System for inspecting a linear infrastructure

Abstract

The invention relates to a system for inspecting a linear infrastructure (50, 60, 70). The system comprises a first unmanned aircraft (10) comprising: a laser-based cartography device, which generates map data of the surroundings of the first unmanned aircraft (10), and a transmitting device, which transmits the map data, in particular in real time, to a base station (40), and a second unmanned aircraft (20) comprising: a receiving device (21), which receives from the base station (40)a specified flight path for inspection of the linear infrastructure (50, 60, 70), the second unmanned aircraft (20) flying along the specified flight path.

Description

[0001] E.ON SE - 1 - 30A-167 534

[0002] Central German Electricity Network Company Ltd.

[0003] System for inspecting linear infrastructure

[0004] TECHNICAL AREA

[0005] The present disclosure relates to a system for inspecting linear infrastructure. In particular, a system for inspecting linear infrastructure is disclosed comprising a first unmanned aerial vehicle and a second unmanned aerial vehicle, wherein the first unmanned aerial vehicle initiates an inspection mission prior to the second unmanned aerial vehicle.

[0006] BACKGROUND

[0007] Linear infrastructure, such as power lines and pylons, requires regular inspection. Overhead power lines, i.e., power lines strung between poles and towers, can be inspected using camera-equipped drones. A drone flies along the power line, controlled by an operator (service personnel) using a control device. To control the drone safely and reliably, for example, to maneuver it around trees, the operator must maintain visual contact with the drone. If the drone flies at a speed higher than walking pace (e.g., 40 km / h), the operator will quickly lose visual contact, and the drone will have to wait for the operator, thus increasing the time required for the power line inspection.

[0008] SUMMARY OF THE REVELATION

[0009] The present disclosure is based on the objective of providing a system for inspecting linear infrastructure, which can reduce the time required for inspecting the linear infrastructure.

[0010] To solve this problem, a system for inspecting linear infrastructure is proposed. The system comprises: a first unmanned aerial vehicle (UAV) comprising: a laser-based mapping device that generates map data of the UAV's surroundings and a transmitter that sends the map data, particularly in real time, to a base station; and a second UAV comprising: a receiver that receives a predetermined flight path from the base station for inspecting the linear infrastructure, with the second UAV flying along the predetermined flight path. E.ON SE - 2 - 30A-167 534

[0011] Central German Electricity Network Company Ltd.

[0012] Linear infrastructure refers to a physical or geographical arrangement of facilities, means of transport, or supply systems in a straight line and / or along a route. Examples of linear infrastructure include a road, a power line, a communications network, or a water supply system. Specifically, it can be a railway line, a gas pipeline, an array of industrial chimneys, or a power line with pylons requiring inspection. The power line requiring inspection can be any type of power line, particularly one or more overhead lines strung between pylons and towers.

[0013] The first and second unmanned aerial vehicles (UAVs) can be aircraft that operate without a human pilot on board. Such aircraft are also known as unmanned aerial vehicles or drones. These aircraft can be autonomous, controlled by pre-programmed flight plans and algorithms, or remotely controlled by a human operator. They can be equipped with various sensors and technologies to perform specific tasks such as surveillance, mapping, inspections, and deliveries. They can also be configured to collect and process real-time data. Unmanned aerial vehicles can be equipped with cameras, light detection and ranging (LDR) systems, lidar, global positioning (GPS) systems, and / or other navigation systems to fulfill their missions.

[0014] The laser-based mapping device is a system that uses laser beams to scan the environment of the first unmanned aerial vehicle (UAV) and generate detailed map data. This device can utilize lidar to measure distances by transmitting and receiving laser pulses and create a three-dimensional representation of the surroundings. The captured data can be transmitted in real time to a base station, which analyzes the information and transforms it into detailed maps. These maps can depict topographic features, objects, and obstacles in the UAV's vicinity. By integrating the laser-based mapping device into the UAV, this technology enables autonomous navigation and precise positioning, which can improve the efficiency and safety of missions.

[0015] The transmitter is a component of the first unmanned aerial vehicle, responsible for transmitting the collected data to the base station at E.ON SE - 3 - 30A-167 534

[0016] Mitteldeutsche Netzgesellschaft Strom mbH. This device can utilize various communication protocols and technologies, such as radio frequencies, satellite communication, or mobile networks, to ensure stable and fast data transmission. The transmitting device can be connected to the sensors and data acquisition systems of the first unmanned aerial vehicle, enabling it to continuously transmit the collected information, such as map data generated by lidar technology, to the base station.

[0017] The base station is a central facility that communicates with unmanned aerial vehicles (UAVs) and receives, analyzes, and processes the data they transmit. This station is equipped with computers and software solutions that enable real-time monitoring and evaluation of the information transmitted by the UAVs. In particular, when using laser-based mapping devices, the raw data captured by the sensors is transformed at the base station into detailed maps and three-dimensional models of the environment. The base station serves not only as a data processing center but also as a control and management unit, allowing operators to plan and monitor the UAVs' missions and make adjustments as needed.Through continuous communication with the aircraft, the base station ensures seamless communication and contributes to the efficiency and accuracy of missions. The base station can be mobile, for example in a support vehicle, or permanently installed.

[0018] The second unmanned aerial vehicle (UAV) is equipped with a receiver designed to receive communication signals from the base station. This receiver picks up the predetermined flight path transmitted by the base station. The predetermined flight path is a detailed plan that precisely defines how the UAV should navigate to inspect linear infrastructure, such as a power line, pipeline, or railway line. For example, the flight path may specify trajectories. Once the receiver has received the flight path, the second UAV processes this information and uses it to control its navigation. This means that the UAV flies autonomously along the predetermined flight path, inspecting the linear infrastructure.During the flight, the second unmanned aerial vehicle can use its sensors and cameras to collect data on the condition of the infrastructure, which can then be analyzed. E.ON SE - 4 - 30A-167 534.

[0019] Central German Electricity Network Company Ltd.

[0020] Thus, the first unmanned aerial vehicle is used to explore the area of ​​the linear infrastructure to be inspected for the first time, and the second unmanned aerial vehicle uses the exploration data from the first unmanned aerial vehicle to inspect the linear infrastructure.

[0021] The system's base station can include the following devices: a receiving device that receives the map data from the first unmanned aerial vehicle, a processing device that calculates the predetermined flight path of the second unmanned aerial vehicle based on the map data, and a transmitting device that sends the predetermined flight path to the second unmanned aerial vehicle.

[0022] The base station's receiver is a component responsible for receiving communication signals and data from unmanned aerial vehicles (UAVs). This receiver is designed to reliably capture information transmitted by the UAVs, such as map data, in real time. By utilizing various communication protocols and technologies, such as radio frequencies, satellite links, or cellular networks, the receiver ensures stable and uninterrupted data transmission. The received data is then forwarded to the base station's processing unit, where it is analyzed and used for mission planning and control.

