A method and system for railway line condition detection based on unmanned aerial vehicle (UAV) intelligent monitoring

Abstract

This invention relates to the field of railway line condition detection technology, and discloses a railway line condition detection method and system based on UAV intelligent monitoring. A track parameter encoding set is constructed using a first monitoring UAV. Considering that each target track parameter code obtained by the second monitoring UAV during cruise detection has uniqueness requirements and monitoring time period requirements, the second monitoring UAV is controlled to perform cruise detection in the corresponding railway fault section. The position corresponding to the current orthophoto track image is located through the target track parameter code, improving the accuracy of condition detection for mountainous railway tracks. The initial track parameter encoding set is constructed using a first monitoring UAV equipped with satellite positioning and inertial navigation equipment, and the image acquisition and positioning for railway line condition detection is performed using a second monitoring UAV equipped only with satellite positioning equipment. This significantly reduces operating costs in engineering applications.

Description

Technical Field

[0001] This invention relates to the field of railway line condition monitoring technology, and in particular to a railway line condition monitoring method and system based on unmanned aerial vehicle (UAV) intelligent monitoring. Background Technology

[0002] Railway line condition monitoring is of great significance for ensuring safe railway operation and improving transportation efficiency. It can promptly detect defects and damage to the lines, ensuring the safety of passengers and goods. Furthermore, regular inspections allow for proactive maintenance and upkeep, reducing train delays and service disruptions caused by line faults and ensuring the efficiency and punctuality of railway transportation. In addition, accurate understanding of the line's condition enables the rational allocation of maintenance plans and resources, avoiding over- or under-maintenance, extending the service life of lines and equipment, and reducing long-term maintenance costs.

[0003] Currently, railway line condition inspections typically employ methods such as manual inspection, track inspection vehicles, and drone-based intelligent inspection. Manual inspections primarily involve staff walking or riding in track vehicles, observing and inspecting the railway line with the naked eye or simple tools. While this can directly identify some obvious line problems, it is inefficient, labor-intensive, and prone to overlooking hidden issues. Track inspection vehicles utilize specialized vehicles equipped with various inspection devices to quickly and accurately detect parameters such as track geometry and rail damage. However, the equipment is costly and has limited adaptability to complex environments. Drone-based intelligent inspection, on the other hand, leverages the flexibility of drones to quickly reach areas inaccessible to humans. Furthermore, the ability to flexibly adjust flight paths and altitudes improves inspection efficiency. Due to its flexibility, efficiency, and accuracy, drone-based intelligent inspection has become a popular method for railway line condition monitoring.

[0004] However, in practical applications, the use of UAV intelligent detection methods in certain scenarios (such as condition detection of railway tracks in mountainous areas) still has limitations. Specifically, mountainous terrain is highly undulating. When UAVs fly into valleys or areas close to mountains, the dense and tall trees and shrubs in the mountains may block satellite signals, causing the satellite signals received by the UAVs to weaken or be lost. This results in low positioning accuracy of the railway track images captured by the UAVs. When performing single-point periodic monitoring of high-probability fault points, inaccurate positioning can easily lead to monitoring position deviations, making it impossible to detect the condition of the railway line in a timely and accurate manner, thus posing potential risks to railway line maintenance and safety management. Although increasing the flight altitude of UAVs can avoid the impact of mountainous terrain and vegetation on satellite positioning accuracy, higher flight altitudes mean that UAVs need to be equipped with more precise image acquisition equipment, which increases detection costs and also affects the accuracy of image analysis-based railway condition detection methods. Meanwhile, some drones are equipped with high-precision satellite positioning and inertial navigation systems, which can switch to inertial navigation when the drone's satellite positioning signal is weak. This can mitigate the impact of positioning accuracy on railway line inspection to some extent. However, drones equipped with high-precision satellite positioning and inertial navigation systems significantly increase operating costs. Currently, industrial-grade drones with both GPS and inertial navigation systems cost between tens of thousands and hundreds of thousands of yuan, while industrial-grade drones with only GPS positioning typically cost between several thousand and tens of thousands of yuan. Therefore, in practical engineering applications, increasing the flight altitude of drones and equipping them with high-precision satellite positioning and inertial navigation systems presents significant feasibility drawbacks.

[0005] Therefore, how to improve the accuracy of condition detection of railway tracks in mountainous areas, enhance detection efficiency and positioning accuracy, and reduce operating costs in engineering applications is a technical problem that urgently needs to be solved. Summary of the Invention

[0006] The main objective of this invention is to provide a railway line condition detection method and system based on unmanned aerial vehicle (UAV) intelligent monitoring, aiming to solve at least one of the above-mentioned technical problems.

[0007] To achieve the above objectives, this invention provides a railway line condition detection method based on unmanned aerial vehicle (UAV) intelligent monitoring, comprising the following steps:

[0008] The first monitoring drone is controlled to collect track layout information of each transport railway within the target area and construct a track parameter code set for the target area; wherein, the first monitoring drone is configured to have at least a satellite positioning device and an inertial navigation device;

[0009] Call the railway status association parameter library of the target area and use the railway status prediction model to predict the section location and fault type set of several railway fault sections in each transport railway;

[0010] The status information of several distributed second monitoring drones within the target area is queried. Based on the track parameter encoding set and the section location of each railway fault section, the status detection line of each second monitoring drone during the monitoring period is determined.

[0011] Based on the status detection line and the fault type set, generate status detection control instructions for each second monitoring UAV during the monitoring period;

[0012] Based on the status detection control command, each second monitoring drone is controlled to perform status detection of the transport railway according to the corresponding status detection line during the monitoring period, and to obtain the status detection result; wherein, the second monitoring drone is configured to have at least a satellite positioning device.

