CN118092462A - Gas pipeline risk identification method and early warning system based on unmanned aerial vehicle and AI - Google Patents

Abstract

The present invention provides a gas pipeline risk identification method and early warning system based on drones and AI, wherein the method includes: planning a flight inspection route based on a preset gas pipeline network model corresponding to a community; based on the flight inspection route, controlling a drone to perform gas pipeline inspection on the community; continuously acquiring the inspection data sent back by the drone; determining the gas pipeline risk based on the inspection data and a preset AI risk identification model; and outputting an early warning of the gas pipeline risk. The present invention controls a drone to perform gas pipeline inspection on a community based on a planned flight inspection route, determines the gas pipeline risk based on the inspection data sent back by the drone and the AI risk identification model, and automatically and adaptively completes the gas pipeline network risk identification in the community, without the need for gas maintenance personnel to complete it manually, which greatly reduces the labor cost and improves the efficiency of gas pipeline risk identification.

Description

Gas pipeline risk identification method and early warning system based on drones and AI

Technical Field

The present invention relates to the technical field of gas pipeline risk identification, and in particular to a gas pipeline risk identification method and early warning system based on drones and AI.

Background technique

In order to ensure the safety of gas use in the community, it is necessary to identify the risks of gas pipelines. Generally, the risk identification of gas pipelines is done manually by gas maintenance personnel, who go door to door in the community, which is a large workload and has high labor costs. In addition, gas pipelines are mostly laid on the outer walls of community buildings, especially high-rise buildings, which makes the work of gas maintenance personnel more difficult and affects work efficiency. Therefore, a solution is urgently needed.

Summary of the invention

One of the purposes of the present invention is to provide a gas pipeline risk identification method based on drones and AI. Based on the planned flight inspection route, the drone is controlled to conduct gas pipeline inspection in the community. The gas pipeline risk is determined based on the inspection data sent back by the drone and the AI risk identification model. The gas pipeline network risk identification in the community is automatically and adaptively completed without the need for manual completion by gas maintenance personnel, which greatly reduces the labor cost. In addition, the efficiency of gas pipeline risk identification is improved.

The gas pipeline risk identification method based on drone and AI provided in the embodiment of the present invention includes:

Plan flight inspection routes based on the preset gas pipeline network model corresponding to the community;

Based on the flight inspection route, control the drone to inspect the gas pipeline in the community;

Continuously obtain the detection data sent back by the drone;

Determine gas pipeline risks based on detection data and preset AI risk identification models;

Output warning of gas pipeline risks.

Preferably, the gas pipeline risk identification method based on drones and AI also includes:

Determine a first partial route that meets the first partial route condition from the flight detection route;

When the drone flies along the first local route, the drone is controlled to detect a moving route of a person approaching the drone within a first preset range within a first preset time period;

Based on the approaching moving route and the first partial route, determining the avoiding direction; there is no intersection between the ray from the route end point of the approaching moving route to the avoiding direction and the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to a preset distance threshold;

Based on the avoidance direction and the approach movement route, the projection area is determined; the projection area is a preset 1/N circle, the center of the 1/N circle is the end point of the route and the 1/N circle is bisected by the ray; N is greater than 2;

Control the drone to continuously detect the eye position and eye direction of the person approaching;

Constructing a first direction vector based on the eye position and the eye orientation;

Constructing a second direction vector based on the projection position of the UAV and the projection direction from the projection position to the projection area;

When the vector angle between the first direction vector and the second direction vector continues to fall within the preset angle interval for a duration greater than or equal to a preset duration threshold, the drone is controlled to project a preset avoidance prompt pattern corresponding to the avoidance direction to the projection area;

Among them, the first local route conditions include:

A first average flight altitude of the first local route is less than or equal to a preset first altitude threshold;

The first cell facility type passed by the first local route exists in a preset first standard facility type library.

Preferably, the gas pipeline risk identification method based on drones and AI also includes:

Determine a second partial route that meets the second partial route condition from the flight detection route;

When the drone flies along the second local route, the drone is controlled to detect changes in the status of indoor personnel who have entered a second preset range around the drone within a recent second preset time;

When the status change of the person meets the preset change conditions, the drone is controlled to detect the person portrait and window area of the indoor person;

Determine the audio of the operation prompt based on the personnel portrait;

determining a third partial route from the second partial routes that meets the third partial route conditions;

When the UAV flies along the third local route, the UAV is controlled to broadcast the operation prompt audio;

Among them, the second local route conditions include:

The second average flight altitude of the second local route is greater than or equal to a preset second altitude threshold;

The second community facility type passed by the second local route exists in the preset second standard facility type library;

Among them, the third local route conditions include:

The drone will fly along the third local route within a preset third time in the future;

As the drone flies along the third local route, the drone continues to approach the window area.

