METHOD AND ARRANGEMENT FOR CONDITION MONITORING OF A PLANT CONDITIONING PRODUCTS

Description

[0001] The invention relates to a method according to the preamble of claim 1 and an arrangement according to the preamble of claim 9.

[0002] Infrastructure such as high-voltage power lines, wind turbines, or bridge structures are typically large. High-voltage pylons with heights of around 100 meters are not uncommon. Visual inspection by trained personnel, sometimes using optical magnification aids such as binoculars, does not meet the requirement for continuous and verifiable documentation and can generally only be carried out sporadically with reasonable effort. Furthermore, a temporal analysis of damage progression through change detection between different measurement points is not possible.

[0003] If automated image acquisition of equipment such as insulators on high-voltage overhead power lines is to be carried out, current visual inspection and assessment procedures must ensure that even surface details revealing the smallest defects are captured. This requires high image resolution. With current camera technology, a single camera cannot simultaneously capture the entire system and achieve the required surface resolution.

[0004] While a sequential flight over individual altitude ranges seems possible for vertical structures such as wind turbines, this is uneconomical for linear structures such as high-voltage power lines in terms of flight management, as power line networks spanning hundreds of kilometers have to be surveyed.

[0005] The aerial surveying of high-voltage power lines or other installations using laser scanning devices (so-called LIDAR, in particular "Airborne Laser Scanning" (ALS)) and image acquisition (aerial photogrammetry) with subsequent visual inspection by a technician has been standard practice for more than a decade. Within the scope of this invention, the term "aerial photograph" is used synonymously with "aerial image." For full-frame aerial image acquisition, cameras with up to 100 megapixels are now available. By way of example, the iXU / iXU-RS 1000 type aerial cameras are known from the product brochure "Phase One Aerial Cameras, Fully integrated Photography Solutions" by PhaseOne Industrial, Roskildeej 39, DK-2000 Frederiksberg, Denmark, http: / / industrial.phaseone.com / iXU_camera_system.aspx.However, considering the overall height of the infrastructure, the use of such a camera does not result in resolutions sufficient to detect subtle surface alterations, such as soot marks, cracks, or paint chips, in the millimeter range. Therefore, there is no known way to obtain sufficiently high-resolution detailed images of the equipment to be monitored without a trained technician.

[0006] From the publication by Christoph Waldhauser et al., "Automated Classification of Airborne Laser Scanning Point Clouds", Springer Proceedings in Mathematics & Statistics, Vol. 97, pp. 269-292, September 2014, it is known to use an aircraft to capture objects on the ground with a laser scanning device in the form of a point cloud dataset. Furthermore, the authors Jie and Toth describe in "Topographic laser ranging and scanning: principles and processing", published by CRC press, 2008, the use of airborne laser scanning (ALS) from a (manned or unmanned) flying platform to capture underlying objects in the form of a so-called 3D point cloud.

[0007] Document WO 2016 / 149513 A1 describes the process of using a small drone to inspect a power line pylon, recording data on its operational status using cameras and sensors. Document EP 1548908 A1 also deals with infrastructure inspections.

[0008] Based on known methods for capturing high-resolution aerial photographs of plants, the invention aims to provide an automated method that can be used for condition monitoring of a plant in a comparatively quick and cost-effective manner.

[0009] The invention solves this problem by means of a method according to claim 1.

[0010] The goal is to use the detailed camera images to assess damage to the monitored system. This assessment can be performed manually by a technician in a control center or automatically using suitable software. Pre-processing followed by manual evaluation is also possible. If damage is detected, for example, to an insulator on an overhead power line, repairs can be initiated before the equipment fails. This increases the availability of the system and the entire power grid, thereby saving on downtime costs. Furthermore, it eliminates the need for costly manual inspections of systems like overhead power lines.

[0011] Since the overview sensor array captures the entire infrastructure or system, individual infrastructure elements (e.g., insulators) are still relatively coarsely resolved in a point cloud dataset and an image dataset. Therefore, no details of these infrastructure elements are yet recognizable in these images. Within the scope of the invention, the term point cloud dataset refers to a 3D point cloud acquired using a laser scanning device.