[0023] The base station's transmitter is a component responsible for sending communication signals and data to the unmanned aerial vehicle (UAV). This device enables the base station to securely transmit calculated flight paths, control commands, and other relevant information to the UAV in real time. By utilizing various communication protocols and technologies, such as radio frequencies, satellite links, or cellular networks, the transmitter ensures stable and reliable data transmission. The receiver and transmitter can be integrated into a single unit.

[0024] The base station's processing unit can be a computer system designed to analyze received map data and, optionally, other sensor information, and to make decisions for the unmanned aerial vehicles (UAVs) based on this data. The processing unit can take over the control and coordination of the UAVs by processing data, calculating flight paths, and performing optimizations. Initially, the E.ON SE - 5 - 30A-167 534

[0025] Central German Electricity Network Company Ltd.

[0026] The processing device receives the map data transmitted by the first unmanned aerial vehicle. This data includes three-dimensional information about the environment, including topographic features, objects, and potential obstacles.

[0027] The received map data can first be filtered and cleaned to remove noise and improve data accuracy. Using image processing algorithms and / or machine learning, the processing device can identify relevant features and potential obstacles in the environment. Based on the analyzed data, the processing device can define specific points and / or areas for the second unmanned aerial vehicle (UAV) to inspect. The processing device can use route planning algorithms to calculate the most efficient and safest flight path, taking into account factors such as the position of the features, the height of the obstacles, the range of the sensors, and / or the battery capacity of the second UAV.

[0028] The algorithms ensure that the calculated flight path maintains sufficient safety distances from identified obstacles to avoid collisions. The processing unit can adjust the flight path in real time if new information becomes available or the environment changes during the mission. This ensures that the second unmanned aerial vehicle (UAV) always follows the optimal path. All legal or mission-specific restrictions are incorporated into the flight path planning to ensure compliance with regulations and mission objectives. After calculation, the processing unit generates a detailed flight path that includes all necessary navigation points, altitude changes, and inspection targets. The calculated flight path is then transmitted to the base station's transmitter, which in turn sends it to the second UAV.The base station's processing unit is a system that analyzes the received map data and, based on this data, calculates the optimal flight path for the second unmanned aerial vehicle. By employing data analysis and flight path planning algorithms, the processing unit ensures efficient, safe, and precise inspection of the linear infrastructure.

[0029] According to a preferred embodiment, the second unmanned aerial vehicle can begin inspecting the linear infrastructure immediately after receiving the predetermined flight path, thus saving time. Starting the inspection immediately increases the efficiency of the entire mission, as no unnecessary E.ON SE - 6 - 30A-167 534

[0030] Central German Electricity Network Company Ltd.

[0031] Delays may occur. Another advantage is the timeliness of the data. Since the second unmanned aerial vehicle (UAV) begins the inspection immediately after receiving the flight path, the collected data is based on the latest information gathered by the first UAV. This ensures that the inspection is carried out under current conditions and taking into account the latest environmental and infrastructure data. The immediate start of the inspection also improves the accuracy and relevance of the results. Because the data is processed continuously and in real time, the second UAV can react to current conditions and thus perform more precise and detailed inspections. This is particularly beneficial for detecting and assessing damage or anomalies to the infrastructure, which may be time-critical.Furthermore, this approach enhances coordination and synchronization between the unmanned aerial vehicles (UAVs) and the base station. The immediate commencement of the inspection by the second UAV upon receiving the flight path optimizes the use of communication and control systems, thereby increasing the overall performance and effectiveness of the inspection mission. Finally, the immediate start of the inspection also contributes to safety. Continuous monitoring and the use of real-time data enable the second UAV to better identify and avoid potential hazards and obstacles. This minimizes the risk of incidents during the mission and ensures the safe execution of the inspection.

[0032] According to one embodiment of the present disclosure, the second unmanned aerial vehicle (UAV), unlike the first UAV, does not include a laser-based mapping device, resulting in a weight reduction. Lidar systems are often heavy and bulky, so their absence significantly reduces the aircraft's weight. A lighter aircraft requires less energy to operate, which extends flight time and improves mission efficiency. The complexity of the second UAV is also reduced. Without the laser-based mapping device, fewer sensors and less complex data processing systems are required. This leads to simpler maintenance and lower failure rates, as there are fewer potential sources of error. Operation and training of drone pilots or operators are also simplified, as less specialized knowledge is required.Without a laser-based mapping device, the second unmanned aerial vehicle can also be deployed more agilely and flexibly. It can react faster and adapt to different inspection requirements, as it does not rely on processing and analyzing extensive 3D data. This is particularly advantageous in dynamic or rapidly changing environments, as E.ON SE - 7 - 30A-167 534.

[0033] Mitteldeutsche Netzgesellschaft Strom mbH requires quick decisions and adjustments. Since the second unmanned aerial vehicle (UAV) receives its mapping data from the base station, it can concentrate fully on inspecting the linear infrastructure. Data processing and analysis are performed centrally at the base station, which can be equipped with powerful computers and specialized software. This ensures that the second UAV always receives the most up-to-date and precise information without having to perform the complex calculations itself.

[0034] The first unmanned aerial vehicle (UAV) can be equipped without a camera or gimbal. Furthermore, the first UAV can continuously transmit new map data to the base station, particularly in real time. The second UAV can also continuously receive and follow new, predefined flight paths for inspecting the linear infrastructure, again in real time. One advantage of this system is the increased efficiency and accuracy of inspection missions. Because the first UAV continuously transmits updated map data to the base station, any changes in the environment are immediately detected and analyzed. This allows the base station to always process the latest and most accurate information and adjust the flight paths of the second UAV accordingly. As a result, potential problems or obstacles can be identified and avoided more quickly.Another advantage is the improvement in safety standards. Continuous, real-time data transmission allows the second unmanned aerial vehicle (UAV) to react immediately to changes in its environment. For example, new obstacles or hazardous conditions detected by the first UAV can be instantly incorporated into the flight plan, enabling the second UAV to navigate safely. This reduces the risk of collisions and increases the overall safety of the mission. Real-time communication between the two UAVs and the base station also allows for a high degree of flexibility and adaptability. Should the environment to be inspected change or new inspection targets be defined, the base station can immediately process this information and forward it to the second UAV.This allows the system to react dynamically to unforeseen events or requirements and adapt the mission accordingly. Another advantage is the optimization of resource utilization. Continuous updates of the flight paths ensure that the second unmanned aerial vehicle always chooses the most efficient and safest route. This helps to maximize flight time and minimize energy consumption. At the same time, the central processing enables E.ON SE - 8 - 30A-167 534.