[0013] Optionally, the step of controlling the first monitoring drone to collect track layout information for each transport railway within the target area specifically includes:

[0014] Control the first monitoring drone to perform railway patrols along each transport railway within the target area, and obtain several orthophoto images of each transport railway.

[0015] Image analysis is performed on several orthophoto tracks of each transport railway to identify the track contour in each orthophoto track image and extract the track parameters of each transport railway at each monitoring point; wherein, the track parameters include the angle between the track tangent direction and the target direction at that monitoring point;

[0016] The track parameters of each transport railway at each monitoring point within the target area are used as track layout information; wherein any two adjacent monitoring points are configured to have the same first interval distance on the transport railway.

[0017] Optionally, the step of controlling the first monitoring drone to perform a railway patrol along each transport railway within the target area and obtaining several orthophoto images of each transport railway specifically includes:

[0018] Randomly select one of the several transport railways in the target area as the target transport railway, and control the first monitoring drone to head to the cruise starting point of the target transport railway;

[0019] The first monitoring drone collects an orthophoto of the starting point of the cruise. Based on the direction of the transport railway identified in the orthophoto, the cruise direction of the first monitoring drone is controlled until the cruise endpoint of the target transport railway is reached.

[0020] When acquiring each orthophoto image of the target transport railway, the first monitoring UAV is simultaneously acquired based on its satellite positioning and inertial navigation equipment, and its current cruising position is used as additional information for the corresponding orthophoto image.

[0021] After completing the patrol of a transport railway, it is removed from the list of transport railways in the target area, and a new transport railway is randomly selected as the target transport railway for patrol, until all transport railways in the target area have completed their patrols.

[0022] Optionally, the step of constructing the orbital parameter encoding set for the target region specifically includes:

[0023] Obtain the track parameters and cruise sequence of several monitoring points for each transport railway in the target area from the cruise start point to the cruise end point;

[0024] Based on the patrol sequence of the monitoring points, the track parameters of several monitoring points are numerically encoded in the direction from the start point to the end point of the patrol, thereby obtaining the track parameter encoding sequence of each transport railway in the target area and constructing the track parameter encoding set of the target area.

[0025] Optionally, the steps include calling the railway status association parameter library for the target area and using the railway status prediction model to predict the location and fault type set of several railway fault sections in each transport railway, specifically including:

[0026] Call the railway status association parameter library of the target area, and extract the railway status association parameter prediction set for each monitoring point of each transport railway in the monitoring period.

[0027] The railway condition correlation parameter prediction set includes predicted values ​​of several railway condition correlation parameters, which include at least one or more of climate correlation parameters, geological correlation parameters, and railway track load parameters.

[0028] Using the railway condition prediction model, with the railway condition correlation parameter prediction set as the prediction input sample, several fault monitoring points and fault types in each transport railway are predicted.

[0029] The railway status prediction model is configured to train an initially constructed convolutional neural network using training samples constructed from railway status association parameters corresponding to different types of railway historical faults in the railway historical fault database.

[0030] The section between two adjacent fault monitoring points whose interval distance is less than the second interval distance is regarded as the same railway fault section, and the fault types of several fault monitoring points in each railway fault section are summarized as the fault type set of the railway fault section.

[0031] Optionally, query the status information of several distributed second monitoring drones within the target area, and determine the status detection route steps for each second monitoring drone during the monitoring period based on the track parameter encoding set and the section location of each railway fault section. Specifically, this includes:

[0032] The status information of several distributed second monitoring drones within the target area is queried; wherein, the status information includes the cruise waiting position, cruise speed range, and orthophoto trajectory image processing speed of each second monitoring drone;

[0033] Based on the track parameter encoding set and the location of each railway fault section, several second monitoring drones are assigned to the corresponding railway fault sections, and the status detection line of each second monitoring drone during the monitoring period is determined.

[0034] Optionally, the step of querying the status information of several distributed second monitoring drones within the target area includes:

[0035] The detection equipment signals transmitted by several railway line status detection nodes distributed within the target area are obtained; wherein, the detection equipment signals include node identifiers and UAV identifiers of the second monitoring UAVs equipped for performing railway line status detection during the current monitoring period;

[0036] Based on the node identifier, query the node location information of each railway line status detection node, and use the node location information as the cruise waiting position of the corresponding second monitoring drone;

[0037] Based on the UAV identifier, query the cruising speed range and orthogonal orbit image processing speed of each second monitoring UAV; wherein, the cruising speed range is configured as a speed range generated according to the flight parameters provided by the UAV manufacturer and pre-written into the UAV status database, and the orthogonal orbit image processing speed is configured as the number of orthogonal orbit images processed per unit time obtained by the second monitoring UAV when performing the test task of extracting the orbit parameters of each monitoring point through orthogonal orbit image analysis.

[0038] Optionally, based on the track parameter encoding set and the location of each railway fault section, several second monitoring drones are assigned to the corresponding railway fault sections, and the status detection line steps for each second monitoring drone during the monitoring period are determined, specifically including:

[0039] Based on the location of each railway fault section, the cruise waiting position and cruise speed range of each second monitoring drone, calculate the time interval for each second monitoring drone to complete the status detection of different railway fault sections or combinations of different railway fault sections in the second monitoring drone allocation scheme, and eliminate the second monitoring drone allocation scheme whose minimum time interval exceeds the monitoring period.

[0040] The time interval includes the first time taken for the second monitoring drone to fly at its maximum cruising speed to the starting position of different railway fault sections or combinations of different railway fault sections, and the sum of several second times taken for the second monitoring drone to perform state detection of different railway fault sections or combinations of different railway fault sections at different cruising speeds within the cruising speed range.