Preferably, the gas pipeline risk identification method based on drones and AI also includes:

Control drones to track gas pipeline risks.

Preferably, the gas pipeline risk identification method based on drones and AI also includes:

Supplement training of AI risk identification model.

The gas pipeline risk identification system based on drone and AI provided by the embodiment of the present invention includes:

Route planning module, used to plan flight inspection routes based on the preset gas pipeline network model corresponding to the community;

The first control module is used to control the drone to perform gas pipeline inspection in the community based on the flight inspection route;

Data acquisition module, used to continuously acquire detection data sent back by the drone;

The risk determination module is used to determine the gas pipeline risk based on the detection data and the preset AI risk identification model;

The risk warning module is used to output warnings of gas pipeline risks.

Preferably, the gas pipeline risk identification system based on drones and AI also includes:

The second control module is configured to include:

Determine a first partial route that meets the first partial route condition from the flight detection route;

When the drone flies along the first local route, the drone is controlled to detect a moving route of a person approaching the drone within a first preset range within a first preset time period;

Based on the approaching moving route and the first partial route, determining the avoiding direction; there is no intersection between the ray from the route end point of the approaching moving route to the avoiding direction and the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to a preset distance threshold;

Based on the avoidance direction and the approach movement route, the projection area is determined; the projection area is a preset 1/N circle, the center of the 1/N circle is the end point of the route and the 1/N circle is bisected by the ray; N is greater than 2;

Control the drone to continuously detect the eye position and eye direction of the person approaching;

Constructing a first direction vector based on the eye position and the eye orientation;

Constructing a second direction vector based on the projection position of the drone and the projection direction from the projection position to the projection area;

When the vector angle between the first direction vector and the second direction vector continues to fall within the preset angle interval for a duration greater than or equal to a preset duration threshold, the drone is controlled to project a preset avoidance prompt pattern corresponding to the avoidance direction to the projection area;

Among them, the first local route conditions include:

A first average flight altitude of the first local route is less than or equal to a preset first altitude threshold;

The first cell facility type passed by the first local route exists in a preset first standard facility type library.

Preferably, the gas pipeline risk identification system based on drones and AI also includes:

The third control module is used to include:

Determine a second partial route that meets the second partial route condition from the flight detection route;

When the drone flies along the second local route, the drone is controlled to detect changes in the status of indoor personnel who have entered a second preset range around the drone within a recent second preset time;

When the status change of the person meets the preset change conditions, the drone is controlled to detect the person portrait and window area of the indoor person;

Determine the audio of the operation prompt based on the personnel portrait;

determining a third partial route from the second partial routes that meets the third partial route conditions;

When the UAV flies along the third local route, the UAV is controlled to broadcast the operation prompt audio;

Among them, the second local route conditions include:

The second average flight altitude of the second local route is greater than or equal to a preset second altitude threshold;

The second community facility type passed by the second local route exists in the preset second standard facility type library;

Among them, the third local route conditions include:

The drone will fly along the third local route within a preset third time in the future;

As the drone flies along the third local route, the drone continues to approach the window area.

Preferably, the gas pipeline risk identification system based on drones and AI also includes:

The risk tracking module is used to control the drone to track gas pipeline risks.

Preferably, the gas pipeline risk identification system based on drones and AI also includes:

Supplementary training module, used to supplement the training of AI risk identification model.

Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

FIG1 is a schematic diagram of a gas pipeline risk identification method based on drones and AI in an embodiment of the present invention;

FIG2 is a schematic diagram of a gas pipeline risk identification system based on drones and AI in an embodiment of the present invention.

Detailed ways

The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

The embodiment of the present invention provides a gas pipeline risk identification method based on drones and AI, as shown in FIG1 , including:

Step S1: planning a flight inspection route based on a preset gas pipeline network model corresponding to the community;

Step S2: Based on the flight detection route, control the drone to perform gas pipeline detection in the community;

Step S3: Continuously obtain the detection data sent back by the drone;

Step S4: Determine the gas pipeline risk based on the detection data and the preset AI risk identification model;

Step S5: Output warning gas pipeline risk.