[0012] A camera capable of capturing aerial images from a typical overflight distance of a few dozen to a few hundred meters can be used for detail photography, revealing even the smallest defects in the millimeter range. For this purpose, a longer focal length, i.e., a telephoto lens, can be used compared to an overview camera. For example, such damage could be a 2.5 cm long crack in an insulator shield. Using a color camera is particularly advantageous. Furthermore, it is beneficial if the camera is mounted on a swivel and tilt mechanism to allow for precise alignment with the determined positions of the equipment. The detail camera can also be advantageously tracked along the installation as the vehicle moves, thus counteracting motion blur in the detailed images.For this purpose, the vehicle's speed and direction can be determined based on a precise temporal positioning of the vehicle. The detail camera can be a high-resolution camera. Due to the significantly smaller field of view compared to an overview camera, the resolution of the detail camera can potentially even be lower than that of an overview camera.

[0013] A suitable positioning device, such as a satellite-based positioning system like GPS or Galileo, can be used in the vehicle. A position as defined by the invention always specifies not only the geographical latitude and longitude, but also the altitude (e.g., above sea level). This allows for precise temporal positioning, particularly down to the millimeter, as the vehicle moves along the system. It is preferred that the positioning device can determine the position with a resolution of less than one meter. In this respect, the so-called GNSS / RTK system is more accurate than GPS.

[0014] The evaluation system can be provided, for example, using conventional computer equipment and data storage. It can be located on board the vehicle or centrally at a control center.

[0015] The vehicle used could be, for example, a truck driving along a road next to an overhead power line. However, the vehicle could also be an aircraft such as a plane, helicopter, or similar. In the case of an aircraft, aerial photographs are taken. The vehicle can be manned or unmanned, and possibly autonomous.

[0016] The key advantage of the inventive method compared to previous solutions lies in the automation of the recording process. First, the relevant infrastructure elements are automatically located and then automatically recorded in detail. Overview and detailed recordings are cleverly combined to obtain a complete overview of the infrastructure. Potentially particularly relevant areas are selectively captured at a higher resolution, allowing for differentiated analyses and assessments.

[0017] The brochure of the same name mentions the "IGI Urban Mapper" camera, manufactured by IGI mbH, Langenauer Str. 46, D-57223 Kreuztal, www.IGI-SYSTEMS.com, which was developed for creating three-dimensional cityscapes. However, to capture infrastructure such as an overhead power line with its associated equipment, several cameras would need to be combined, for example, to capture an entire pylon. This is comparatively very expensive and, with its correspondingly high weight and space requirements, places high demands on an aircraft.

[0018] As an alternative to the invention, a high-resolution line scan camera could also be used, but this would present the problem of aligning the recorded image lines with extremely high precision, as otherwise distortions would be induced. In summary, it can be stated that the existing approaches are not suitable for the application of automated monitoring of infrastructure with large vertical and horizontal extents, such as an overhead power line.

[0019] According to the inventive method, a laser scanning device is used for the overview sensor arrangement to acquire a point cloud dataset. A laser scanning device is, for example, a LiDAR system, such as those used in (semi-)autonomous vehicles to obtain three-dimensional models of the vehicle's surroundings. The laser measures distance values ​​from the sensor to objects in the environment, so that a point cloud dataset results from a multitude of measurements. If the position of the sensor or the vehicle is known, the position of a point from the point cloud dataset can be reconstructed very accurately by referring to the position of the laser scanning device or the flying object and the direction in which the laser scanning device is aligned. In dynamic measurement methods, such as...Mobile laser scanning (MLS) and airborne laser scanning (ALS) utilize laser scanners in conjunction with a GNSS / INS system (Global Navigation Satellite System or Inertial Navigation System). This serves to determine a vehicle trajectory in order to capture the vehicle's surroundings as a 3D point cloud. If the relative orientation between the GNSS / INS system and the laser scanner is known, a 3D point cloud can be generated by combining the vehicle trajectory with the laser scan measurements (distance and direction).

[0020] It is advisable to scan the entire vertical dimension of the system with the laser scanner, for example, along the entire height of a pylon. Typically, laser scanners are now always used for condition monitoring of systems using aircraft, because they can also measure the sag of a line and its distance from the ground. Furthermore, they can monitor how close vegetation grows to the system, for example, whether trees or shrubs are growing up towards an overhead power line.

