Unmanned aerial vehicle autonomous inspection system and method based on machine vision

Through deep learning analysis technology and FSBi-RRT algorithm, the drone inspection route is adjusted using infrared image change characteristics of power equipment, and the path optimization and safety hazards caused by infrared feature changes in the existing technology are solved, and the drone patrol is realized Check the high accuracy and safety of flights.

CN118915794APending Publication Date: 2024-11-08NANJING QIHE INTELLIGENT TECH CO LTD
View PDF 0 Cites 5 Cited by

Patent Information

Application Number
CN202411109478.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing drone inspection system is difficult to automatically generate and optimize inspection routes based on the changing state of infrared characteristics of power equipment, resulting in hidden dangers in path optimization and safety guarantees.

Method used

Using deep learning analysis technology, different patrol flight routes are preset for the drone through the changing characteristics of infrared images of power equipment, and the flight altitude deviation is judged during the flight, the flight trajectory is adjusted, and the small route segments are divided using the FSBi-RRT algorithm. , generate line data models, update image features in real time, and ensure that the drone returns to the correct flight trajectory.

Benefits of technology

The high accuracy of drone inspection flight is achieved. By adjusting the flight trajectory in real time, it ensures that the drone can return to the correct flight trajectory more accurately, improving the diversity and safety of patrol routes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118915794A_ABST
    Figure CN118915794A_ABST
Patent Text Reader

Abstract

The invention discloses an unmanned aerial vehicle autonomous inspection system and method based on machine vision, and belongs to the technical field of unmanned aerial vehicle inspection. The information data acquisition module is used for acquiring coordinates of a position where preset power equipment is located in a preset region and an equipment area of the preset power equipment, the information data acquisition module is used for acquiring historical infrared data of the preset power equipment, the infrared data comprises infrared image data, and the acquired historical infrared image data is divided; different inspection flight routes are preset for the unmanned aerial vehicle through the change characteristics of the infrared image of the power equipment by using a deep learning analysis technology, and whether the flight height of the unmanned aerial vehicle deviates can be judged in the inspection flight process of the unmanned aerial vehicle, so that the flight track height of the unmanned aerial vehicle is adjusted; and the inspection flight of the unmanned aerial vehicle can more accurately return to a correct flight path.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle inspection, and in particular to an unmanned aerial vehicle autonomous inspection system and method based on machine vision. Background Art

[0002] The drone inspection system is an innovative application that combines machine vision technology and drone technology. It uses machine vision algorithms and cameras or sensors carried by drones to achieve automated monitoring and inspection of targets in outdoor or indoor environments. In real life, with the continuous development of social economy, national electricity load is also growing. Among them, the use of infrared temperature measurement technology to detect heating and / or fire hazards of power equipment has been widely used in the current power system.

[0003] The applicant retrieved a domestic technology application number CN202410022254.7, which provides a substation drone automatic inspection system and method. This technology application uses a directional self-learning strategy and an algorithm to determine the search direction to form the final path optimization, thereby accelerating the search for optimization points, thereby improving efficiency while reducing safety accidents and ensuring the safety of substations. There is also a domestic technology application involving the optimization of drone inspection paths in the power industry, "A method for generating drone distribution network inspection routes", whose technology application number is CN202311082618.2. The technical application calculates the take-off and landing points and optimizes the distribution network inspection route based on the location of key towers and drone parameters, and quickly selects the optimal model and the optimal inspection route based on the fault prediction points within the inspection range, effectively improving the inspection efficiency; in addition, the domestic technical application "A patrol drone path planning system" has a technical application number of CN202111605225.6. The technical application system includes a drone control unit, a route planning unit, an endurance unit, an inspection unit, an early warning unit and a communication unit, which effectively plan feasible routes and realize large-scale inspections of complex terrain.

[0004] The above technical solutions all use the target coordinates as the main body for inspection route planning. However, as an entity, the target coordinates have different infrared characteristics depending on the fault, which makes their external visual states different. However, it is difficult for the existing technology to select and automatically generate various inspection routes according to the different changing states of the target coordinates, which leads to hidden dangers in path optimization and ensuring the safety of the target coordinates. Summary of the invention

[0005] In view of the above-mentioned problems existing in the existing drone inspections, the present invention is proposed.