[0041] In the remaining second monitoring drone allocation scheme, when each second monitoring drone allocation scheme performs state detection of each railway fault section using each cruise speed value within the corresponding cruise speed range, it determines the section track parameter code of each railway fault section in the track parameter coding sequence of the railway to which it belongs based on the third interval distance between two adjacent selected monitoring points determined by the orthophoto track image processing speed and the cruise speed value. It then determines whether the target track parameter code obtained by performing the parameter coding bit-by-bit extraction action of the target sampling window according to the third interval distance meets the coding positioning accuracy requirements.

[0042] Among them, the coding positioning accuracy requirement is configured such that each target track parameter code obtained by bit-by-bit extraction of parameter coding according to the third interval distance in the segment track parameter coding is different;

[0043] Based on the judgment results of the coding positioning accuracy requirements, the remaining second monitoring UAV allocation schemes are removed from the cruise speed range of each railway fault section where the cruise speed value that does not meet the coding positioning accuracy requirements and the cruise speed value that takes longer to perform status detection at the corresponding cruise speed for each railway fault section or combination of railway fault sections than the monitoring period. This results in the railway fault section allocation set for each second monitoring UAV when performing cruise task allocation.

[0044] The railway fault section allocation set includes several candidate railway fault sections or combinations of railway fault sections for each second monitoring UAV, as well as the cruise speed range adopted by the second monitoring UAV in each railway fault section or the combination of cruise speed ranges adopted by different railway fault sections in the combination of railway fault sections.

[0045] From the railway fault section allocation set of each second monitoring UAV, select the railway fault section allocation strategy of each second monitoring UAV that covers all railway fault sections, and generate the status detection line of each second monitoring UAV during the monitoring period.

[0046] Optionally, the step of generating status detection control instructions for each second monitoring UAV during the monitoring period, based on the status detection line and the fault type set, specifically includes:

[0047] Based on the status detection line of each second monitoring UAV and the cruise altitude requirements corresponding to the fault type of each railway fault section, a spatial cruise route for each second monitoring UAV is generated.

[0048] Based on the cruising speed range of each railway fault section corresponding to the space cruise route and the status detection line, status detection control commands are generated for each second monitoring UAV during the monitoring period.

[0049] Furthermore, to achieve the above objectives, the present invention also provides a railway line condition detection system based on unmanned aerial vehicle (UAV) intelligent monitoring, comprising:

[0050] The control module is used to control the first monitoring drone to collect track layout information of each transport railway within the target area and construct a track parameter code set for the target area; wherein, the first monitoring drone is configured to have at least a satellite positioning device and an inertial navigation device;

[0051] The prediction module is used to call the railway status association parameter library of the target area and use the railway status prediction model to predict the section location and fault type set of several railway fault sections in each transport railway.

[0052] The query module is used to query the status information of several distributed second monitoring drones within the target area. Based on the track parameter encoding set and the section location of each railway fault section, the status detection line of each second monitoring drone during the monitoring period is determined.

[0053] The generation module is used to generate status detection control instructions for each second monitoring UAV during the monitoring period based on the status detection line and the fault type set;

[0054] The detection module is used to control each second monitoring drone to perform status detection of the transport railway according to the corresponding status detection line during the monitoring period based on the status detection control command, and to obtain the status detection result; wherein, the second monitoring drone is configured to have at least a satellite positioning device.

[0055] The beneficial effects of this invention are as follows: It proposes a railway line condition detection method and system based on UAV intelligent monitoring. By analyzing the target track parameters of the orthorectified track image collected during cruise detection and determining the target track parameter encoding, the location information corresponding to the current orthorectified track image is determined, realizing the mapping between railway line condition detection results and location information, improving the accuracy of railway track condition detection in mountainous areas, and improving detection efficiency and positioning accuracy. At the same time, it utilizes a first monitoring UAV equipped with satellite positioning equipment and inertial navigation equipment to perform the initial track parameter encoding set construction, and then uses a second monitoring UAV equipped only with satellite positioning equipment to perform image acquisition and positioning for railway line condition detection, which can significantly reduce the operating cost in engineering applications. Attached Figure Description

[0056] Figure 1 This is a flowchart of the railway line condition detection method based on UAV intelligent monitoring according to the present invention;

[0057] Figure 2 This is a structural diagram of the railway line condition detection system based on UAV intelligent monitoring according to the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0059] This invention provides a method for railway line status detection based on unmanned aerial vehicle (UAV) intelligent monitoring, referring to... Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the railway line condition detection method based on unmanned aerial vehicle (UAV) intelligent monitoring according to the present invention.

[0060] In this embodiment, a railway line status detection method based on UAV intelligent monitoring includes the following steps:

[0061] S100: Control the first monitoring drone to collect track layout information of each transport railway within the target area and construct a track parameter code set for the target area; wherein, the first monitoring drone is configured to have at least a satellite positioning device and an inertial navigation device;

[0062] S200: Call the railway status association parameter library of the target area, and use the railway status prediction model to predict the section location and fault type set of several railway fault sections in each transport railway;

[0063] S300: Query the status information of several distributed second monitoring drones within the target area, and determine the status detection route of each second monitoring drone during the monitoring period based on the track parameter encoding set and the section location of each railway fault section;

[0064] S400: Generate status detection control instructions for each second monitoring UAV during the monitoring period based on the status detection line and the fault type set;

[0065] S500: Based on the state detection control command, control each second monitoring drone to perform state detection of the transport railway according to the corresponding state detection line during the monitoring period, and obtain state detection results; wherein, the second monitoring drone is configured to have at least a satellite positioning device.