In the above scheme, the gas pipeline network model is a three-dimensional model of the community, on which the layout of the gas pipeline network in the community is marked; based on the gas pipeline network model, a flight inspection route for drones to conduct gas pipeline inspections in the community can be planned, and the flight inspection route passes through each route point on the flight inspection route; when the drone conducts gas pipeline inspections in the community, the laser methane telemeter, methane sensor, etc. equipped on the drone will start working, detect and send back the detection data; the AI risk identification model is an artificial intelligence model that uses a large amount of detection data from laser methane telemeters, methane sensors, etc. representing gas risks as training samples to train the neural network model until convergence; the AI risk identification model can identify gas pipeline risks based on the detection data; when the gas pipeline risk is determined, an early warning is output.

This application controls drones to conduct gas pipeline inspections in residential areas based on planned flight inspection routes, determines gas pipeline risks based on the inspection data sent back by drones and AI risk identification models, and automatically and adaptively completes gas pipeline network risk identification in the residential area. This eliminates the need for manual work by gas maintenance personnel, greatly reducing labor costs. In addition, it improves the efficiency of gas pipeline risk identification.

In one embodiment, the gas pipeline risk identification method based on drones and AI further includes:

Determine a first partial route that meets the first partial route condition from the flight detection route;

When the drone flies along the first local route, the drone is controlled to detect the approach movement route of the approaching person who enters the first preset range around the drone within the latest preset first time; the first time may be, for example, 100 seconds; the first range is a circular range with a radius of 15 meters around the drone; the approaching person may be a resident in a residential area, etc.; the approach movement route is the route generated by the feet of the approaching person moving on the ground; the approach movement route may be obtained by a camera, a millimeter wave radar sensor, etc. equipped by the drone;

Based on the approaching moving route and the first partial route, the avoidance direction is determined; there is no intersection between the ray from the end point of the approaching moving route to the avoidance direction and the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to a preset distance threshold; the distance threshold may be, for example, 0.5 meters; there is no intersection between the ray and the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to the distance threshold, which can ensure that when the approaching person avoids in the avoidance direction, there will be no collision with the drone in the future, etc.;

The projection area is determined based on the avoidance direction and the approaching movement route; the projection area is a preset 1/N circle, the center of the 1/N circle is the route end point and the 1/N circle is bisected by the ray; N is greater than 2; the approaching movement route is the route generated by the feet of the approaching person moving on the ground, then the route end point of the approaching movement route is on the ground, and the ray bisected the 1/N circle, then the projection area is on the ground;

Control the drone to continuously detect the eye position and eye orientation of the person approaching; the eye position and eye orientation can be obtained through the camera equipped on the drone;

Constructing a first direction vector based on the eye position and the eye orientation;

Constructing a second direction vector based on the projection position of the UAV and the projection direction from the projection position to the projection area;

When the vector angle between the first direction vector and the second direction vector continuously falls within the preset angle interval for a duration greater than or equal to a preset duration threshold, the drone is controlled to project a preset avoidance prompt pattern corresponding to the avoidance direction to the projection area; the angle interval can be 30 degrees to 180 degrees; the duration threshold can be, for example: 2 seconds; the avoidance prompt pattern can be an arrow animation indicating the avoidance direction; when the vector angle falls within the angle interval, it can be ensured that when the drone starts to project to the projection area, the nearby personnel can perceive it through vision, so that they will look at the projection area, see the avoidance prompt pattern, and get the avoidance prompt, thereby improving the avoidance prompt effect and accuracy; the projection can be achieved through the projection device equipped by the drone;

Among them, the first local route conditions include:

The first average flight height of the first partial route is less than or equal to a preset first height threshold; the first height threshold may be, for example, 3 meters; the first average flight height is an average of the flight heights of the route points on the first partial route;

The first community facility type that the first partial route passes through exists in a preset first standard facility type library. The first standard facility type library contains a large number of first standard facility types, which are facility types where people on the facilities in the community may collide with low-flying drones, such as slides.

When drones are inspecting gas pipelines in a residential area, they are usually carried out during the day to avoid affecting the rest of the residents. During the day, there will be more residents walking around in the community. When they see a drone, they will get closer to check it out out of curiosity. However, residents do not know where the drone will move to in the future, and may not have time to avoid it, resulting in a collision accident. In addition, since the drone is in the air, it cannot project its future flight route, and residents have no way to learn about it. The embodiments of the present invention can solve these two problems: control the drone to detect approaching personnel, reasonably determine the projection area, project an avoidance prompt pattern, and give avoidance prompts to approaching personnel, which greatly improves the safety of drone gas pipeline inspection operations and improves the applicability of drone operations in the community. In addition, the first local route condition is introduced, the first local route is determined, and the drone is automatically triggered to give an avoidance prompt, which reduces the working resources of the drone and improves the efficiency of the avoidance prompt.