[0006] Therefore, one of the objects of the present invention is to provide a UAV autonomous inspection system and method based on machine vision, which utilizes deep learning analysis technology to preset different inspection flight routes for UAVs through the changing characteristics of infrared images of power equipment. During the UAV inspection flight, it can determine whether the UAV flight altitude deviates, and then adjust the UAV flight trajectory altitude so that the UAV inspection flight can more accurately return to the correct flight trajectory.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] On the one hand, the present invention provides a drone autonomous inspection system based on machine vision, comprising:

[0009] A comprehensive information acquisition module, used to acquire the range of a preset UAV and geographic information of a preset area, and also includes a module for acquiring the coordinates of a location of a preset power device in a preset area and the equipment area of ​​the preset power device;

[0010] An information data acquisition module is used to collect historical infrared data of preset power equipment, the infrared data includes infrared image data, divide the collected historical infrared image data, obtain the change law of the infrared image data, and generate a prediction model;

[0011] An image data processing unit, the image data processing unit is used to respond to the infrared image data collected by the information data collection module, including an image data compression module for compressing and processing the infrared image data, and an image interception module for intercepting image data of a preselected area in the image data compression module;

[0012] A route planning generation system, wherein the route planning generation unit includes an analysis unit, a three-dimensional modeling unit and a calculation unit;

[0013] The analysis unit is used to analyze the height difference between the preset power equipment and the surrounding buildings based on the position coordinates of the preset power equipment in the preset area, and obtain the average height difference therebetween, and is also used to obtain the angle difference between different infrared image areas on the preset power equipment and the surrounding buildings, and distinguish the surrounding buildings;

[0014] The three-dimensional modeling unit is used to draw and generate a real-life three-dimensional inspection flight route for different infrared image areas and surrounding buildings on the preset power equipment, and divide the real-life three-dimensional inspection flight route into Wherein, n represents the nth inspection flight route, wherein an infrared image area is selected and preset as a reference area, and route features of different inspection flight routes and the reference area are obtained, and a route data model is generated at the same time;

[0015] The calculation unit is used to divide the entire flight segment of each different inspection flight route into at least three small route segments according to the FSBi-RRT algorithm, wherein the control input of each small route segment is regarded as an independent variable, and the state variable is regarded as a dependent variable, and the image features of the reference area obtained by the different inspection flight routes are calculated and obtained, and the image features are uploaded to the route data model, and updated at the same time;

[0016] The monitoring terminal is used to record the communication base station and / or communication base station operator that matches the preset UAV flight, receive the field data uploaded by the collection monitoring point, and update the background data at the same time; when the preset UAV inspection flight in the preset area deviates from the preset inspection flight route, a request for obtaining information on the preset UAV inspection flight status is initiated to the communication base station and / or communication base station operator.

[0017] As a preferred solution of the present invention, in the information data acquisition module, the historical infrared image data is divided, including dividing the historical infrared image data according to the seasonal changes of the preset area and / or the changes in the power consumption of the preset area according to θ1, θ2, ..., θ n The infrared image data are divided, wherein n represents the nth type of infrared image data, data features between different infrared image data are analyzed, the data features are uploaded to the prediction model, and the prediction model is updated.

[0018] As a preferred solution of the present invention, wherein: according to seasonal changes and / or changes in electricity consumption in a preset area, change characteristics of infrared images of preset power equipment from the central area to the edge area are collected, the change characteristics are divided into at least 6 change circles, and the inspection flight route of the preset UAV is planned according to the infrared image characteristics of different change circles that need to be collected. When the infrared image collected by the preset UAV after arriving at the collection point does not meet the infrared image characteristics of the different change circles that need to be collected, the system determines that the inspection flight route of the preset UAV is an invalid route, otherwise, it will not be determined.