[0066] It should be noted that in practical applications, the use of UAV intelligent detection methods still has limitations in some scenarios (such as condition detection of railway tracks in mountainous areas). Specifically, mountainous terrain is highly undulating. When UAVs fly into valleys or areas close to mountains, the dense and tall trees and shrubs in the mountains may block satellite signals, causing the satellite signals received by the UAV to weaken or be lost. This results in low positioning accuracy of the railway track images captured by the UAV. When performing single-point periodic monitoring of high-probability fault points, inaccurate positioning can easily lead to monitoring position deviation, making it impossible to detect the condition of the railway line in a timely and accurate manner, thus posing a hidden danger to railway line maintenance and safety management. Although increasing the flight altitude of UAVs can avoid the impact of mountainous terrain and vegetation on satellite positioning accuracy, higher flight altitudes mean that more precise image acquisition equipment is required for the UAVs. This increases detection costs and also affects the accuracy of image analysis-based railway condition detection methods. Meanwhile, some drones are equipped with high-precision satellite positioning and inertial navigation systems, which can switch to inertial navigation when the drone's satellite positioning signal is weak. This can mitigate the impact of positioning accuracy on railway line inspection to some extent. However, drones equipped with high-precision satellite positioning and inertial navigation systems significantly increase operating costs. Currently, industrial-grade drones with both GPS and inertial navigation systems cost between tens of thousands and hundreds of thousands of yuan, while industrial-grade drones with only GPS positioning typically cost between several thousand and tens of thousands of yuan. Therefore, in practical engineering applications, increasing the flight altitude of drones and equipping them with high-precision satellite positioning and inertial navigation systems presents significant feasibility drawbacks.

[0067] To address the aforementioned issues, this embodiment constructs a track parameter encoding set using a first monitoring drone. Considering that each target track parameter code obtained by the second monitoring drone during its cruise detection has uniqueness and monitoring time requirements, the second monitoring drone is controlled to perform cruise detection in the corresponding railway fault section. The target track parameter code is used to locate the position corresponding to the current orthophoto track image, improving the accuracy of track condition detection in mountainous areas. The first monitoring drone, equipped with satellite positioning and inertial navigation equipment, performs the initial track parameter encoding set construction, and the second monitoring drone, equipped only with satellite positioning equipment, performs image acquisition and positioning for railway line condition detection, which can significantly reduce the operating costs in engineering applications.

[0068] In a preferred embodiment, the step of controlling the first monitoring drone to collect track layout information for each transport railway within the target area specifically includes:

[0069] S110: Control the first monitoring drone to perform railway patrol along each transport railway within the target area and obtain several orthophoto images of each transport railway.

[0070] S120: Perform image analysis on several orthophoto images of each transport railway, identify the track contour in each orthophoto image, and extract the track parameters of each transport railway at each monitoring point; wherein, the track parameters include the angle between the track tangent direction and the target direction at that monitoring point;

[0071] S130: The track parameters of each transport railway within the target area at each monitoring point are used as track layout information; wherein any two adjacent monitoring points are configured to have the same first interval distance on the transport railway.

[0072] In this embodiment, the track layout information of each transport railway is collected by controlling the first monitoring drone to patrol along each transport railway in the target area, collecting each orthophoto track image collected during the patrol, and then extracting the track contour through image recognition and analysis technology, and then calculating the angle between the track tangent and the target direction at each monitoring point, and using the track parameters of each transport railway at each monitoring point as track layout information.

[0073] In a preferred embodiment, the step of controlling the first monitoring drone to perform railway patrol along each transport railway within the target area and obtain several orthophoto images of each transport railway specifically includes:

[0074] S111: Randomly select one of the several transport railways in the target area as the target transport railway, and control the first monitoring drone to go to the cruise starting point of the target transport railway;

[0075] S112: The first monitoring drone collects an orthophoto image of the starting point of the cruise. Based on the direction of the transport railway identified in the orthophoto image, the cruise direction of the first monitoring drone is controlled until the cruise endpoint of the target transport railway is reached.

[0076] S113: When collecting each orthophoto image of the target transport railway, the first monitoring UAV is simultaneously collected at its current cruising position based on the satellite positioning and inertial navigation equipment configured on the first monitoring UAV, and the cruising position is used as additional information for the corresponding orthophoto image.

[0077] S114: After completing the patrol of a transport railway, remove it from the list of transport railways in the target area, and randomly select another transport railway as the target transport railway to perform the patrol, until all the transport railways in the target area have completed the patrol.

[0078] In this embodiment, when controlling the first monitoring drone to cruise along the railway line, the drone's cruise direction can be adjusted based on the changes in the railway's direction along the line as captured in the orthorectified orbit images at each moment, ensuring that the railway remains within the image during the cruise. Simultaneously, the satellite positioning equipment of the first monitoring drone adds additional positioning information to each orthorectified orbit image. When the satellite positioning signal is unstable or weak, the drone switches to inertial navigation equipment to collect and add additional positioning information until the railway cruise is complete, thus obtaining several orthorectified orbit images for each railway line. Afterward, by analyzing the additional positioning information in each orthorectified orbit image, the orbital parameters of several monitoring points spaced at a first interval length along each railway line can be extracted.

[0079] In a preferred embodiment, the step of constructing the orbital parameter encoding set for the target region specifically includes:

[0080] S140: Obtain the track parameters and cruise sequence of several monitoring points for each transport railway in the target area from the cruise start point to the cruise end point;

[0081] S150: Based on the cruise sequence of the monitoring points, the track parameters of several monitoring points are numerically encoded in the direction from the cruise start point to the cruise end point to obtain the track parameter encoding sequence of each transport railway in the target area, and to construct the track parameter encoding set of the target area.

[0082] In this embodiment, after obtaining the track parameters of several monitoring points for each transport railway, the track parameter values ​​are encoded according to the patrol sequence of the monitoring points to obtain the track parameter encoding sequence for each transport railway, thereby constructing a track parameter encoding set for the target area. The construction of this track parameter encoding set can guide the generation of the patrol route of the second monitoring UAV, and achieve precise positioning of the state detection location when obtaining several sets of track parameters within the target sampling window collected and analyzed by the second monitoring UAV.