[0035] Central German Electricity Network Company Ltd.

[0036] Map data in the base station enables efficient use of computing resources, as the second aircraft itself does not need to perform complex calculations.

[0037] The second unmanned aerial vehicle (UAV) can also be configured to receive coordinate data from a supply vehicle and, depending on the battery charge level of the second UAV and its distance from the supply vehicle, initiate an approach to the supply vehicle. The supply vehicle's coordinate data can be represented as precise geographic positions in a coordinate system. This data can consist of latitude and longitude values, which specify the vehicle's exact position on the Earth's surface. Additionally, the coordinate data can include altitude information to define the vehicle's vertical position. This data can be provided in the form of GPS coordinates.

[0038] According to this embodiment, the second unmanned aerial vehicle (UAV) can continuously monitor its battery charge level. It can receive the current coordinates of a support vehicle specifically designed to recharge or swap UAV batteries during a mission. If the battery charge level falls below a certain threshold and the distance to the support vehicle is within an acceptable range, the UAV can automatically initiate an approach to the support vehicle, for example, to recharge its battery. One advantage of this system is the extended operating time of the second UAV. By being able to recharge or swap its battery during the mission, the UAV can remain airborne longer and perform its inspection tasks without prolonged interruptions.This increases mission efficiency by reducing time spent landing, swapping, or recharging batteries on the ground. The automated approach to the supply vehicle also reduces the need for human intervention. The unmanned aerial vehicle can autonomously determine when a battery swap or recharge is required, based on the current charge level and distance to the supply vehicle. This enhances the system's autonomy and frees up operators to focus on other tasks. Continuous monitoring of the battery charge level and automatic initiation of the approach to the supply vehicle minimizes the risk of unforeseen battery failure in flight. This contributes to mission safety, as the unmanned aerial vehicle always has sufficient power to safely perform its tasks and, if necessary, return to a safe location.This embodiment is offered by E.ON SE - 9 - 30A-167 534.

[0039] Mitteldeutsche Netzgesellschaft Strom mbH also benefits from high flexibility and adaptability, enabling the unmanned aerial vehicle to dynamically adjust to mission requirements and environmental conditions. Should the mission be unexpectedly extended or the environment change, the unmanned aerial vehicle can continue operating by regularly recharging its battery. By integrating a support vehicle that allows for battery charging during the mission, the system can operate almost continuously. This is particularly advantageous for lengthy inspection tasks or missions in remote areas where frequent returns to base are impractical.

[0040] The second unmanned aerial vehicle (UAV) can determine its distance to the supply vehicle using GPS. Both the UAV and the supply vehicle can be equipped with GPS receivers, allowing them to determine their precise positions in real time. The UAV can receive the supply vehicle's coordinates via a communication link and compare its own position with that of the supply vehicle. By calculating the distance between the two GPS coordinates, the UAV can accurately determine its distance to the supply vehicle. Another method for determining the distance is the use of radio communication. The supply vehicle can transmit a signal that is received by the second UAV.By measuring the signal strength and the time it takes for the signal to reach the second unmanned aerial vehicle (UAV), the distance can be estimated. This method can be refined by using signal strength values ​​and distance estimation algorithms. Lidar and radar systems can also be used for distance determination. These systems emit light or radio waves that are reflected by objects. By measuring the time it takes for the waves to travel to the support vehicle and back, the distance can be calculated accurately. This method offers high accuracy and is particularly useful in environments where GPS signals are unreliable.

[0041] The support vehicle can be configured in various ways to meet the requirements of battery charging or replacement for unmanned aerial vehicles (UAVs). Specifically, the support vehicle, which travels along a road, can be equipped to charge or replace the battery of the second UAV. The support vehicle can also be configured to replace the second UAV with another UAV. This additional UAV can perform the mission of E.ON SE - 10 - 30A-167 534

[0042] Mitteldeutsche Netzgesellschaft Strom mbH will continue with a second unmanned aerial vehicle (UAV). The support vehicle can be designed as a specially equipped ground vehicle or as a mobile platform capable of providing power and potentially other maintenance services to the UAVs. Initially, the support vehicle can be equipped with an energy storage system consisting of batteries or fuel cells. This energy storage system allows the vehicle to store a large amount of energy, which can be used to quickly recharge the UAVs' batteries. To maximize efficiency and flexibility, the vehicle can have multiple charging ports, enabling it to recharge several UAVs simultaneously or in rapid succession.The supply vehicle can also be equipped with a positioning system that uses GPS and other location technologies to determine its precise position and transmit it to the unmanned aerial vehicles (UAVs). This allows the UAVs to quickly and accurately locate the supply vehicle, facilitating approach and docking. For the actual charging process, the supply vehicle can be equipped with automated docking stations or charging arms that enable the UAVs to dock safely and efficiently. These docking systems can be designed to stabilize the UAVs during charging while ensuring fast and secure power transfer. Additionally, the supply vehicle can have communication systems that continuously exchange data with the UAVs and the base station.These systems can transmit information about the battery charge level, the charging status, and any maintenance requirements. This ensures that all involved systems are always aware of the current status and can react accordingly. To be operational in diverse environments, the supply vehicle can also be designed to be robust and off-road capable. This enables deployment in varied terrains, from urban areas to remote or difficult-to-access regions. Furthermore, the vehicle can be designed for quick and easy relocation to different deployment locations to flexibly respond to changing mission requirements.

[0043] The route and flight path planning algorithms can also plan an approach of the second unmanned aerial vehicle (UAV) to the supply vehicle based on coordinate data of the supply vehicle, the battery charge level of the second UAV, and / or the distance between the second UAV and the supply vehicle. E.ON SE - 11 - 30A-167 534

[0044] Central German Electricity Network Company Ltd.

[0045] The base station can be located in the supply vehicle or in a trailer connected to the supply vehicle. The supply vehicle or the trailer connected to the supply vehicle can provide a landing pad for unmanned aerial vehicles.

[0046] The base station can also be configured to specify flight paths for a large number of unmanned aerial vehicles and to change the specified flight paths, especially in real time.