[0019] As a preferred solution of the present invention, the areas occupied by different change circles of the infrared image are analyzed, and the sizes of the infrared images that can be taken by the preset drone at different flight altitudes are obtained, and the areas occupied by the different change circles are divided into the following order: ω1, ω2, ..., ω n Divide it into categories, where n represents the area occupied by the nth category, and calculate the safe distance between different occupied areas and surrounding buildings according to the following formula:

[0020] Where t represents the historical time when the occupied area changes the most in different seasons and / or under different electricity consumption conditions;

[0021] In the formula, ω v j represents the jth spacing size data measured for the vth collection point in the different change circles, χ is the independent variable, which represents the change function of the area occupied by different change circles, and a safety threshold is preset based on the measured minimum spacing data. When the drone is preset to collect infrared images at the end position of the selected inspection flight route in the future period, if the distance between the end position and the surrounding buildings is lower than the safety threshold, the system determines that the selected inspection flight route is a dangerous route, and the system automatically deletes the route, otherwise, it will not be determined.

[0022] As a preferred solution of the present invention, the total number of the dangerous routes is counted, and different routes are divided into at least 5 small route segments in the total number of routes, and the flight trajectory coordinate data of different small route segments are obtained, and the height between the initial flight path point of the preset drone and the reference terrain ground of the preset area is recorded as m i o , and according to the height m i o Calculate the data difference between it and the flight trajectory coordinate data of different small route segments, and calculate it according to the following formula:

[0023] Among them, m i represents the ith height value, and o represents the trajectory coordinate corresponding to the height value;

[0024] In the formula, e i represents the collected flight trajectory coordinate data of the ith small route segment, and ρ represents the distance value, which is the distance value of the collected small route segment.

[0025] As a preferred solution of the present invention, wherein: a patrol flight route is preset based on the infrared image data of the different changing circles that need to be obtained, a patrol flight altitude value of the reference terrain ground is given in the patrol flight route, and a critical threshold is preset based on the patrol flight altitude value. When the actual flight altitude of the preset UAV is lower than and / or exceeds the critical threshold, the system determines that a deviation has occurred in the patrol flight, and issues an early warning to the communication base station and / or the communication base station operator that matches the preset UAV flight.

[0026] As a preferred solution of the present invention, wherein: based on the critical threshold, the change characteristics of the flight altitude of the preset UAV during the inspection flight are collected, the altitude value corresponding to the change characteristics is marked at least three times, and the data characteristics of each marked value are analyzed to obtain the increase and / or decrease value of the marked data at the critical threshold. When the actual flight altitude of the preset UAV is lower than and / or exceeds the critical threshold, the system automatically inputs the obtained increase and / or decrease value into the flight control instruction of the preset UAV, otherwise, it is not input.

[0027] As a preferred solution of the present invention, the acquired increase value and / or decrease value data are divided into several evaluation indicators, the evaluation indicator data during the inspection flight are statistically calculated, and the data are normalized to calculate the weight of each evaluation indicator.

[0028] On the other hand, the present invention provides a method for autonomous inspection of unmanned aerial vehicles based on machine vision, which is applied to an autonomous inspection system of unmanned aerial vehicles based on machine vision, and comprises the following steps:

[0029] Obtaining the range of a preset UAV and geographic information of a preset area, as well as the coordinates of a location of a preset power device in the preset area and the equipment area of ​​the preset power device;

[0030] Collecting historical infrared data of preset power equipment, the infrared data including infrared image data, dividing the collected historical infrared image data, obtaining the changing rules of the infrared image data, and generating a prediction model;

[0031] Analyze the height difference between the preset power equipment and surrounding buildings based on the position coordinates of the preset power equipment in the preset area, and obtain the average height difference therebetween, and at the same time obtain the angle difference between different infrared image areas on the preset power equipment and surrounding buildings, and distinguish the surrounding buildings;

[0032] Draw and generate real-life three-dimensional inspection flight routes for different infrared image areas of the preset power equipment and surrounding buildings, and divide the real-life three-dimensional inspection flight routes into Wherein, n represents the nth inspection flight route, wherein an infrared image area is selected and preset as a reference area, and route features of different inspection flight routes and the reference area are obtained, and a route data model is generated at the same time;

[0033] According to the FSBi-RRT algorithm, the entire flight segment of each different inspection flight route is divided into at least three small route segments, wherein the control input of each small route segment is regarded as an independent variable, and the state variable is regarded as a dependent variable, and the image features of the reference area obtained by the different inspection flight routes are calculated and obtained, and the image features are uploaded to the route data model and updated at the same time;

[0034] Record the communication base station and / or communication base station operator that matches the preset UAV flight, and receive the field data uploaded by the collection monitoring point. When the preset UAV inspection flight in the preset area deviates from the preset inspection flight route, initiate an information acquisition request for the preset UAV inspection flight status to the communication base station and / or communication base station operator.