[0083] In a preferred embodiment, the steps of calling the railway status association parameter library of the target area and using the railway status prediction model to predict the segment location and fault type set of several railway fault sections in each transport railway specifically include:

[0084] S210: Call the railway status association parameter library of the target area, and extract the railway status association parameter prediction set for each monitoring point of each transport railway in the monitoring period.

[0085] The railway condition correlation parameter prediction set includes predicted values ​​of several railway condition correlation parameters, which include at least one or more of climate correlation parameters, geological correlation parameters, and railway track load parameters.

[0086] S220: Using the railway condition prediction model, with the railway condition correlation parameter prediction set as the prediction input sample, predict and obtain several fault monitoring points and fault types in each transport railway.

[0087] The railway status prediction model is configured to train an initially constructed convolutional neural network using training samples constructed from railway status association parameters corresponding to different types of railway historical faults in the railway historical fault database.

[0088] S230: The section between two adjacent fault monitoring points with an interval distance less than the second interval distance is regarded as the same railway fault section, and the fault types of several fault monitoring points in each railway fault section are summarized as the fault type set of the railway fault section.

[0089] In this embodiment, the prediction of the location and type of railway fault sections in each transport railway can be achieved by referring to a historical fault database and using railway status association parameters corresponding to historical faults of different fault types to train a railway status prediction model. Then, the prediction set of railway status association parameters for the monitoring period is input into the railway status prediction model to predict several fault monitoring points and fault types in each transport railway. Finally, the fault monitoring points are summarized to form railway fault sections, and the fault type set of the railway fault sections is determined.

[0090] In a preferred embodiment, the steps of querying the status information of several distributed second monitoring drones within the target area, and determining the status detection route of each second monitoring drone during the monitoring period based on the track parameter encoding set and the section location of each railway fault section, specifically include:

[0091] S310: Query the status information of several second monitoring drones distributed within the target area; wherein, the status information includes the cruise waiting position, cruise speed range, and orthophoto trajectory image processing speed of each second monitoring drone;

[0092] S320: Based on the track parameter encoding set and the section location of each railway fault section, several second monitoring drones are assigned to the corresponding railway fault sections, and the status detection line of each second monitoring drone during the monitoring period is determined.

[0093] In this embodiment, by querying the status information of several second monitoring drones distributed within the target area (usually set along the railway line), the status information is used to limit the allocation of the second monitoring drones to railway fault sections. Considering the track parameter encoding set and the section location of each railway fault section, the final status detection line of each second monitoring drone during the monitoring period is generated.

[0094] Furthermore, the steps of querying the status information of several distributed second monitoring drones within the target area specifically include:

[0095] S311: Obtain detection equipment signals transmitted by a plurality of railway line status detection nodes distributed within the target area; wherein, the detection equipment signals include node identifiers and drone identifiers of the second monitoring drone equipped for performing railway line status detection during the current monitoring period;

[0096] S312: Based on the node identifier, query the node location information of each railway line status detection node, and use the node location information as the cruise waiting position of the corresponding second monitoring drone;

[0097] S313: Based on the UAV identifier, query the cruise speed range and orthogonal orbit image processing speed of each second monitoring UAV; wherein, the cruise speed range is configured as a speed range generated according to the flight parameters provided by the UAV manufacturer and pre-written into the UAV status database, and the orthogonal orbit image processing speed is configured as the number of orthogonal orbit images processed per unit time obtained by the second monitoring UAV when performing the test task of extracting the orbit parameters of each monitoring point through orthogonal orbit image analysis.

[0098] In practical applications, the status information query of the second monitoring drone can be obtained by obtaining the node location information of the status detection node of the railway line to determine the cruise waiting position, by obtaining the flight parameters provided by the drone manufacturer to determine the cruise speed range, and by performing orthophoto image analysis tests on the same scenario to determine the orthophoto image processing speed.

[0099] Furthermore, based on the track parameter encoding set and the location of each railway fault section, several second monitoring drones are assigned to the corresponding railway fault sections, and the status detection line steps for each second monitoring drone during the monitoring period are determined, specifically including:

[0100] S321: Based on the location of each railway fault section, the cruise waiting position and cruise speed range of each second monitoring drone, calculate the time interval for each second monitoring drone to complete the status detection of the second monitoring drone allocation scheme for different railway fault sections or combinations of different railway fault sections, and eliminate the second monitoring drone allocation scheme whose minimum time interval exceeds the monitoring period.

[0101] The time interval includes the first time taken for the second monitoring drone to fly at its maximum cruising speed to the starting position of different railway fault sections or combinations of different railway fault sections, and the sum of several second times taken for the second monitoring drone to perform state detection of different railway fault sections or combinations of different railway fault sections at different cruising speeds within the cruising speed range.

[0102] It should be noted that when the second monitoring drone travels from the cruise waiting position to the assigned railway fault section, if the railway fault section is adjacent to the cruise waiting position, the second monitoring drone is directly controlled to perform status detection during the cruise at the corresponding cruise speed range. The duration at this time is only the cruise time. If the railway fault section is not adjacent to the cruise waiting position, the second monitoring drone needs to be controlled to travel to the cruise position at a higher position (to avoid the impact of mountainous terrain and vegetation on satellite positioning accuracy) at the maximum cruise speed. After reaching the cruise position, the status detection during the cruise is performed according to the corresponding cruise altitude and cruise speed. The duration at this time is the combined time consumed by drone scheduling and cruise.