In one embodiment, the gas pipeline risk identification method based on drones and AI further includes:

Determine a second partial route that meets the second partial route condition from the flight detection route;

When the drone flies along the second local route, the drone is controlled to detect changes in the status of indoor personnel who enter a second preset range around the drone within a recent preset second time; the second time may be, for example, 40 seconds; the second range is a circular range with a radius of 5 meters around the drone; changes in the status of personnel include: changes in the movements of indoor personnel, etc.; changes in the status of personnel may be obtained by a camera equipped with the drone; indoor personnel may be personnel in the kitchen, etc.;

When the change of the person's status meets the preset change conditions, the drone is controlled to detect the person's portrait and window area of the indoor person; the change condition is that the change of the person's status represents the change of the action of the indoor person being affected by the drone, curious about the legality of the drone, etc., such as opening the window to look at the drone, etc.; the person's portrait includes: gender, predicted age range, etc.; the window area is the area of the window of the room where the indoor person is located, such as: the kitchen window area, etc.; the person's portrait and window area can be obtained through the camera equipped by the drone;

Based on the personnel portrait, the operation prompt audio is determined; the operation prompt audio is used to remind the indoor personnel that the drone is operating legally and there is no need to worry. The operation prompt audio is adapted to the personnel portrait. For example, if the personnel portrait is female and the predicted age range is 40 to 50 years old, the operation prompt audio is "Hello, Auntie, I am inspecting the gas pipeline. Don't worry, I will leave after the inspection!"

determining a third partial route from the second partial routes that meets the third partial route conditions;

When the UAV flies along the third local route, the UAV is controlled to broadcast the operation prompt audio;

Among them, the second local route conditions include:

The second average flight height of the second partial route is greater than or equal to a preset second height threshold; the second height threshold may be, for example, 10 meters; the second average flight height is an average of the flight heights of the route points on the second partial route;

The second community facility type that the second local route passes through exists in the preset second standard facility type library; the second standard facility type library contains a large number of second standard facility types, and the second standard facility type is a facility in the community where people may be affected by high-altitude drones, such as a kitchen, etc.;

Among them, the third local route conditions include:

The drone will fly along the third local route within a preset third time in the future; the third time may be, for example, 20 seconds;

As the drone flies along the third local route, the drone continues to approach the window area.

Generally, when drones are identifying gas pipeline risks in a residential area, they will fly at the entrance gas pipelines on the outer wall of the residential corridor, and may get close to residents' kitchens, etc. If residents are in the kitchen at this time, they may monitor the drones out of concern that the drones may leak their privacy or pose a security threat due to illegal flight, affecting their normal lives. Some residents may even call the police, etc. This is especially true in residential areas that have never used drones to identify gas pipeline risks. An embodiment of the present invention can solve this problem: when it is determined that indoor personnel are affected by drone flight, an operation prompt audio is reasonably determined to be broadcast to remind indoor personnel that there is no need to worry; in addition, a second local route is introduced, and the second local route is determined to automatically trigger the drone to issue an operation prompt, thereby reducing the drone's working resources and improving the efficiency of the operation prompt; secondly, the introduction of

The third local route conditions can ensure that indoor personnel have a feeling that the drone flight is receiving targeted prompts, thereby improving the accuracy and efficiency of operation prompts.

In one embodiment, the gas pipeline risk identification method based on drones and AI further includes:

Control drones to track gas pipeline risks.

Drones can also be controlled to track gas pipeline risks. For example, drones can be controlled to continuously monitor locations where gas pipeline risks occur and send back monitoring data.

In one embodiment, the gas pipeline risk identification method based on drones and AI further includes:

Supplement training of AI risk identification model.

It is also possible to supplement the training of the AI risk identification model to enhance its ability to identify gas risks.

The embodiment of the present invention provides a gas pipeline risk identification system based on drones and AI, as shown in FIG2 , including:

Route planning module 1, used to plan a flight inspection route based on a preset gas pipeline network model corresponding to the community;

The first control module 2 is used to control the drone to perform gas pipeline inspection in the community based on the flight inspection route;

Data acquisition module 3, used to continuously acquire the detection data sent back by the drone;

Risk determination module 4, used to determine gas pipeline risks based on detection data and a preset AI risk identification model;

The risk warning module 5 is used to output warning gas pipeline risks.