[0035] Beneficial effects:

[0036] The present invention utilizes deep learning analysis technology to preset different inspection flight routes for UAVs through the changing characteristics of infrared images of power equipment, and divides the preset inspection flight routes into multiple small route segments. At the same time, flight trajectory coordinate data can be collected during inspection flights in different small route segments, and the flight trajectory coordinate data is compared and analyzed with the original coordinate data without flight deviation to determine whether the flight altitude of the UAV deviates, and then adjust the flight trajectory altitude of the UAV. In the process of adjustment, the system can calculate the data obtained before and after the flight trajectory deviation occurs, and input the data difference into the preset flight control command of the UAV, so that the inspection flight of the UAV can return to the correct flight trajectory more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0038] Figure 1 Schematic diagram of the modular structure of the autonomous inspection system of a UAV based on machine vision in an embodiment of the present invention;

[0039] Figure 2 A schematic diagram of a method flow chart of an embodiment of the present invention;

[0040] Figure 3 A schematic diagram of an inspection flight trajectory according to an embodiment of the present invention;

[0041] Figure 4 A schematic diagram of a flight based on terrain ground height trajectory coordinates according to an embodiment of the present invention;

[0042] Numbers in the figure: 110 - comprehensive information acquisition module; 120 - information data acquisition module; 130 - image data processing unit; 1301 - image data compression module; 1302 - image capture module; 140 - route planning generation system; 1401 - analysis unit; 1402 - three-dimensional modeling unit; 1403 - calculation unit; 150 - monitoring terminal. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0044] Since the existing technologies all use the target coordinates as the main body for inspection route planning, however, the target coordinates, as an entity, have different infrared characteristics depending on the fault, so that their external visual states are also different. However, it is difficult for the existing technologies to select and automatically generate various inspection routes according to the different changing states of the target coordinates, which leads to hidden dangers in path optimization and ensuring the safety of the target coordinates.

[0045] Based on this, the present invention proposes a UAV autonomous inspection system and method based on machine vision. It uses deep learning analysis technology to preset different inspection flight routes for UAVs through the changing characteristics of infrared images of power equipment. During the UAV inspection flight, it can determine whether the UAV flight altitude deviates, and then adjust the UAV flight trajectory altitude, so that the UAV inspection flight can more accurately return to the correct flight trajectory.

[0046] The present invention is further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0047] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides a UAV autonomous inspection system based on machine vision, including:

[0048] The comprehensive information acquisition module 110 is used to acquire the range of the preset UAV and the geographical information of the preset area, and also includes a module for acquiring the coordinates of the location of the preset power equipment in the preset area and the equipment area of ​​the preset power equipment;

[0049] Furthermore, the geographical information includes the topography and altitude of the preset area;

[0050] The information data acquisition module 120 is used to collect historical infrared data of preset power equipment, the infrared data includes infrared image data, and divide the collected historical infrared image data, obtain the change law of the infrared image data, and generate a prediction model. The information data acquisition module includes a CCD camera;

[0051] It should be noted that in this embodiment, the historical infrared image data is divided into the following categories: θ1, θ2, ..., θ nDivide, where n represents the nth type of infrared image data, analyze the data features between different infrared image data, upload the data features to the prediction model, and update the prediction model;

[0052] In reality, infrared thermal images will be different due to temperature differences inside the power equipment, so using temperature as one of the reference factors can facilitate judgment;

[0053] Further, according to seasonal changes and / or changes in power consumption in a preset area, change characteristics of infrared images of preset power equipment from the central area to the edge area are collected, and the change characteristics are divided into at least 6 change circles, and the inspection flight route of the preset UAV is planned according to the infrared image characteristics of different change circles that need to be collected. When the infrared image collected by the preset UAV after arriving at the collection point does not meet the infrared image characteristics of different change circles that need to be collected, the system determines that the inspection flight route of the preset UAV is an invalid route, otherwise, it will not be determined;