[0103] S322: In the remaining second monitoring drone allocation scheme, when each second monitoring drone allocation scheme performs state detection of each railway fault section using each cruise speed value within the corresponding cruise speed range, it determines the section track parameter code of each railway fault section in the track parameter coding sequence of the railway to which it belongs based on the third interval distance between the orthophoto track image processing speed and the cruise speed value, and whether the target track parameter code obtained by performing the parameter coding bit-by-bit extraction action of the target sampling window according to the third interval distance meets the coding positioning accuracy requirements;

[0104] Among them, the coding positioning accuracy requirement is configured such that each target track parameter code obtained by bit-by-bit extraction of parameter coding according to the third interval distance in the segment track parameter coding is different;

[0105] S323: Based on the judgment result of the coding positioning accuracy requirement, the remaining second monitoring UAV allocation schemes are removed from the cruise speed range of each railway fault section where the cruise speed value that does not meet the coding positioning accuracy requirement and the cruise speed value that takes longer to perform status detection at the corresponding cruise speed for each railway fault section or combination of railway fault sections than the monitoring period. This results in the railway fault section allocation set for each second monitoring UAV when performing cruise task allocation.

[0106] The railway fault section allocation set includes several candidate railway fault sections or combinations of railway fault sections for each second monitoring UAV, as well as the cruise speed range adopted by the second monitoring UAV in each railway fault section or the combination of cruise speed ranges adopted by different railway fault sections in the combination of railway fault sections.

[0107] S324: From the railway fault section allocation set of each second monitoring UAV, select the railway fault section allocation strategy of each second monitoring UAV that covers all railway fault sections, and generate the status detection line of each second monitoring UAV during the monitoring period.

[0108] In this embodiment, a first monitoring drone equipped with satellite positioning and inertial navigation devices is controlled to collect track parameters of each transport railway in the target area at each monitoring point, construct a track parameter code set, predict the location and fault type set of railway fault sections of each transport railway in the target area, and based on the status information of the distributed second monitoring drones, considering that each target track parameter code obtained by each second monitoring drone in the corresponding railway fault section by extracting the parameter code bit by bit from the target sampling window at different cruise speeds is unique (for example, if the target sampling window is 20 monitoring points, it is necessary to ensure that each target track code composed of 20 consecutive monitoring points extracted bit by bit is unique) and the requirement that the cruise task of all railway fault sections be completed within the monitoring period (i.e., the time taken by each second monitoring drone to complete the cruise task is within the monitoring period), each railway fault section is assigned to the corresponding second monitoring drone, and a status detection line for each second monitoring drone in the monitoring period is generated. In this way, the second monitoring drones equipped only with satellite positioning devices are controlled to perform cruise detection in the corresponding railway fault sections in the target area.

[0109] Therefore, this invention determines the target track parameter encoding by analyzing the track parameters of the orthorectified track image of the target sampling window collected during cruise detection, thereby determining the position information corresponding to the current orthorectified track image. This enables the mapping between railway line status detection results and position information, improving the accuracy of railway track status detection in mountainous areas. While improving detection efficiency and positioning accuracy, the invention also utilizes a first monitoring UAV equipped with satellite positioning and inertial navigation equipment to construct the initial track parameter encoding set, and then uses a second monitoring UAV equipped only with satellite positioning equipment to perform image acquisition and positioning for railway line status detection. This significantly reduces the operating costs in engineering applications.

[0110] In a preferred embodiment, the step of generating a status detection control command for each second monitoring UAV during a monitoring period, based on the status detection line and the fault type set, specifically includes:

[0111] S410: Generate the spatial cruise route for each second monitoring UAV based on the status detection line of each second monitoring UAV and the cruise altitude requirements corresponding to the fault type of each railway fault section;

[0112] S420: Based on the cruising speed range of each railway fault section corresponding to the space cruise route and the status detection line, generate status detection control instructions for each second monitoring UAV during the monitoring period.

[0113] In this embodiment, after obtaining the status detection line of each second monitoring UAV, the lowest cruise height value is selected as the cruise height of the railway fault section based on the cruise height requirements of each fault type in the fault type set of each railway fault section (considering that different faults of the railway line require different image resolution requirements, such as small cracks in the rails and broken sleepers requiring a lower cruise height, while the railway line covered by mudslides and landslides does not require such a low cruise height). Finally, based on the cruise speed range of each railway fault section, the status detection control command of each second monitoring UAV during the monitoring period is generated.

[0114] Reference Figure 2 , Figure 2 This is a structural block diagram of an embodiment of the railway line condition monitoring system based on unmanned aerial vehicle (UAV) intelligent monitoring according to the present invention.

[0115] like Figure 2 As shown in the figure, the railway line condition detection system based on UAV intelligent monitoring proposed in this embodiment of the invention includes:

[0116] Control module 10 is used to control the first monitoring drone to collect track layout information of each transport railway within the target area and construct a track parameter code set for the target area; wherein, the first monitoring drone is configured to have at least a satellite positioning device and an inertial navigation device;

[0117] Prediction module 20 is used to call the railway status association parameter library of the target area and use the railway status prediction model to predict the section location and fault type set of several railway fault sections in each transport railway.

[0118] The query module 30 is used to query the status information of several second monitoring drones distributed in the target area, and to determine the status detection line of each second monitoring drone during the monitoring period based on the track parameter encoding set and the section location of each railway fault section.

[0119] The generation module 40 is used to generate a status detection control command for each second monitoring UAV during the monitoring period based on the status detection line and the fault type set.

[0120] The detection module 50 is used to control each second monitoring drone to perform status detection of the transport railway according to the corresponding status detection line during the monitoring period based on the status detection control command, and to obtain the status detection result; wherein, the second monitoring drone is configured to have at least a satellite positioning device.