[0047] Furthermore, a third unmanned aerial vehicle (UAV) can be provided, which, after the second UAV begins its approach to the supply vehicle, continues the flight of the second UAV along the predetermined flight path. An advantage of this configuration is mission continuity. If the second UAV begins its approach to the supply vehicle to recharge its battery due to a low battery level, the third UAV takes over the inspection of the linear infrastructure. This ensures that no inspection gap occurs and the mission continues without interruption. This is particularly important for tasks that are time-critical or require continuous monitoring. Another advantage is the increased efficiency of the overall system.By transferring inspection tasks to the third unmanned aerial vehicle (UAV), the second UAV can recharge its battery without extending the overall duration of the inspection mission. This allows for optimal resource utilization, as both UAVs are active simultaneously—one charging and the other conducting inspections. The system's flexibility and adaptability are also enhanced. The third UAV can be seamlessly integrated into the mission by taking over the predefined flight paths and inspection tasks. This enables the system to dynamically respond to varying mission requirements and unexpected events without compromising inspection effectiveness. The use of a third UAV also improves the overall performance of the inspection.Since the third aircraft takes over the task of the second, the inspection process remains consistent and the quality of the collected data is maintained. Continuous monitoring and data acquisition ensure that no critical information is overlooked and that the integrity of the infrastructure can be fully assessed. This approach also contributes to mission safety. The third unmanned aerial vehicle can seamlessly continue the inspection while the second aircraft safely approaches the supply vehicle. This minimizes E.ON SE - 12 - 30A-167 534.

[0048] Mitteldeutsche Netzgesellschaft Strom mbH assesses the risk of incidents or outages that could be caused by low battery levels and ensures that the inspection is carried out under safe and controlled conditions.

[0049] According to a further development of the present disclosure, the second unmanned aerial vehicle comprises a camera, a gimbal carrying the camera, a lidar system configured to scan the linear infrastructure and generate SD point cloud data, and a processing device configured to receive the 3D point cloud data from the lidar system, analyze the 3D point cloud data to detect features of interest on the linear infrastructure, calculate, based on the detected features of interest, camera angles and / or distances to the features of interest, and generate, based on the calculated camera angles and / or distances to the features of interest, commands for the gimbal, wherein the gimbal is configured to change the orientation of the camera, in particular in real time, based on the commands of the processing device.

[0050] Features of interest can include, for example, structural components, potential defects, and specific points requiring closer examination. For instance, an insulator on a power pole might be inspected as a feature of interest.

[0051] The camera on the second unmanned aerial vehicle (UAV) can be an optical device that captures images and videos of the linear infrastructure. The camera is designed to take high-resolution images to enable detailed inspections of the infrastructure. It can capture various types of images, including visual images and, depending on the equipment, infrared or thermographic images, to analyze different aspects of the infrastructure.

[0052] The gimbal, which supports the camera, is a mechanical stabilization device that holds the camera in a stable position and can control its orientation on various axes. The gimbal can be configured to compensate for aircraft movements and keep the camera in the optimal position regardless of the aircraft's flight maneuvers. It can orient the camera in real time according to instructions from the processing device. This is done based on the calculated optimal angles and distances to the detected features of interest. This ensures that the camera always maintains the best viewing angle and the correct distance to E.ON SE - 13 - 30A-167 534

[0053] Mitteldeutsche Netzgesellschaft Strom mbH has the areas to be inspected in order to guarantee high-quality and precise recordings.

[0054] Together, the camera and gimbal, in combination with the lidar system and processing unit, enable comprehensive and efficient inspection of linear infrastructure. The lidar system scans the environment and generates 3D point cloud data, which is analyzed by the processing unit to identify relevant features. Based on this analysis, the processing unit calculates the optimal camera angles and distances and sends corresponding commands to the gimbal, which aligns the camera accordingly. This technology allows for precise and dynamic camera adjustment, resulting in improved inspection quality and efficiency.

[0055] Analyzing 3D point cloud data to identify features of interest in linear infrastructure can be accomplished through a series of sequential steps, potentially supported by data processing algorithms and / or machine learning. First, the 3D point cloud data acquired by the lidar system is preprocessed. This preprocessing includes noise and outlier removal, data calibration, and point cloud normalization to ensure consistent scaling and orientation. These steps improve data quality and establish a consistent foundation for subsequent analysis. Following preprocessing, the point cloud is segmented to identify distinct areas and objects within it. This segmentation can be based on geometric properties such as surfaces, edges, and volumes.Area detection algorithms or clustering methods such as Density-Based Spatial Clustering of Applications with Noise (DBSCAN) can be used to identify contiguous structures within the point cloud. The next step involves feature detection within the segmented point cloud. This includes identifying specific structures and details relevant for inspecting linear infrastructure, such as cables, poles, insulators, or anomalies like cracks and deformations. Machine learning algorithms, particularly neural networks trained on large amounts of training data, can be used to automatically detect and classify these features. The detected features are then geometrically analyzed to determine their precise dimensions, positions, and orientations within the point cloud.This enables precise localization of features and the calculation of relevant distances and angles. Computer vision methods and mathematical models can be used to perform these geometric analyses. (Subsequently E.ON SE - 14 - 30A-167 534.)

[0056] Mitteldeutsche Netzgesellschaft Strom mbH classifies and evaluates the identified characteristics to determine their significance and the condition of the infrastructure. This can include categorizing characteristics as "normal," "abnormal," or "critical," which helps in prioritizing inspection and maintenance measures.

[0057] Based on the detected and analyzed features, the processing unit calculates the optimal camera angles and distances required to capture detailed and relevant images of the features of interest. These calculations take into account the spatial arrangement of the features as well as the movement and position of the unmanned aerial vehicle. First, the processing unit identifies the precise positions of the features of interest within the 3D point cloud data generated by the lidar system. These positions are then defined as coordinates in three-dimensional space.

[0058] The processing unit then analyzes the spatial relationship between the unmanned aerial vehicle (UAV) and the detected features. This analysis takes into account the UAV's current position and orientation, as well as the positions of the features. Spatial analysis algorithms can calculate the optimal viewing angles from which the camera should capture the features to obtain the best visual information. These calculations are based on geometric principles and consider factors such as distance to the feature, height, and angle to minimize distortion and ensure a clear view.

[0059] In addition to the viewing angles, the optimal distances between the camera and the features of interest are calculated. This ensures that the camera is close enough—that is, the unmanned aerial vehicle flies close enough to the feature of interest to capture detailed images—but at the same time maintains sufficient distance to capture the entire feature of interest within the field of view. The calculations also take into account the movements of the unmanned aerial vehicle to enable stable and shake-free images.

[0060] The calculated camera angles and distances are then updated in real time while the unmanned aerial vehicle is flying. Finally, the calculated camera angles and distances are converted into commands for the gimbal. These commands control the camera's orientation and position in real time to ensure that the camera is always optimally aligned with the features of interest. The processing unit continuously sends adjustments to the gimbal to ensure that the camera is always optimally aligned. E.ON SE - 15 - 30A-167 534

[0061] The system is designed for Mitteldeutsche Netzgesellschaft Strom mbH. These dynamic adjustments allow the system to react flexibly to changes in the environment and the position of features, thus ensuring consistent image quality. This precise and continuous calculation process ensures that the camera always captures the best images of the detected features of interest, significantly improving the efficiency and accuracy of the inspection mission.