[0054] It should be noted that the change characteristics of the infrared image of the preset power equipment from the central area to the edge area include color tone change characteristics;

[0055] The present embodiment requires to emphasize that the area occupied by different change circles of the infrared image is analyzed, and the area size of the infrared image that can be taken by the preset drone at different flight altitudes is obtained, and the area occupied by the different change circles is divided into ω1, ω2, ..., ω n Divide it into categories, where n represents the area occupied by the nth category, and calculate the safe distance between different occupied areas and surrounding buildings according to the following formula:

[0056] Where t represents the historical time when the occupied area changes the most in different seasons and / or under different electricity consumption conditions;

[0057] In the formula, ω vj Indicates the jth spacing size data measured for the vth collection point in different change circles, χ is the independent variable, which represents the change function of the area occupied by different change circles, and presets the safety threshold based on the measured minimum spacing data. When the drone is preset to collect infrared images at the end position of the selected inspection flight route in the future period, if the distance between the end position and the surrounding buildings is lower than the safety threshold, the system determines that the selected inspection flight route is a dangerous route, and the system automatically deletes the route, otherwise, it will not be determined;

[0058] On the basis of the above, this embodiment specifically counts the total number of dangerous routes, and divides different routes into at least 5 small route segments in the total number of routes, obtains the flight trajectory coordinate data of different small route segments, and records the height between the initial flight path point of the preset drone and the reference terrain ground of the preset area as m. i o , and according to the height m i o Calculate the data difference between it and the flight trajectory coordinate data of different small route segments, and calculate it according to the following formula:

[0059] Among them, m i represents the i-th height value, and o represents the trajectory coordinate corresponding to the height value;

[0060] In the formula, e i represents the flight trajectory coordinate data of the i-th small route segment collected, ρ represents the distance value, and the distance value is the distance value of the small route segment collected;

[0061] The image data processing unit 130 is used to respond to the infrared image data collected by the information data collection module 120, including an image data compression module 1301 for compressing and processing the infrared image data, and an image capture module 1302 for capturing image data of a preselected area in the image data compression module 1301;

[0062] The route planning generation system 140, the route planning generation unit includes an analysis unit 1401, a three-dimensional modeling unit 1402 and a calculation unit 1403;

[0063] The analysis unit 1401 is used to analyze the height difference between the preset power equipment and the surrounding buildings based on the position coordinates of the preset power equipment in the preset area, and obtain the average height difference therein, and is also used to obtain the angle difference between different infrared image areas on the preset power equipment and the surrounding buildings, and distinguish the surrounding buildings;

[0064] It should be noted that this embodiment includes analyzing the area of ​​different angle differences, distinguishing the surrounding buildings into primary buildings and secondary buildings, and analyzing the angle differences formed based on the primary buildings, while the surrounding buildings include trees;

[0065] The 3D modeling unit 1402 is used to draw and generate real-life 3D inspection flight routes for different infrared image areas of preset power equipment and surrounding buildings, and divide the real-life 3D inspection flight routes into Wherein, n represents the nth inspection flight route, wherein an infrared image area is selected and preset as a reference area, and route features of different inspection flight routes and the reference area are obtained, and a route data model is generated at the same time;

[0066] The calculation unit 1403 is used to divide the entire flight segment of each different inspection flight route into at least three small route segments according to the FSBi-RRT algorithm, wherein the control input of each small route segment is regarded as an independent variable, and the state variable is regarded as a dependent variable, and the image features of the reference area obtained by the different inspection flight routes are calculated and obtained, and the image features are uploaded to the route data model, and updated at the same time;

[0067] The monitoring terminal 150 is used to record the communication base station and / or the communication base station operator that matches the preset UAV flight, receive the field data uploaded by the collection monitoring point, and update the background data at the same time; when the inspection flight of the preset UAV in the preset area deviates from the preset inspection flight route, a request for obtaining information on the inspection flight status of the preset UAV is initiated to the communication base station and / or the communication base station operator;