[0021] According to the inventive method, an overview camera is used for the overview sensor arrangement to acquire an image data set. The overview camera used for acquiring an image data set can be a camera which, by design, possibly through a suitable magnification mechanism or a corresponding lens, is capable of imaging a similarly sized section of the vehicle's surroundings as the laser scanning device. The resolution should be sufficient to make objects ranging in size from a few centimeters to decimeters recognizable, at least in outline. These could be, for example, equipment such as insulators, but also foreign objects such as bird nests. Preferably, a wide-angle lens is used for the overview camera. Furthermore, it is advantageous if the overview camera is oriented essentially the same way as the laser scanning device if the latter is also used in the overview sensor arrangement.This can be achieved, for example, by mounting the laser scanner and overview camera together on a mounting device so that they always point in the same direction. The orientation of the laser scanner and the overview camera, or, if a shared mounting device is used, the orientation of the mounting device itself, can be fixed relative to the vehicle when using a vehicle, preventing any change in the "viewing direction" during data collection. Alternatively, a view that can be changed during data collection can be achieved by mounting the mounting device or the two individual sensors (laser scanner and overview camera) so that they can be rotated and swiveled. This would allow, for example, when using an aircraft as a vehicle, the direction relative to the flight direction or to a flight trajectory known from previous flight planning to be adjusted using electric motors or similar devices.becomes possible.

[0022] As an alternative to essentially aligning the laser scanner and overview camera in the same direction, a laser scanner designed for a comparatively large directional range, e.g., 330°, can be used. This allows the laser scanner to always collect sufficient 3D point cloud data, even without being aligned with the viewing direction of an overview camera, to combine it with camera images.

[0023] The orientation of the laser scanning device and / or overview camera can be selected, for example, with respect to a longitudinal axis through the aircraft, such that when the aircraft is flying straight, i.e., in a straight line and parallel to the Earth's surface, the viewing angle is inclined downwards (e.g., between 30° and 70° to the Earth's surface). This has the advantage that a structure covering a large area, such as an overhead power line, is easily captured. Furthermore, the orientation can be such that the viewing angle is tilted to one side of the flight trajectory, resulting in an angle between 30° and 70° to the Earth's surface, for example, to the right or left. This has the advantage that the equipment and other objects are captured from the side instead of directly from above, making their three-dimensional shape more easily recognizable and the objects easier to identify.For example, when viewed from an oblique angle during a flyover, an insulator appears, to a first approximation, as a tube between the pylon and the overhead power line. Directly from the front, however, an insulator appears only as a rectangle, and from directly above, only as a circle.

[0024] When an overview camera and a laser scanning device are used together, it is advantageous to spatially superimpose or register the image data set and the point cloud data set in such a way that, for example, the point cloud data and the image data from the overview camera are stored in relation to each other in an internal data storage device of the vehicle, so that an object in the image can be assigned an exact position using the point cloud. In a preferred embodiment of the method according to the invention, visible light is captured by the overview camera. This is advantageous because many objects in the visible spectrum are easily recognizable, for example, objects with a different color than their surroundings.

[0025] In a further preferred embodiment of the method according to the invention, infrared and / or ultraviolet light is detected by means of the overview camera. This is advantageous because, outside the visible spectrum, and especially by means of thermal radiation in the IR range, electrical equipment heated during operation, for example, is easily recognizable as a heat source from the surroundings.

[0026] According to the inventive method, overview data is acquired at a first time point, and later, at a second time point and after determining the positions of the equipment, detailed images are acquired. This is advantageous because this approach is particularly simple. First, overview data, i.e., point cloud data, is acquired using a single vehicle. This data can then be evaluated, for example, offline after the vehicle returns to the control center, and the positions of the relevant equipment can be determined. Afterward, the system can be surveyed again with the same vehicle to capture detailed images using the detail camera.

[0027] In a further preferred embodiment of the method according to the invention, the evaluation unit is provided in the single vehicle. This is advantageous because, with appropriate design of the evaluation unit, processing of the overview data is possible almost in real time, so that in a single data acquisition process, e.g., a single flight over an overhead power line system, the positions of the equipment can be determined directly and the detailed images captured. This saves time, costs, and effort for a second data acquisition process for the detailed images. The evaluation unit can be implemented with conventional computer systems and data storage devices.

[0028] The vehicle collects and stores data sets during an initial data acquisition process using the laser scanning device. After this first acquisition, the stored data sets can be read and processed to determine the positions of the equipment and subsequently plan a second data acquisition process for the system. During this second acquisition process, the detailed aerial camera is used. For this configuration, it is sufficient to equip the single vehicle with the laser scanning device for the first acquisition process and with the detailed aerial camera for the second.