[0062] The processing device can also be configured to continuously update the camera angle and / or distances to the features of interest based on 3D point cloud data received and analyzed in real time, while the second unmanned aerial vehicle flies around the linear infrastructure.

[0063] According to an alternative embodiment, the second unmanned aerial vehicle can also receive the 3D point cloud data from the first unmanned aerial vehicle. For this purpose, the first unmanned aerial vehicle can include a lidar system that generates the SD point cloud data.

[0064] Alternatively, the following process steps can be performed in the first unmanned aerial vehicle or the base station by a processing device: receiving the 3D point cloud data, analyzing the 3D point cloud data to identify features of interest on the linear infrastructure, calculating, based on the identified features of interest, camera angles and / or distances to the features of interest, and generating, based on the calculated camera angles and / or distances to the features of interest, commands for the gimbal.

[0065] The generated commands for the gimbal can then be sent from the first unmanned aerial vehicle (UAV) or from the base station via the first UAV to the second UAV. The gimbal then adjusts the camera's orientation based on the received commands.

[0066] The second unmanned aircraft may further include a navigation system designed to control, based on the camera angles and / or distances to the features of interest calculated by the processing device, a flight path of the second unmanned aircraft.

[0067] The navigation system can be a combination of various technologies and sensors that work together to ensure the precise control and navigation of the unmanned aerial vehicle. Initially, the E.ON SE - 16 - 30A-167 534

[0068] Central German Electricity Network Company Ltd.

[0069] The navigation system must be equipped with a Global Positioning System (GPS) that provides accurate, real-time position data. This GPS data enables the unmanned aerial vehicle (UAV) to determine its precise geographic position and move according to its calculated flight path. In addition to GPS, the navigation system may be equipped with inertial sensors such as gyroscopes and accelerometers. These sensors detect the aircraft's movements and orientation, allowing it to continuously monitor and adjust its attitude and stability. These inertial sensors are particularly useful for the precise control and stabilization of the aircraft, especially in situations where GPS signals may be unreliable.

[0070] Another important element of the navigation system can be a barometer, which measures the aircraft's altitude above the ground. This is particularly important for precise altitude control, ensuring the aircraft always flies at the optimal altitude for inspecting linear infrastructure. Additionally, a lidar or radar system can be integrated to detect and avoid obstacles. These systems scan the environment in real time and provide data on potential obstacles, allowing the navigation system to adjust the aircraft's flight path accordingly to prevent collisions. The navigation system can also be equipped with a communication module that enables continuous data transmission between the aircraft and the base station.This module ensures that the calculated camera angles and distances, as well as other relevant data, are transmitted to the aircraft in real time and that the aircraft's status and position data are sent back to the base station. A component of the navigation system may also be a processing unit capable of executing complex algorithms for flight path planning and control. This processing unit processes data from GPS, inertial measurement, barometric, and / or lidar / radar sensors and uses this information to calculate the optimal flight path and send control commands to the aircraft's flight control systems.

[0071] The navigation system can work in conjunction with the processing unit to ensure the aircraft maintains the optimal position and orientation to capture the best images of relevant features. To do this, the navigation system continuously receives calculated optimal camera angles and distances from the processing unit and uses this information to dynamically adjust the aircraft's flight path. This means the unmanned aerial vehicle can change its position and direction in real time to keep the camera in the best position for capturing clear and detailed images of the features. E.ON SE - 17 - 30A-167 534

[0072] Central German Electricity Network Company Ltd.

[0073] The navigation system takes into account the spatial relationship between the aircraft and the features of interest and controls the aircraft's movements accordingly. The advantage of this system lies in the improved efficiency and accuracy of the inspection mission. By continuously adjusting the aircraft's flight path to the calculated camera angles and distances, the navigation system ensures that the camera is always optimally aligned. This results in high-quality images that are beneficial for analyzing and evaluating the infrastructure. Furthermore, the system allows for a flexible and rapid response to changes in the environment, enabling reliable inspections even in dynamic and complex situations. Another advantage is the reduction of manual intervention.The automated navigation system takes over the complex tasks of flight control and camera positioning, allowing human operators to focus on analyzing the collected data. This increases the overall efficiency of the inspection process and reduces the likelihood of errors that could arise from manual control.

[0074] The processing device can further be configured to determine obstacles and / or the structure of the linear infrastructure, and the navigation system can further be configured to control the flight path of the second unmanned aerial vehicle based on the obstacles and / or the structure of the linear infrastructure determined by the processing device. This means that the processing device not only analyzes the 3D point cloud data to detect features of interest, but is also designed to identify potential obstacles in the environment as well as the precise structure of the infrastructure to be inspected. To determine obstacles, the processing device analyzes the 3D data to detect objects such as trees, buildings, other vehicles, or any barriers that might be located near the flight path of the unmanned aerial vehicle.These obstacles are marked in the data, and their positions and dimensions are recorded. Simultaneously, the processing unit analyzes the structure of the linear infrastructure, such as the shape and arrangement of power lines or pipelines, to ensure that the aircraft can inspect them accurately and comprehensively.

[0075] The navigation system can also be configured to control the flight path of the second unmanned aerial vehicle based on obstacles determined by the processing device and / or the structure of the linear infrastructure. This means that the navigation system uses information about the detected obstacles and the infrastructure structure to dynamically adjust the aircraft's flight path. E.ON SE - 18 - 30A-167 534

[0076] Central German Electricity Network Company Ltd.

[0077] For example, if the processing device identifies an obstacle near the current flight path, the navigation system calculates a new flight path that safely avoids the obstacle while maintaining the optimal position for infrastructure inspection. The advantage of this system lies in the improved safety and efficiency of the inspection mission. By continuously monitoring and adjusting the flight path based on detected obstacles, the risk of collisions is minimized, thus increasing the safety of the aircraft and its surroundings. Furthermore, the precise determination of the infrastructure structure enables a detailed and accurate inspection, as the aircraft is able to optimally adjust its position and orientation to capture clear and comprehensive data.

[0078] The processing unit and navigation system can also be configured to synchronize the camera's orientation with the flight path of the second unmanned aerial vehicle (UAV) so that the camera captures the features of interest. This means that both units work together to ensure the camera is always optimally oriented as the UAV flies along its predetermined flight path. The processing unit continuously analyzes the SD point cloud data and identifies features of interest, such as structural details or potential anomalies in the linear infrastructure. Based on this analysis, the processing unit calculates the optimal camera angles and distances to capture these features in the best possible way. Simultaneously, the navigation system controls the UAV's flight path so that it remains in the ideal position relative to the detected features.This involves adjusting the aircraft's altitude, heading, and speed to ensure the camera is positioned at the optimal angle to capture the features. Synchronization of the camera's orientation with the flight path can occur in real time. The processing unit continuously sends data on optimal camera angles and distances to the navigation system. The navigation system uses this information to adjust the aircraft's movements accordingly. Simultaneously, the navigation system controls the gimbal supporting the camera to ensure it remains focused on the features of interest. The advantage of this synchronization lies in the improved quality and precision of the captured images and videos. Because the camera is always optimally aligned and the aircraft's flight path is dynamically adjusted, clear and detailed images of the features can be captured.This is particularly important for the precise inspection and analysis of the infrastructure, as it allows for the early detection of potential problems and accurate assessments. Additionally, this approach increases the efficiency of inspection missions. (E.ON SE - 19 - 30A-167 534.)