[0121] Other embodiments or specific implementations of the railway line condition detection system based on UAV intelligent monitoring of the present invention can refer to the above-described method embodiments, and will not be repeated here.

[0122] It is understood that in the description of this specification, references to terms such as "one embodiment," "another embodiment," "other embodiments," or "first embodiment to Nth embodiment," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0123] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0124] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for detecting the condition of railway lines based on intelligent monitoring by unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: The first monitoring drone is controlled to collect track layout information of each transport railway within the target area and construct a track parameter code set for the target area; wherein, the first monitoring drone is configured to have at least a satellite positioning device and an inertial navigation device; Specifically, constructing the orbital parameter encoding set for the target region includes: Obtain the track parameters and cruise sequence of several monitoring points for each transport railway in the target area from the cruise start point to the cruise end point; Based on the patrol sequence of the monitoring points, the track parameters of several monitoring points are numerically encoded in the direction from the start point to the end point of the patrol, thereby obtaining the track parameter encoding sequence of each transport railway in the target area and constructing the track parameter encoding set of the target area; Call the railway status association parameter library of the target area and use the railway status prediction model to predict the section location and fault type set of several railway fault sections in each transport railway; The status information of several distributed second monitoring drones within the target area is queried. Based on the track parameter encoding set and the section location of each railway fault section, the status detection line of each second monitoring drone during the monitoring period is determined. Specifically, querying the status information of several distributed secondary monitoring drones within the target area includes: The detection equipment signals transmitted by several railway line status detection nodes distributed within the target area are obtained; wherein, the detection equipment signals include node identifiers and UAV identifiers of the second monitoring UAVs equipped for performing railway line status detection during the current monitoring period; Based on the node identifier, query the node location information of each railway line status detection node, and use the node location information as the cruise waiting position of the corresponding second monitoring drone; Based on the UAV identifier, query the cruising speed range and orthogonal orbit image processing speed of each second monitoring UAV; wherein, the cruising speed range is configured as a speed range generated according to the flight parameters provided by the UAV manufacturer and pre-written into the UAV status database, and the orthogonal orbit image processing speed is configured as the number of orthogonal orbit images processed per unit time obtained by the second monitoring UAV when performing the test task of extracting the orbit parameters of each monitoring point by orthogonal orbit image analysis; Based on the status detection line and the fault type set, generate status detection control instructions for each second monitoring UAV during the monitoring period; Based on the status detection control command, each second monitoring drone is controlled to perform status detection of the transport railway according to the corresponding status detection line during the monitoring period, and to obtain the status detection result; wherein, the second monitoring drone is configured to have at least a satellite positioning device.

2. The railway line condition detection method based on UAV intelligent monitoring as described in claim 1, characterized in that, The steps for controlling the first monitoring drone to collect track layout information for each transport railway within the target area specifically include: Control the first monitoring drone to perform railway patrols along each transport railway within the target area, and obtain several orthophoto images of each transport railway. Image analysis is performed on several orthophoto tracks of each transport railway to identify the track contour in each orthophoto track image and extract the track parameters of each transport railway at each monitoring point; wherein, the track parameters include the angle between the track tangent direction and the target direction at that monitoring point; The track parameters of each transport railway at each monitoring point within the target area are used as track layout information; wherein any two adjacent monitoring points are configured to have the same first interval distance on the transport railway.

3. The railway line condition detection method based on UAV intelligent monitoring as described in claim 2, characterized in that, The steps of controlling the first monitoring drone to perform railway patrols along each transport railway within the target area and obtain several orthophoto images of each transport railway specifically include: Randomly select one of the several transport railways in the target area as the target transport railway, and control the first monitoring drone to head to the cruise starting point of the target transport railway; The first monitoring drone collects an orthophoto of the starting point of the cruise. Based on the direction of the transport railway identified in the orthophoto, the cruise direction of the first monitoring drone is controlled until the cruise endpoint of the target transport railway is reached. When acquiring each orthophoto image of the target transport railway, the first monitoring UAV is simultaneously acquired based on its satellite positioning and inertial navigation equipment, and its current cruising position is used as additional information for the corresponding orthophoto image. After completing the patrol of a transport railway, it is removed from the list of transport railways in the target area, and a new transport railway is randomly selected as the target transport railway for patrol, until all transport railways in the target area have completed their patrols.

4. The railway line condition detection method based on UAV intelligent monitoring as described in claim 1, characterized in that, The steps of calling the railway status correlation parameter library of the target area and using the railway status prediction model to predict the location and fault type set of several railway fault sections in each transport railway include: Call the railway status association parameter library of the target area, and extract the railway status association parameter prediction set for each monitoring point of each transport railway in the monitoring period. The railway condition correlation parameter prediction set includes predicted values ​​of several railway condition correlation parameters, which include at least one or more of climate correlation parameters, geological correlation parameters, and railway track load parameters. Using the railway condition prediction model, with the railway condition correlation parameter prediction set as the prediction input sample, several fault monitoring points and fault types in each transport railway are predicted. The railway status prediction model is configured to train an initially constructed convolutional neural network using training samples constructed from railway status association parameters corresponding to different types of railway historical faults in the railway historical fault database. The section between two adjacent fault monitoring points whose interval distance is less than the second interval distance is regarded as the same railway fault section, and the fault types of several fault monitoring points in each railway fault section are summarized as the fault type set of the railway fault section.

5. The railway line condition detection method based on UAV intelligent monitoring as described in claim 1, characterized in that, The process involves querying the status information of several distributed second monitoring drones within a target area, and determining the status detection route for each second monitoring drone during the monitoring period based on the track parameter encoding set and the location of each railway fault section. Specifically, this includes: The status information of several distributed second monitoring drones within the target area is queried; wherein, the status information includes the cruise waiting position, cruise speed range, and orthophoto trajectory image processing speed of each second monitoring drone; Based on the track parameter encoding set and the location of each railway fault section, several second monitoring drones are assigned to the corresponding railway fault sections, and the status detection line of each second monitoring drone during the monitoring period is determined.