[0029] Alternatively, the evaluation unit can be located in a control center for condition monitoring. This is advantageous because the determination of the equipment positions described above does not need to be performed in the vehicle. This saves weight and energy consumption in the vehicle, making it comparatively more cost-effective, smaller, and suitable for longer distances. In a preferred embodiment, a communication link between the vehicle and the evaluation unit or control center can be maintained essentially continuously, so that the vehicle transmits overview data to the evaluation unit and receives the estimated positions of the equipment back from it.In this way, detailed images can be obtained directly in a single data acquisition process, saving time, costs and effort for a second data acquisition process for the detailed images.

[0030] In another preferred embodiment of the method according to the invention, a cloud evaluation unit is provided. The cloud evaluation unit can, for example, be provided as a cloud application and access external data centers for position detection. For this purpose, it is advantageous if the aircraft enables permanent data communication with the cloud, e.g., via a satellite network.

[0031] In a further preferred embodiment of the method according to the invention, a drone is used as the sole vehicle. A drone, as defined in the invention, is an unmanned aerial vehicle. A suitable example is a multicopter drone, which flies relatively slowly at a low altitude, e.g., 30-40 km / h, along the site for data acquisition and / or detailed recording. This has the advantage that an unmanned drone eliminates the risk that a human pilot would have to bear. If, for example, a crash occurs, only material damage is generally to be expected. The drone can be remotely controlled from a control center via radio, as is common practice with military drones today. It is preferred if the drone automatically flies a predetermined route at a predetermined altitude. This is particularly cost-effective. The flight route can be planned automatically, for example, based on the site plans of the facility operator.

[0032] Alternatively, a conventional flying object such as an airplane or a helicopter with a pilot can also be used.

[0033] Both the first and second vehicles used are a single unit equipped with the overview sensor array and the detail camera. This is advantageous because only one vehicle is needed to perform the procedure, saving on manufacturing and configuration costs.

[0034] In a further preferred embodiment of the method according to the invention, the detail camera is used in the same acquisition process as the laser scanning device and the overview camera. This is advantageous because it eliminates the need for a second acquisition process, such as another overflight with an aircraft. A prerequisite for this approach is that the evaluation unit provides a sufficiently fast evaluation of the position of the objects to be filmed, such as equipment, so that the detail camera can be pointed directly at them.

[0035] In a further preferred embodiment of the method according to the invention, the dimensions of the equipment are additionally determined. This is advantageous because, in addition to the position, the three-dimensional dimensions are then also known, so that the detail camera can be guided in such a way that particularly large objects can be photographed completely.

[0036] The system consists of an overhead electrical line with masts and cables, with electrical equipment being arranged on the masts.

[0037] Electrical insulators are used as operating equipment on the masts.

[0038] In a further preferred embodiment of the method according to the invention, the evaluation unit uses an automatic classification method trained on example data to detect the conductors, create a simplified line model of the detected conductors, and identify the positions of the insulators at the contact point of two modeled conductors. This is done by first modeling the conductor path, i.e., by mapping it using a so-called catenary or rope curve. This has been known for a long time and is explained, for example, on Wikipedia: https: / / de.wikipedia.org / wiki / Kettenlinie_(Mathematik). For example, the connection point of an insulator can be determined from the transition points between points of the class "conductor" and "pylon" or from the intersection points of two catenary curves of successive conductor sections. The automatic classification method can, for example, use so-called...Semantic classification of the captured point cloud is performed; that is, the points in the point cloud dataset are assigned to predefined classes—in particular, the class "line." This step can be accomplished using various approaches. One possible method is machine learning. Here, features describing the spatial structure are first collected for each point cloud data point, and then a classifier is trained on this feature space. A suitable approach for an automated classification procedure is described in the publication by Christoph Waldhauser et al., "Automated Classification of Airborne Laser Scanning Point Clouds," Springer Proceedings in Mathematics & Statistics, Vol. 97, pp. 269–292, September 2014.

[0039] Another suitable approach is described in the publication "Fast semantic segmentation of 3D point clouds using a dense CRF with learned parameters" by Daniel Wolf et al., International Conference on Robotics and Automation, pp. 4867-4873, 2015. The described method, based on a dataset of images of an interior space, is able to recognize objects, taking into account, in particular, that certain objects, such as a chair and a table, are often located close together. Applied to the present example of an overhead power line, an insulator would be expected, for instance, where a pylon and a line were also detected.