[0079] Mitteldeutsche Netzgesellschaft Strom mbH and synchronization ensure that the camera always captures the best images without requiring the aircraft to perform unnecessary maneuvers. This saves time and energy, which is particularly advantageous for extensive inspection tasks.

[0080] According to a further development of the present disclosure, the camera is configured to automatically adjust its focus and / or image settings based on the detected features of interest in order to improve image and / or video quality. The camera may have an autofocus mechanism and variable settings such as exposure, white balance, ISO, and aperture, which can be adjusted automatically. The aircraft's processing device continuously analyzes the SD point cloud data generated by the lidar system to identify features of interest. This processing can be performed using algorithms that employ machine learning and image processing methods. Once a feature is detected, its position and size are determined. Based on this information, the processing device calculates the optimal camera angles and distances to ensure the best possible images.The processing unit then sends control commands to the camera to adjust focus and image settings. For example, the focus point is automatically set to the detected feature, and the exposure can be adjusted to the current lighting conditions to optimize image quality. Simultaneously, the gimbal receives commands to orient the camera based on calculated angles and distances, ensuring that the camera is always optimally aligned. A continuous feedback loop between the camera, gimbal, and processing unit ensures that all adjustments are made in real time. This allows the system to react quickly to changes in the environment or the position of the features.This integration of hardware and software improves image and video quality, as the camera automatically adjusts focus and settings to always take the best shots.

[0081] The processing device can also be configured to utilize machine learning algorithms to optimize the detection of features of interest and, based on this, the camera angles. This can be achieved through a structured integration of machine learning and data-driven approaches. Initially, a comprehensive collection of training data is required, encompassing images, videos, and 3D point clouds representing various features of interest. This data is annotated to identify the relevant features, enabling the algorithms to learn to recognize these features during the training process. Once the training data is collected, several machine learning models are used. E.ON SE - 20 - 30A-167 534

[0082] Mitteldeutsche Netzgesellschaft Strom mbH develops and trains image processing models such as Convolutional Neural Networks (CNNs). These models are trained on annotated image data to identify specific features in images and videos. For 3D point cloud data, models like PointNet are used to detect features within the point clouds by analyzing their spatial structure.

[0083] After training, these models are integrated into the processing unit. The processing unit continuously receives 3D point cloud data from the lidar system, as well as image and video data from the camera, and processes this data in real time. First, the data is preprocessed to reduce noise and improve quality. Then, the preprocessed data is fed into the trained machine learning models, which identify features of interest. The models provide information about the position, size, and type of features. Based on the identified features, the processing unit calculates the optimal camera angles and distances. These calculations take into account the spatial arrangement of the features, the aircraft's current position and orientation, and the gimbal's movements.The calculated camera angles and distances are transmitted to the gimbal and camera in real time, allowing the gimbal to adjust the camera's orientation accordingly, ensuring it remains focused on the features of interest. This continuous real-time processing and adjustment significantly improves image and video quality, resulting in more precise and detailed inspections. This enables efficient and accurate data collection, enhancing the overall quality and efficiency of inspection missions.

[0084] The second unmanned aerial vehicle may also include communication equipment for sending images and / or videos captured by the camera to the base station.

[0085] The aspects and variants described above can be combined without this being explicitly stated. Each of the described design variants is therefore optional to any other design variant or combination thereof. This disclosure is thus not limited to the individual designs and variants in the described order or to any specific combination of aspects and design variants.

[0086] BRIEF DESCRIPTION OF THE DRAWINGS E.ON SE - 21 - 30A-167 534

[0087] Central German Electricity Network Company Ltd.

[0088] Further advantages, details and features of the devices and systems described here will become apparent from the following description of exemplary embodiments and the figures.

[0089] Fig. 1 shows a schematic representation of an embodiment of a system for inspecting linear infrastructure;

[0090] Fig. 2 shows a schematic representation of an embodiment of a first unmanned aerial vehicle;

[0091] Fig. 3 shows a schematic representation of an embodiment of a second unmanned aerial vehicle;

[0092] Fig. 4 shows a schematic representation of an embodiment of a base station for unmanned aerial vehicles; and

[0093] Fig. 5 shows a schematic representation of an embodiment of the second unmanned aerial vehicle inspecting an insulator on a power pole of a high-voltage power line.

[0094] DETAILED DESCRIPTION

[0095] Fig. 1 shows a schematic representation of an embodiment of a system for inspecting linear infrastructure.

[0096] The linear infrastructure consists of a high-voltage cable 50, which is strung between two power pylons 60 and 70. Various vegetation, such as a tree 95, is located in the vicinity of the linear infrastructure. This vegetation presents obstacles for unmanned aerial vehicles.

[0097] The system for inspecting linear infrastructure comprises a first unmanned aerial vehicle (UAV) 10 and a second unmanned aerial vehicle (UAV) 20. The system may also include additional unmanned aerial vehicles, of which only unmanned aerial vehicle 30 is shown in Fig. 1. A base station 40, which communicates with the unmanned aerial vehicles 10, 20, and 30, is provided in a supply vehicle 80 traveling along a road 90. A trailer 85, which serves as a take-off and landing site for the unmanned aerial vehicles 10, 20, and 30, is coupled to the supply vehicle 80. (Furthermore, E.ON SE - 22 - 30A-167 534)

[0098] The trailer 85 belonging to Mitteldeutsche Netzgesellschaft Strom mbH includes a charging station (not shown in Fig. 1) for charging the batteries of the unmanned aerial vehicles 10, 20 and 30.

[0099] Fig. 2 shows a schematic representation of an embodiment of the first unmanned aerial vehicle 10. The first unmanned aerial vehicle 10 comprises a laser-based mapping device 11, which generates map data of the environment of the first unmanned aerial vehicle 10, a transmitter and receiver 12, which sends the map data to the base station 40 in real time, and a battery 19. Further devices may be provided in the unmanned aerial vehicle 10.

[0100] Figure 3 shows a schematic representation of an embodiment of the second unmanned aerial vehicle 20. The second unmanned aerial vehicle 20 comprises one or more of the following devices: a transmitter and receiver 21, a camera 23, a gimbal 24 supporting the camera 23, a lidar system 25, a processing device 26, a navigation system 27, and a battery 29. Further devices may be provided in the unmanned aerial vehicle 20. The second unmanned aerial vehicle 20 does not include a laser-based mapping device.