The gas pipeline risk identification system based on drones and AI also includes:

The second control module is configured to include:

Determine a first partial route that meets the first partial route condition from the flight detection route;

When the drone flies along the first local route, the drone is controlled to detect a moving route of a person approaching the drone within a first preset range within a first preset time period;

Based on the approaching moving route and the first partial route, determining the avoiding direction; there is no intersection between the ray from the route end point of the approaching moving route to the avoiding direction and the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to a preset distance threshold;

Based on the avoidance direction and the approach movement route, the projection area is determined; the projection area is a preset 1/N circle, the center of the 1/N circle is the end point of the route and the 1/N circle is bisected by the ray; N is greater than 2;

Control the drone to continuously detect the eye position and eye direction of the person approaching;

Constructing a first direction vector based on the eye position and the eye orientation;

Constructing a second direction vector based on the projection position of the UAV and the projection direction from the projection position to the projection area;

When the vector angle between the first direction vector and the second direction vector continues to fall within the preset angle interval for a duration greater than or equal to a preset duration threshold, the drone is controlled to project a preset avoidance prompt pattern corresponding to the avoidance direction to the projection area;

Among them, the first local route conditions include:

A first average flight altitude of the first local route is less than or equal to a preset first altitude threshold;

The first cell facility type passed by the first local route exists in a preset first standard facility type library.

The gas pipeline risk identification system based on drones and AI also includes:

The third control module is used to include:

Determine a second partial route that meets the second partial route condition from the flight detection route;

When the drone flies along the second local route, the drone is controlled to detect changes in the status of indoor personnel who have entered a second preset range around the drone within a recent second preset time;

When the status change of the person meets the preset change conditions, the drone is controlled to detect the person portrait and window area of the indoor person;

Determine the audio of the operation prompt based on the personnel portrait;

determining a third partial route from the second partial routes that meets the third partial route conditions;

When the UAV flies along the third local route, the UAV is controlled to broadcast the operation prompt audio;

Among them, the second local route conditions include:

The second average flight altitude of the second local route is greater than or equal to a preset second altitude threshold;

The second community facility type passed by the second local route exists in the preset second standard facility type library;

Among them, the third local route conditions include:

The drone will fly along the third local route within a preset third time in the future;

As the drone flies along the third local route, the drone continues to approach the window area.

The gas pipeline risk identification system based on drones and AI also includes:

The risk tracking module is used to control the drone to track gas pipeline risks.

The gas pipeline risk identification system based on drones and AI also includes:

Supplementary training module, used to supplement the training of AI risk identification model.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims (10)