[0040] Furthermore, the publication "Fast semantic segmentation of 3d point clouds with strongly varying density" by Timo Hackel et al., ISPRS Annals - ISPRS Congress, Prague, 2016, also addresses the topic of object detection in point cloud datasets.

[0041] The evaluation unit uses an automatic classification procedure, trained on sample data, to detect the lines and pylons. A simplified line model of the detected lines and pylons is then created, and the positions of the insulators are identified at the contact point between the pylons and the lines. The same solution approaches can be used as described for the previous embodiment.

[0042] In a further preferred embodiment of the method according to the invention, the evaluation unit estimates the position of a piece of equipment on a mast by determining the position of the equipment at the intersection of a line extrapolated from the line segment with a vertical line through the mast's position, based on a previously known position of the mast and a previously detected section of a line leading to that mast. This embodiment has the advantage of predictively estimating the relevant positions on a mast to be monitored, even before the overview sensor arrangement has detected the mast. This allows the detailed camera to begin collecting detailed images as soon as it approaches the mast in question. This increases the probability that at least one of the detailed images will depict the equipment to be monitored with high resolution and good visibility.Thus, the predictive alignment of the detail camera increases the reliability of the condition monitoring and therefore the availability of the system.

[0043] In a further development of the aforementioned embodiment, a previously determined installation height of the cable is taken into account by defining the position of the device at an intersection of the extrapolated cable path with a horizontal plane at the installation height. This is advantageous because it improves the accuracy of the position determination.

[0044] In an alternative preferred embodiment of the method according to the invention, the evaluation unit estimates the position of a piece of equipment on a pylon by determining the position of the equipment at the intersection of a vertical line through the pylon's position and a horizontal plane at the mounting height, based on a previously known position of the pylon and a previously recorded mounting height. This embodiment has the advantage that, after data acquisition from one pylon, the positions of the equipment in a series of pylons of an overhead power line can be estimated if the mounting height is similar. This is the case, for example, in flat terrain. In hilly or mountainous terrain, an elevation profile can be taken into account in such a way that even mounting heights that differ from pylon to pylon can be processed.

[0045] In a further preferred embodiment of the method according to the invention, at least one of the following objects is detected and captured with detailed images, either as an alternative to or in addition to the operating equipment: signs on the poles, anomalies on the lines, attachments to the line and / or the poles, anomalies on a pole, bird nests. The other objects are detected and their positions are stored in the same manner as described above for insulators.

[0046] In a further preferred embodiment of the method according to the invention, the detected equipment is assigned to spatial clusters based on its position by means of the evaluation unit, in order to reduce the number of alignment operations required by the detail camera. The detail camera or the lens is designed such that, if, for example, a typical insulator dimension is known, the insulator is captured in its entirety and in its entirety. Capturing insulator groups or clusters when insulators are located close together, thus reducing the required positioning of the detail sensor, saves time and costs during data acquisition.

[0047] In a further preferred embodiment of the method according to the invention, the evaluation unit detects the equipment based on the image data set, whereby the respective positions of the equipment are determined using example data, taking the point cloud data set into account. This approach of direct detection of insulators or other objects in the image data uses backprojection into 3D space via the known relationship between the image and the coordinates provided by the point cloud data set. For example, an approach can be followed as described by Joseph Redmon et al. in the publication "You Only Look Once: Unified, Real-Time Object Detection.", Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779-788. The algorithm can be trained directly using example images of equipment, etc., taken from different perspectives.

[0048] A similar and equally suitable approach is known from the publication "Faster R-CNN: Towards real-time object detection with region proposal networks" by Shaoquing Ren, Advances in neural information processing systems (pp. 91-99), 2015.

[0049] Furthermore, starting from known arrangements for the manual acquisition of high-resolution aerial photographs of plants, the invention aims to provide an arrangement with which condition monitoring of a plant can be carried out automatically, relatively quickly and cost-effectively.

[0050] The invention solves this problem by means of an arrangement according to claim 9. Preferred embodiments are set forth in claims 10 to 14. The arrangement according to the invention and its embodiments offer the same advantages as explained above for the method according to the invention.