[0101] The transmitting and receiving device 21 receives a predefined flight path from the base station 40 for inspecting the linear infrastructure 50, 60, 70. The lidar system 25 scans the linear infrastructure 50, 60, 70 and generates 3D point cloud data. The processing device 26 receives the 3D point cloud data from the lidar system 25, analyzes the 3D point cloud data to detect features of interest on the linear infrastructure 50, 60, 70, calculates camera angles and / or distances to the features of interest based on the detected features of interest, and generates commands for the gimbal 24 based on the calculated camera angles and / or distances to the features of interest.

[0102] The navigation system 27 controls a flight path of the second unmanned aerial vehicle 20 based on the camera angles and / or distances to the features of interest calculated by the processing device 26. The gimbal 24 is configured to change the orientation of the camera 23 in real time based on the commands of the processing device 26.

[0103] As can be seen in Fig. 1, the first unmanned aerial vehicle 10 initially starts a new inspection mission on its own. To do this, it takes off from the trailer 85, flies to the linear infrastructure 50, 60 and 70 and generates map data of the surroundings of the first E.ON SE - 23 - 30A-167 534

[0104] Mitteldeutsche Netzgesellschaft Strom mbH unmanned aerial vehicle 10. The transmitting and receiving device 12 sends the map data in real time to the base station 40.

[0105] Fig. 4 shows a schematic representation of an embodiment of the base station 40. The base station 40 comprises a receiving device 41, which receives the map data from the first unmanned aerial vehicle 10, a processing device 42, which calculates the predetermined flight path of the second unmanned aerial vehicle 20 based on the map data, and a transmitting device 43, which sends the predetermined flight path to the second unmanned aerial vehicle 20.

[0106] Immediately after receiving the specified flight path, the second unmanned aerial vehicle (UAV) 20 begins inspecting the linear infrastructure 50, 60, 70. Thus, after receiving the specified flight path, the second UAV 20 can take off from the trailer 85 and fly to the linear infrastructure 50, 60, 70 to inspect it, for example, with camera 23. However, it is also conceivable that the second UAV 20 is already airborne and flies to the linear infrastructure 50, 60, 70 after receiving the specified flight path. This results in a time saving in completing the inspection mission.

[0107] The first unmanned aerial vehicle (UAV) 10 continuously transmits new map data to base station 40 in real time, and the second UAV 20 continuously receives new, predetermined, or modified flight paths for inspecting the linear infrastructure 50, 60, and 70 in real time and flies accordingly. This continuous synchronization allows the inspection mission to be constantly optimized.

[0108] According to a further development of the embodiment of Fig. 1, the second unmanned aerial vehicle 20 receives coordinate data of the supply vehicle 80 and begins an approach to the supply vehicle 80 depending on its battery charge level and distance to the supply vehicle 80.

[0109] The supply vehicle 80 with trailer 85 travels along road 90 and is equipped to change or charge the battery 29 of the second unmanned aerial vehicle 20. The second unmanned aerial vehicle 20 can also be replaced by another unmanned aerial vehicle. E.ON SE - 24 - 30A-167 534

[0110] Central German Electricity Network Company Ltd.

[0111] The third unmanned aerial vehicle 30 shown in Fig. 1 can be identical in construction to the second unmanned aerial vehicle 20. Thus, after the second unmanned aerial vehicle 20 has begun its approach to the supply vehicle 80, i.e., after the second unmanned aerial vehicle 20 has begun its return, the third unmanned aerial vehicle 30 can continue the flight of the second unmanned aerial vehicle 20 along the predetermined flight path.

[0112] Base station 40 is located in supply vehicle 80. Alternatively, base station 40 can also be located in trailer 85 connected to supply vehicle 80. Base station 40 is configured to define flight paths for a large number of unmanned aerial vehicles (UAVs) 10, 20, or 30 and to modify these flight paths in real time. This allows for further optimization of the flight paths.

[0113] According to another further development of the embodiment of Fig. 1, the second unmanned aerial vehicle 20 uses the camera 23, the gimbal 24, the lidar system 25, the processing device 26 and the navigation system 27 for an optimization of the recording of features of interest by the camera 23.

[0114] The processing device 26 can, in particular, perform the following tasks: receiving the 3D point cloud data from the lidar system 25, analyzing the SD point cloud data to identify features of interest in the linear infrastructure, calculating, based on the identified features of interest, camera angles and / or distances to the features of interest, and generating, based on the calculated camera angles and / or distances to the features of interest, commands for the gimbal 24. The processing device 26 can also be configured to determine obstacles and / or a structure of the linear infrastructure 50, 60, 70.

[0115] Based on commands from the processing device 26, the gimbal 24 changes the orientation of the camera 23 in real time, ensuring optimal alignment of the camera 23 with respect to the features of interest. For example, the second unmanned aerial vehicle 20 flies to the power pylon 70 to inspect an insulator, and the processing device 26 optimizes the camera angle of the camera 23 with respect to the insulator and the distance flown by the camera 23 to the insulator during the flight, so that the camera 23 can capture the insulator in the best possible way. In particular, the processing device 26 can continuously adjust the camera angle and / or distances to the features of interest based on real-time received and E.ON SE - 25 - 30A-167 534

[0116] Mitteldeutsche Netzgesellschaft Strom mbH analyzed 3D point cloud data and updated it while the second unmanned aerial vehicle flew 20 around the insulator.

[0117] For even better imaging of features of interest, for example a possibly defective insulator on a power pole, the second unmanned aircraft 20 includes the navigation system 27, which is designed to control the flight path of the second unmanned aircraft 20 based on the camera angles and / or distances to the features of interest calculated by the processing device.

[0118] The navigation system 26 can further control the flight path of the second unmanned aerial vehicle 20, based on the obstacles determined by the processing device 25 and / or the structure of the linear infrastructure 50, 60, 70, in such a way that the imaging of the feature of interest, for example the insulator on the power pole, by the camera 23 is further optimized.

[0119] For further improved imaging of the feature of interest, for example the insulator on the power pole, by the camera 23, the processing device 26 and the navigation system 26 can be configured to synchronize the alignment of the camera 23 with the flight path of the second unmanned aerial vehicle 20 in such a way that the camera 23 records the features of interest with the best possible angle and distance.

[0120] The camera 23 can also be set up to automatically adjust its focus and / or image settings based on the detected features of interest in order to optimize its image and / or video quality.

[0121] The processing device 25 can also use machine learning algorithms to optimize the recognition of features of interest and, based on this, the camera angles.