1. A gas pipeline risk identification method based on drones and AI, characterized by including: Plan flight inspection routes based on the preset gas pipeline network model corresponding to the community; Based on the flight inspection route, the drone is controlled to inspect the gas pipeline in the community; Continuously obtain the detection data sent back by the drone; Determine the gas pipeline risk based on the detection data and the preset AI risk identification model; Output warning of the gas pipeline risk. 2. The method for gas pipeline risk identification based on drone and AI according to claim 1, further comprising: Determine a first partial route that meets the first partial route condition from the flight detection route; When the drone flies along the first local route, the drone is controlled to detect a moving route of a person approaching the drone within a first preset range around the drone within a recent first preset time; Based on the approaching moving route and the first partial route, determining an escaping direction; a ray from the route end point of the approaching moving route to the escaping direction has no intersection with the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to a preset distance threshold; Based on the escaping direction and the approaching moving route, a projection area is determined; the projection area is a preset 1/N circle, the center of the 1/N circle is the end point of the route and the 1/N circle is bisected by the ray; N is greater than 2; Controlling the drone to continuously detect the eye position and eye orientation of the approaching person; Constructing a first direction vector based on the eye position and the eye orientation; Constructing a second direction vector based on the projection position of the drone and the projection direction from the projection position to the projection area; When the vector angle between the first direction vector and the second direction vector continues to fall within the preset angle interval for a duration greater than or equal to a preset duration threshold, controlling the drone to project a preset avoidance prompt pattern corresponding to the avoidance direction to the projection area; Wherein, the first local route condition includes: A first average flight altitude of the first local route is less than or equal to a preset first altitude threshold; The first cell facility type passed by the first local route exists in a preset first standard facility type library. 3. The method for gas pipeline risk identification based on drone and AI according to claim 1, further comprising: Determine a second partial route that meets the second partial route condition from the flight detection route; When the drone flies along the second local route, the drone is controlled to detect a change in the status of a person who has entered an indoor area within a preset second range around the drone within a recently preset second time; When the state change of the person meets the preset change condition, the drone is controlled to detect the person portrait and the window area of the indoor person; Based on the personnel portrait, determine the operation prompt audio; Determine a third partial route that meets the third partial route condition from the second partial routes; When the drone flies along the third local route, controlling the drone to broadcast the operation prompt audio; Wherein, the second local route condition includes: The second average flight altitude of the second local route is greater than or equal to a preset second altitude threshold; The second cell facility type passed by the second local route exists in a preset second standard facility type library; Wherein, the third local route condition includes: The drone will fly along the third local route within a third preset time in the future; When the drone flies along the third local route, the drone continues to approach the window area. 4. The method for gas pipeline risk identification based on drone and AI according to claim 1, further comprising: Control the drone to track the gas pipeline risks. 5. The method for gas pipeline risk identification based on drone and AI according to claim 1, further comprising: Supplement training of the AI risk identification model. 6. The gas pipeline risk identification system based on drones and AI is characterized by including: Route planning module, used to plan flight inspection routes based on the preset gas pipeline network model corresponding to the community; A first control module is used to control the drone to perform gas pipeline inspection on the community based on the flight inspection route; A data acquisition module, used to continuously acquire the detection data transmitted back by the drone; A risk determination module, used to determine the gas pipeline risk based on the detection data and a preset AI risk identification model; The risk warning module is used to output the gas pipeline risk warning. 7. The gas pipeline risk identification system based on drone and AI according to claim 6, characterized in that it also includes: The second control module is configured to include: Determine a first partial route that meets the first partial route condition from the flight detection route; When the drone flies along the first local route, the drone is controlled to detect a moving route of a person approaching the drone within a first preset range around the drone within a recent first preset time; Based on the approaching moving route and the first partial route, determining an escaping direction; a ray from the route end point of the approaching moving route to the escaping direction has no intersection with the first partial route, and the shortest distance between the ray and the first partial route is greater than or equal to a preset distance threshold; Based on the escaping direction and the approaching moving route, a projection area is determined; the projection area is a preset 1/N circle, the center of the 1/N circle is the end point of the route and the 1/N circle is bisected by the ray; N is greater than 2; Controlling the drone to continuously detect the eye position and eye orientation of the approaching person; Constructing a first direction vector based on the eye position and the eye orientation; Constructing a second direction vector based on the projection position of the drone and the projection direction from the projection position to the projection area; When the vector angle between the first direction vector and the second direction vector continues to fall within the preset angle interval for a duration greater than or equal to a preset duration threshold, controlling the drone to project a preset avoidance prompt pattern corresponding to the avoidance direction to the projection area; Wherein, the first local route condition includes: A first average flight altitude of the first local route is less than or equal to a preset first altitude threshold; The first cell facility type passed by the first local route exists in a preset first standard facility type library. 8. The gas pipeline risk identification system based on drone and AI according to claim 6, characterized in that it also includes: The third control module is used to include: Determine a second partial route that meets the second partial route condition from the flight detection route; When the drone flies along the second local route, the drone is controlled to detect a change in the status of a person who has entered an indoor area within a preset second range around the drone within a recently preset second time; When the state change of the person meets the preset change condition, the drone is controlled to detect the person portrait and the window area of the indoor person; Based on the personnel portrait, determine the operation prompt audio; Determining a third partial route that meets the third partial route condition from the second partial routes; When the drone flies along the third local route, controlling the drone to broadcast the operation prompt audio; Wherein, the second local route condition includes: The second average flight altitude of the second local route is greater than or equal to a preset second altitude threshold; The second cell facility type passed by the second local route exists in a preset second standard facility type library; Wherein, the third local route condition includes: The drone will fly along the third local route within a third preset time in the future; When the drone flies along the third local route, the drone continues to approach the window area. 9. The gas pipeline risk identification system based on drone and AI according to claim 6, characterized in that it also includes: The risk tracking module is used to control the drone to track the gas pipeline risk. 10. The gas pipeline risk identification system based on drone and AI according to claim 6, characterized in that it also includes: A supplementary training module is used to supplement the training of the AI risk identification model.