6. The railway line condition detection method based on UAV intelligent monitoring as described in claim 5, characterized in that, Based on the track parameter encoding set and the location of each railway fault section, several second monitoring drones are assigned to the corresponding railway fault sections. The status detection line steps for each second monitoring drone during the monitoring period are determined, specifically including: Based on the location of each railway fault section, the cruise waiting position and cruise speed range of each second monitoring drone, calculate the time interval for each second monitoring drone to complete the status detection of different railway fault sections or combinations of different railway fault sections in the second monitoring drone allocation scheme, and eliminate the second monitoring drone allocation scheme whose minimum time interval exceeds the monitoring period. The time interval includes the first time taken for the second monitoring drone to fly at its maximum cruising speed to the starting position of different railway fault sections or combinations of different railway fault sections, and the sum of several second times taken for the second monitoring drone to perform state detection of different railway fault sections or combinations of different railway fault sections at different cruising speeds within the cruising speed range. In the remaining second monitoring drone allocation scheme, when each second monitoring drone allocation scheme performs state detection of each railway fault section using each cruise speed value within the corresponding cruise speed range, it determines the section track parameter code of each railway fault section in the track parameter coding sequence of the railway to which it belongs based on the third interval distance between two adjacent selected monitoring points determined by the orthophoto track image processing speed and the cruise speed value. It then determines whether the target track parameter code obtained by performing the parameter coding bit-by-bit extraction action of the target sampling window according to the third interval distance meets the coding positioning accuracy requirements. Among them, the coding positioning accuracy requirement is configured such that each target track parameter code obtained by bit-by-bit extraction of parameter coding according to the third interval distance in the segment track parameter coding is different; Based on the judgment results of the coding positioning accuracy requirements, the remaining second monitoring UAV allocation schemes are removed from the cruise speed range of each railway fault section where the cruise speed value that does not meet the coding positioning accuracy requirements and the cruise speed value that takes longer to perform status detection at the corresponding cruise speed for each railway fault section or combination of railway fault sections than the monitoring period. This results in the railway fault section allocation set for each second monitoring UAV when performing cruise task allocation. The railway fault section allocation set includes several candidate railway fault sections or combinations of railway fault sections for each second monitoring UAV, as well as the cruise speed range adopted by the second monitoring UAV in each railway fault section or the combination of cruise speed ranges adopted by different railway fault sections in the combination of railway fault sections. From the railway fault section allocation set of each second monitoring UAV, select the railway fault section allocation strategy of each second monitoring UAV that covers all railway fault sections, and generate the status detection line of each second monitoring UAV during the monitoring period.

7. The railway line condition detection method based on UAV intelligent monitoring as described in claim 1, characterized in that, Based on the status detection line and the fault type set, the steps for generating status detection control instructions for each second monitoring UAV during the monitoring period specifically include: Based on the status detection line of each second monitoring UAV and the cruise altitude requirements corresponding to the fault type of each railway fault section, a spatial cruise route for each second monitoring UAV is generated. Based on the cruising speed range of each railway fault section corresponding to the space cruise route and the status detection line, status detection control commands are generated for each second monitoring UAV during the monitoring period.

8. A railway line condition monitoring system based on unmanned aerial vehicle (UAV) intelligent monitoring, characterized in that, include: The control module is used to control the first monitoring drone to collect track layout information of each transport railway within the target area and construct a track parameter code set for the target area; wherein, the first monitoring drone is configured to have at least a satellite positioning device and an inertial navigation device; Specifically, constructing the orbital parameter encoding set for the target region includes: Obtain the track parameters and cruise sequence of several monitoring points for each transport railway in the target area from the cruise start point to the cruise end point; Based on the patrol sequence of the monitoring points, the track parameters of several monitoring points are numerically encoded in the direction from the start point to the end point of the patrol, thereby obtaining the track parameter encoding sequence of each transport railway in the target area and constructing the track parameter encoding set of the target area; The prediction module is used to call the railway status association parameter library of the target area and use the railway status prediction model to predict the section location and fault type set of several railway fault sections in each transport railway. The query module is used to query the status information of several distributed second monitoring drones within the target area. Based on the track parameter encoding set and the section location of each railway fault section, the status detection line of each second monitoring drone during the monitoring period is determined. Specifically, querying the status information of several distributed secondary monitoring drones within the target area includes: The detection equipment signals transmitted by several railway line status detection nodes distributed within the target area are obtained; wherein, the detection equipment signals include node identifiers and UAV identifiers of the second monitoring UAVs equipped for performing railway line status detection during the current monitoring period; Based on the node identifier, query the node location information of each railway line status detection node, and use the node location information as the cruise waiting position of the corresponding second monitoring drone; Based on the UAV identifier, query the cruising speed range and orthogonal orbit image processing speed of each second monitoring UAV; wherein, the cruising speed range is configured as a speed range generated according to the flight parameters provided by the UAV manufacturer and pre-written into the UAV status database, and the orthogonal orbit image processing speed is configured as the number of orthogonal orbit images processed per unit time obtained by the second monitoring UAV when performing the test task of extracting the orbit parameters of each monitoring point by orthogonal orbit image analysis; The generation module is used to generate status detection control instructions for each second monitoring UAV during the monitoring period based on the status detection line and the fault type set; The detection module is used to control each second monitoring drone to perform status detection of the transport railway according to the corresponding status detection line during the monitoring period based on the status detection control command, and to obtain the status detection result; wherein, the second monitoring drone is configured to have at least a satellite positioning device.

Images