[0122] The second unmanned aerial vehicle 20 may also have communication means for transmitting images and / or videos recorded by the camera 23 to the base station 40. This can be done, for example, by the transmitter and receiver 21. The images and / or videos can then be further processed in the base station 40. E.ON SE - 26 - 30A-167 534

[0123] Central German Electricity Network Company Ltd.

[0124] Fig. 5 shows a schematic representation of an embodiment of the second unmanned aerial vehicle 20 inspecting an insulator 61 on a power pole 60 of a high-voltage line, see also Fig. 1.

[0125] The second unmanned aerial vehicle 20 flies around the insulator 61 of the power pylon 60. Simultaneously, the gimbal 23 adjusts the camera angle of camera 23 so that the camera can always optimally capture the insulator 61. In particular, the flight path of the second unmanned aerial vehicle 20 is continuously synchronized with the camera angle of camera 23, enabling the acquisition of comprehensive image and video data of the insulator 61. Because the second unmanned aerial vehicle 20 not only changes its flight path around the insulator 61 but also adjusts the camera angle of camera 23, an optimized digital image of the insulator is possible, allowing for better detection of defects in the insulator 61.

[0126] In the examples presented, different features and functions of the present disclosure have been described separately as well as in specific combinations. It is understood, however, that many of these features and functions can be freely combined with one another, unless explicitly excluded.

[0127] Although the embodiments described above relate to power lines with pylons and insulators, the present invention can be applied to any type of linear infrastructure.

Claims

E.ON SE - 27 - 30A-167 534 Central German Electricity Network Company Ltd. REQUIREMENTS 1. System for inspecting linear infrastructure (50, 60, 70) comprising: a first unmanned aerial vehicle (10) comprising: a laser-based mapping device (11) that maps the The first unmanned aerial vehicle (10) generates a surroundings and comprises a transmitting device (12) which sends the map data, in particular in real time, to a base station (40) and a second unmanned aerial vehicle (20) comprising: a receiving device (21) which receives from the base station (40) a predetermined flight path for inspecting the linear infrastructure (50, 60, 70), wherein the second unmanned aerial vehicle (20) flies along the predetermined flight path.

2. System according to claim 1, further comprising: the base station (40) comprising: a receiving device (41) that receives the map data from the first unmanned aircraft (10), a processing device (42) that calculates the predetermined flight path of the second unmanned aircraft (20) based on the map data, and a transmitting device (43) that sends the predetermined flight path to the second unmanned aircraft (20).

3. System according to one of the preceding claims, wherein the second unmanned aircraft (20) begins the inspection of the linear infrastructure (50, 60, 70) immediately after receiving the predetermined flight path and / or the second unmanned aircraft (20) does not include a laser-based mapping device (11).

4. System according to one of the preceding claims, wherein the first unmanned aerial vehicle (10) continuously, in particular in real time, transmits new map data to the base station (40) and the second unmanned aerial vehicle (20) continuously, in particular in real time, receives new, predetermined flight paths for inspecting the linear infrastructure (50, 60, 70) and flies according to them.

5. System according to one of the preceding claims, wherein E.ON SE - 28 - 30A-167 534 Mitteldeutsche Netzgesellschaft Strom mbH the second unmanned aircraft (20) receives coordinate data of a supply vehicle (80) and, depending on the battery charge level of the second unmanned aircraft (20) and the distance of the second unmanned aircraft (20) to the supply vehicle (80), begins an approach to the supply vehicle (80).

6. System according to claim 5, further comprising: the supply vehicle (80) which travels along a road (90) and is equipped to change the battery of the second unmanned aircraft, to charge the battery of the second unmanned aircraft and / or to replace the second unmanned aircraft (20) with another unmanned aircraft (30).

7. System according to claim 5 or 6, wherein the base station (40) is arranged in the supply vehicle (80) or in a trailer (85) connected to the supply vehicle (80), the supply vehicle (80) or the trailer (85) connected to the supply vehicle (80) provides a landing site for unmanned aerial vehicles and / or the base station (40) is configured to specify flight paths for a plurality of unmanned aerial vehicles and to change the specified flight paths, in particular in real time.

8. System according to one of claims 5 to 7, further comprising: a third unmanned aircraft (30) which, after the start of the approach of the second unmanned aircraft (20) to the supply vehicle (80), continues the flight of the second unmanned aircraft (20) along the predetermined flight path.

9. System according to one of the preceding claims, wherein the second unmanned aerial vehicle (20) comprises: a camera (23), a gimbal (24) carrying the camera (23), a light detection and ranging (Lidar) system (25) configured to scan the linear infrastructure and generate 3D point cloud data, and a processing device (26) configured to receive the 3D point cloud data from the Lidar system (25), analyze the 3D point cloud data to detect features of interest on the linear infrastructure, and calculate, based on the detected features of interest, camera angles and / or distances to the E.ON SE - 29 - 30A-167 534 Central German Electricity Network Company Ltd. Features of interest, and generating, based on the calculated camera angles and / or distances to the features of interest, commands for the gimbal (24), wherein the gimbal (24) is configured to change the orientation of the camera (23), in particular in real time, based on the commands of the processing device (26).

10. System according to claim 9, wherein the processing device (26) is further configured to continuously update the camera angle and / or distances to the features of interest based on 3D point cloud data received and analyzed in real time while the second unmanned aerial vehicle (20) flies around the linear infrastructure (50, 60, 70).

11. System according to claim 9 or 10, wherein the second unmanned aircraft (20) further comprises a navigation system (27) configured to control, based on the camera angles and / or distances to the features of interest calculated by the processing device (26), the flight path of the second unmanned aircraft (20).

12. System according to claim 11, wherein the processing device (26) is further configured to determine obstacles and / or a structure of the linear infrastructure (50, 60, 70), and the navigation system (27) is further configured to control, based on the obstacles and / or the structure of the linear infrastructure (50, 60, 70) determined by the processing device (26), the flight path of the second unmanned aircraft (20).

13. System according to claim 11 or 12, wherein the processing device (26) and the navigation system (27) are configured to synchronize the orientation of the camera (23) with the flight path of the second unmanned aircraft (20) such that the camera (23) captures the features of interest.

14. System according to any one of claims 9 to 13, wherein the camera (23) is configured to automatically adjust its focus and / or image settings based on the detected features of interest in order to improve image and / or video quality. E.ON SE - 30 - 30A-167 534 Central German Electricity Network Company Ltd.

15. System according to any one of claims 9 to 14, wherein the processing device (26) is configured to use machine learning algorithms to optimize the recognition of the features of interest and, based on this, the camera angles.

16. System according to one of claims 9 to 15, wherein the second unmanned aerial vehicle (20) further comprises communication means (21) for Includes sending images and / or videos captured by the camera (23) to the base station (40).

Images