[0068] Furthermore, a patrol flight route is preset according to the infrared image data of different change circles that need to be obtained, a patrol flight altitude value of a given reference terrain ground is given in the patrol flight route, and a critical threshold is preset based on the patrol flight altitude value. When the actual flight altitude of the preset UAV is lower than and / or exceeds the critical threshold, the system determines that a deviation has occurred in the patrol flight, and issues an early warning to the communication base station and / or the communication base station operator that matches the preset UAV flight;

[0069] On the above basis, the changing characteristics of the flight altitude during the inspection flight of the preset UAV are collected based on the critical threshold, the altitude value corresponding to the changing characteristics is marked at least three times, and the data characteristics of each marked value are analyzed to obtain the increase and / or decrease value of the marked data at the critical threshold. When the actual flight altitude of the preset UAV is lower than and / or exceeds the critical threshold, the system automatically inputs the obtained increase and / or decrease value into the flight control instruction of the preset UAV, otherwise it is not input.

[0070] The acquired value increase and / or value decrease data are divided into several evaluation indicators, the evaluation indicator data during the inspection flight process are statistically calculated, and the data are normalized to calculate the weight of each evaluation indicator.

[0071] Based on the above, this application uses deep learning analysis technology to preset different inspection flight routes for drones through the changing characteristics of infrared images of power equipment. During the drone inspection flight, it can determine whether the drone's flight altitude deviates, and then adjust the drone's flight trajectory altitude, so that the drone's inspection flight can more accurately return to the correct flight trajectory.

[0072] This embodiment combines the above-mentioned UAV autonomous inspection system based on machine vision and also proposes a working method of the system, as follows:

[0073] Obtaining the flight range of the preset UAV and the geographical information of the preset area, the coordinates of the location of the preset power equipment in the preset area and the equipment area of ​​the preset power equipment;

[0074] Collect historical infrared data of preset power equipment, the infrared data includes infrared image data, divide the collected historical infrared image data, obtain the change law of the infrared image data, and generate a prediction model;

[0075] Analyze the height difference between the preset power equipment and surrounding buildings based on the position coordinates of the preset power equipment in the preset area, and obtain the average height difference therein, and at the same time obtain the angle difference between different infrared image areas on the preset power equipment and surrounding buildings, and distinguish the surrounding buildings;

[0076] Draw and generate real-life 3D inspection flight routes for different infrared image areas of preset power equipment and surrounding buildings, and divide the real-life 3D inspection flight routes into Wherein, n represents the nth inspection flight route, wherein an infrared image area is selected and preset as a reference area, and route features of different inspection flight routes and the reference area are obtained, and a route data model is generated at the same time;

[0077] According to the FSBi-RRT algorithm, the entire flight section of each different inspection flight route is divided into at least three small route segments, wherein the control input of each small route segment is regarded as an independent variable, and the state variable is regarded as a dependent variable, and the image features of the reference area obtained by the different inspection flight routes are calculated and obtained, and the image features are uploaded to the route data model and updated at the same time;

[0078] Record the communication base station and / or communication base station operator that matches the preset UAV flight, and receive the field data uploaded by the collection monitoring point. When the preset UAV inspection flight in the preset area deviates from the preset inspection flight route, initiate an information acquisition request for the preset UAV inspection flight status to the communication base station and / or communication base station operator.

[0079] In summary, the present invention utilizes deep learning analysis technology to preset different inspection flight routes for UAVs through the changing characteristics of infrared images of power equipment, and divides the preset inspection flight routes into multiple small route segments. At the same time, flight trajectory coordinate data can be collected during inspection flights in different small route segments, and the flight trajectory coordinate data is compared and analyzed with the original coordinate data without flight deviation to determine whether the flight altitude of the UAV deviates, and then adjust the flight trajectory altitude of the UAV. In the process of adjustment, the system can calculate the data obtained before and after the flight trajectory deviation occurs, and input the data difference into the preset flight control instructions of the UAV, so that the inspection flight of the UAV can more accurately return to the correct flight trajectory.

