.
  4. 4. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-breakage prevention monitoring device according to claim 2 or 3, wherein the deep learning model is a YOLOv s lightweight model which is cut and optimized, and the optimization direction comprises calculation force adaptation, recognition speed improvement and target detection accuracy reinforcement.
  5. 5. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-break prevention monitoring device of claim 1, 2 or 3, wherein the structured AI identification result comprises a target type, GPS coordinates, a time stamp, and a confidence level.
  6. 6. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-breaking prevention monitoring device according to claim 1, 2 or 3, wherein the control instruction issued by the edge intelligent layer comprises directing the unmanned aerial vehicle to fly to the upper air of the target, and starting an onboard loudspeaker to perform high-altitude shouting warning.
  7. 7. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-breaking prevention monitoring device according to claim 1,2 or 3, wherein the cloud management platform further comprises a visual display module for visually displaying a target type, GPS coordinates, a risk level and a time stamp in the early warning information.
  8. 8. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-breakage prevention monitoring device according to claim 1,2 or 3, wherein the edge intelligent layer is used for determining the movement track and trend of the target by analyzing the position change of the target in a continuous time period in combination with the historical position data of the target when evaluating the risk level.
  9. 9. The 5G edge AI-based power transmission line unmanned aerial vehicle external damage prevention monitoring device according to claim 1, 2 or 3, wherein the terminal perception layer is used for independently completing local recognition and basic early warning when the terminal perception layer is interrupted with a cloud network for a short time.
  10. 10. The 5G edge AI-based power transmission line unmanned aerial vehicle outward-breaking prevention monitoring device according to claim 1,2 or 3, wherein the edge intelligent layer receives the structured AI recognition results of the plurality of terminal perception layers simultaneously and performs fusion analysis.

Description

5G edge AI-based transmission line unmanned aerial vehicle anti-external-damage monitoring device Technical Field The invention relates to the technical field of on-line monitoring and safety protection of electric power facilities, in particular to an anti-external-damage monitoring device for an unmanned aerial vehicle of a power transmission line based on a 5G edge AI. Background With the rapid development of the ultra-high voltage power grid, the coverage area of the power transmission line is continuously enlarged, and the safe and stable operation of the ultra-high voltage power grid has important significance for guaranteeing power supply and supporting social and economic development. However, external damage has become the primary factor that threatens the safe operation of transmission line, and various external broken events such as crane line, off-line violation construction, mountain fire frequently occur, can not only lead to line trouble outage, causes huge economic loss, and serious consequences such as large-area power failure can also be triggered, influence the normal production life order of society. In the prior art, the solution for monitoring the power transmission line against external damage has a plurality of obvious defects. According to the traditional unmanned aerial vehicle image transmission technical scheme, after the unmanned aerial vehicle is relied on to collect high-definition videos, all original video data are transmitted back to a remote cloud platform through a 4G network or wireless image transmission for analysis and processing. The mode has extremely high requirements on network bandwidth, and the problems of large data transmission delay and even interruption are extremely easy to occur in areas with poor signal coverage in remote mountain areas, suburbs and the like. More importantly, the remote transmission of massive video data can lead to slow analysis response speed, delay of a few seconds to tens of seconds is usually required, the core requirement of real-time early warning in external damage risk monitoring cannot be met at all, and accidents often occur when early warning information is sent out. The pure cloud AI analysis mode concentrates all AI identification tasks on the cloud server for processing, when the number of the accessed unmanned aerial vehicles increases, the calculation pressure of the cloud server can be increased sharply, performance bottlenecks are easy to form, the system processing efficiency is greatly reduced, and the expansibility is extremely poor. In addition, the mode has extremely high stability dependence on network connection, once the network is fluctuated or interrupted, the whole monitoring system can fall into a paralytic state, any monitoring and early warning effect can not be exerted, and the reliability of the system is seriously affected. In addition, some schemes that attempt data processing on an unmanned aerial vehicle are difficult to deploy due to limitations in the computational power, power consumption and weight of the on-board equipment. The existing airborne computing equipment is limited in computing power, the requirements of high-precision and high-speed identification cannot be met, meanwhile, the endurance of the unmanned aerial vehicle can be seriously influenced by excessive power consumption, and the burden and the safety risk of the unmanned aerial vehicle in flying can be increased by the larger weight. Therefore, the scheme can only perform simple target detection, cannot perform deep analysis on complex behaviors of the target, is not ideal in recognition accuracy and speed, and is difficult to effectively cope with diversified external damage risk scenes. The defects of the prior art are interwoven, so that the transmission line external damage prevention monitoring always faces the problems of outstanding contradiction between data transmission bandwidth and time delay, slow cloud centralized processing response, poor reliability under severe network environment, insufficient intelligent level of the machine-mounted environment and the like, the real-time external damage prevention monitoring with low delay, high reliability and wide area coverage cannot be realized, and the safe and stable operation of the transmission line is difficult to effectively guarantee. Disclosure of Invention The invention provides a 5G edge AI-based transmission line unmanned aerial vehicle anti-external damage monitoring device, which overcomes the defects of the prior art, and can effectively solve the problems of slow response, poor reliability and insufficient recognition precision in the existing transmission line anti-external damage monitoring. The technical scheme of the invention is realized by the following measures that the transmission line unmanned aerial vehicle anti-external-damage monitoring device based on the 5G edge AI comprises a terminal sensing layer, an edge intelligent layer and a cloud mana