[0080] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A UAV autonomous inspection system based on machine vision, It is characterized in that include: A comprehensive information acquisition module, used to acquire the range of a preset UAV and geographic information of a preset area, and also includes a module for acquiring the coordinates of a preset power device in a preset area and the equipment area of ​​the preset power device; An information data acquisition module is used to collect historical infrared data of preset power equipment, the infrared data includes infrared image data, divide the collected historical infrared image data, obtain the change law of the infrared image data, and generate a prediction model; An image data processing unit, the image data processing unit is used to respond to the infrared image data collected by the information data collection module, including an image data compression module for compressing and processing the infrared image data, and an image interception module for intercepting image data of a preselected area in the image data compression module; A route planning generation system, wherein the route planning generation unit includes an analysis unit, a three-dimensional modeling unit and a calculation unit; The analysis unit is used to analyze the height difference between the preset power equipment and the surrounding buildings based on the position coordinates of the preset power equipment in the preset area, and obtain the average height difference therein, and is also used to obtain the angle difference between different infrared image areas on the preset power equipment and the surrounding buildings, and distinguish the surrounding buildings; The three-dimensional modeling unit is used to draw and generate real-life three-dimensional inspection flight routes of different infrared image areas and surrounding buildings on the preset power equipment, and divide the real-life three-dimensional inspection flight routes into Wherein, n represents the nth inspection flight route, wherein an infrared image area is selected and preset as a reference area, and route features of different inspection flight routes and the reference area are obtained, and a route data model is generated at the same time; The calculation unit is used to divide the entire flight segment of each different inspection flight route into at least three small route segments according to the FSBi-RRT algorithm, wherein the control input of each small route segment is regarded as an independent variable, and the state variable is regarded as a dependent variable, and the image features of the reference area obtained by the different inspection flight routes are calculated and obtained, and the image features are uploaded to the route data model, and updated at the same time; The monitoring terminal is used to record the communication base station and / or communication base station operator that matches the preset UAV flight, receive the field data uploaded by the collection monitoring point, and update the background data at the same time; when the preset UAV inspection flight in the preset area deviates from the preset inspection flight route, a request for obtaining information on the preset UAV inspection flight status is initiated to the communication base station and / or communication base station operator.

2. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 1, It is characterized in that In the information data acquisition module, the historical infrared image data is divided, including dividing the historical infrared image data according to the seasonal changes of the preset area and / or the changes in the power consumption of the preset area according to θ 1 ,θ 2 , ..., θ n The prediction model is divided into the following types: n represents the nth type of infrared image data; data features between different infrared image data are analyzed; the data features are uploaded to the prediction model; and the prediction model is updated.

3. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 2, It is characterized in that According to seasonal changes and / or changes in electricity consumption in a preset area, change characteristics of infrared images of preset power equipment from the central area to the edge area are collected, and the change characteristics are divided into at least 6 change circles. The inspection flight route of the preset UAV is planned according to the infrared image characteristics of different change circles that need to be collected. When the infrared image collected by the preset UAV after arriving at the collection point does not conform to the infrared image characteristics of the different change circles that need to be collected, the system determines that the inspection flight route of the preset UAV is an invalid route, otherwise, it will not be determined.

4. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 3, It is characterized in that Analyze the area occupied by different change circles of the infrared image, and obtain the area size of the infrared image that can be taken by the preset drone at different flight altitudes. 1 ,ω 2 , ..., ω n Divide it into categories, where n represents the area occupied by the nth category, and calculate the safe distance between different occupied areas and surrounding buildings, which can be calculated according to the following formula: Where t represents the historical time when the occupied area changes the most in different seasons and / or under different electricity consumption conditions; In the formula, ω v j represents the jth spacing size data measured for the vth collection point in the different change circles, χ is the independent variable, which represents the change function of the area occupied by different change circles, and a safety threshold is preset based on the measured minimum spacing data. When the drone is preset to collect infrared images at the end position of the selected inspection flight route in the future period, if the distance between the end position and the surrounding buildings is lower than the safety threshold, the system determines that the selected inspection flight route is a dangerous route, and the system automatically deletes the route, otherwise, it will not be determined.

5. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 4, It is characterized in that Count the total number of dangerous routes, and divide different routes into at least 5 small route segments in the total number of routes, obtain the flight trajectory coordinate data of different small route segments, and record the height between the initial flight path point of the preset drone and the reference terrain ground of the preset area as m i o , and according to the height m i o Calculate the data difference between it and the flight trajectory coordinate data of different small route segments, and calculate it according to the following formula: Among them, m i represents the ith height value, and o represents the trajectory coordinate corresponding to the height value; In the formula, e i represents the collected flight trajectory coordinate data of the ith small route segment, and ρ represents the distance value, which is the distance value of the collected small route segment.

6. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 5, It is characterized in that An inspection flight route is preset based on the infrared image data of the different changing circles that need to be obtained, an inspection flight altitude value of the reference terrain ground is given in the inspection flight route, and a critical threshold is preset based on the inspection flight altitude value. When the actual flight altitude of the preset UAV is lower than and / or exceeds the critical threshold, the system determines that a deviation has occurred in the inspection flight, and issues an early warning to the communication base station and / or communication base station operator that matches the preset UAV flight.

7. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 6, It is characterized in that Based on the critical threshold, the change characteristics of the flight altitude of the preset UAV during the inspection flight are collected, the altitude value corresponding to the change characteristics is marked at least three times, and the data characteristics of each marked value are analyzed to obtain the increase and / or decrease value of the marked data at the critical threshold. When the actual flight altitude of the preset UAV is lower than and / or exceeds the critical threshold, the system automatically inputs the obtained increase and / or decrease value into the flight control instruction of the preset UAV, otherwise, it is not input.

8. The autonomous inspection system for unmanned aerial vehicles based on machine vision as claimed in claim 7, It is characterized in that The acquired increase value and / or decrease value data are divided into a number of evaluation indicators, the evaluation indicator data during the inspection flight process are statistically calculated, and the data are normalized to calculate the weight of each evaluation indicator.

9. A method for autonomous inspection of unmanned aerial vehicles based on machine vision, applied to the autonomous inspection system of unmanned aerial vehicles based on machine vision as claimed in claim 1, It is characterized in that The following steps are involved: Obtaining the range of a preset UAV and geographic information of a preset area, as well as the coordinates of a location of a preset power device in the preset area and the equipment area of ​​the preset power device; Collecting historical infrared data of preset power equipment, the infrared data including infrared image data, dividing the collected historical infrared image data, obtaining the changing rules of the infrared image data, and generating a prediction model; Analyze the height difference between the preset power equipment and surrounding buildings based on the position coordinates of the preset power equipment in the preset area, and obtain the average height difference therebetween, and at the same time obtain the angle difference between different infrared image areas on the preset power equipment and surrounding buildings, and distinguish the surrounding buildings; Draw and generate real-life three-dimensional inspection flight routes for different infrared image areas of the preset power equipment and surrounding buildings, and divide the real-life three-dimensional inspection flight routes into Wherein, n represents the nth inspection flight route, wherein an infrared image area is selected and preset as a reference area, and route features of different inspection flight routes and the reference area are obtained, and a route data model is generated at the same time; According to the FSBi-RRT algorithm, the entire flight segment of each different inspection flight route is divided into at least three small route segments, wherein the control input of each small route segment is regarded as an independent variable, and the state variable is regarded as a dependent variable, and the image features of the reference area obtained by the different inspection flight routes are calculated and obtained, and the image features are uploaded to the route data model and updated at the same time; Record the communication base station and / or communication base station operator that matches the preset UAV flight, and receive the field data uploaded by the collection monitoring point. When the preset UAV inspection flight in the preset area deviates from the preset inspection flight route, initiate an information acquisition request for the preset UAV inspection flight status to the communication base station and / or communication base station operator.

Citation Information

Cited By

  • Logistics inspection unmanned aerial vehicle based on artificial intelligence

    CN119940867A

  • Unmanned aerial vehicle inspection planning method and system for power fault feature perception

    CN120803020A

  • Intelligent inspection system and method for power system based on machine vision

    CN121033711A

  • A machine vision-based intelligent inspection system and method for power systems

    CN121033711B

  • Unmanned aerial vehicle autonomous inspection operation monitoring and analysis system

    CN121330378A