[0051] The invention will now be illustrated with exemplary embodiments and accompanying schematic figures for better explanation. Figure 1 shows an embodiment of an arrangement according to the invention, and Figure 2 shows a detailed view of a flying object according to the arrangement. Figure 1 , and Figure 3 shows an enlarged section of a high-voltage pylon with insulators according to Figure 1 , and Figure 4 shows an embodiment of a predictive position estimation of equipment.

[0052] The Figure 1 Figure 1 shows an embodiment of an arrangement according to the invention for condition monitoring of a system with equipment 7. In this example, the system consists of high-voltage pylons 1 with high-voltage overhead lines 29, which are strung between the pylons 1. The pylons 1 stand freely on the ground 2. The pylons 1 have a typical height 3. In the left part of the Figure 1A mast 1 is shown schematically. It has three cross members 4, 5, 6 at different heights, each with two insulators 7 suspended from it. The high-voltage overhead lines 29 are attached to the insulators 7. For clarity, only two overhead lines 29 are shown.

[0053] A flying object 9 is used to obtain detailed images of the insulators 7. Based on these detailed images, damage to the insulators 7 can be detected manually or mechanically, so that maintenance or replacement of a damaged insulator 7 can be carried out in a timely manner before an insulator fails.

[0054] The aircraft 9 is an unmanned drone that automatically flies over the installation with masts 1 and overhead power lines 29. The drone 9 has a position detection device 10 that receives Global Positioning System (GPS) signals 16 from a GPS satellite 8 and can thus determine the aircraft's position with high accuracy at any given time. The aircraft 9 has an overview sensor array 17, 18, which is configured as an overview camera 17 and a laser scanner 18. Furthermore, the aircraft 9 has a detail camera 19. The detail camera 19, the overview camera 17, the laser scanner 18, and the position detection device 10 are connected to a computer unit 20 via data communication lines 22. The computer unit 20 is connected to a data storage device 21 via a data communication line 22.The computer unit 20 is further connected via a data communication line 22 to a communication unit 11, which is connected via radio signals 15 to a communication network 12. The communication network 12 includes, for example, a cloud platform where data from the aircraft can be evaluated. The communication network 12 is connected via a communication link 14 to a control center 13, where, for example, flight data, overview data obtained by means of the overview sensor unit 17, 18, and detailed images obtained by means of the detail camera 19 are permanently stored and made available for further processing.

[0055] The following section briefly explains the functionality of the flying object 9 in more detail. When flying over a high-voltage pylon 1, the appropriately aligned overview sensor array 17, 18 has a wide-angle field of view 23, capable of capturing the entire height of the pylon 1. Using the overview data—that is, images from the overview camera 17 and point cloud data from the laser scanning device 18—the computer unit 20 can estimate the positions of the insulators 7. The detail camera 19 can then be aligned with a comparatively narrow field of view (not shown) to capture detailed images of the insulators. For example, a telephoto lens can be used. The detailed images are transmitted via the communication unit 11 to the control center 13 over the network 12.

[0056] The Figure 2Figure 1 shows a detailed view of the aircraft 9, which is equipped with a first pan-tilt assembly 52 at the front in the direction of flight 37. The first pan-tilt assembly 52 is attached to a first suspension point 51 and enables the joint alignment of the overview camera 17 and the laser scanning device 18 mounted on the pan-tilt assembly 52. ​​By appropriately aligning the first pan-tilt assembly 52, a viewing angle 23 is obtained with which an entire mast 1 can be captured. Once the position of the equipment of interest, or insulators 7, has been estimated, the detail camera 19, which is mounted somewhat further back on the aircraft, can be aligned. For this purpose, the detail camera 19 is arranged on a second pan-tilt assembly 25 with a second suspension point 26 and can be aligned with high precision to the presumed position 53 of an equipment in order to obtain detailed images.

[0057] Figure 3Figure 1 shows an enlarged detail view of the presumed position 53, where two insulators 7 are located. One of the insulators 7 exhibits damage 54, which is visible in correspondingly high-resolution detail images. By evaluating the detail images, it is possible to detect the damage 54 early and replace the insulator 7. This increases the reliability and availability of the power grid, of which the overhead line is a part.

[0058] Figure 4Figure 1 shows an embodiment for a predictive estimation of the position 39 of an insulator 7. The flying object 9 flies in direction 37 parallel to the ground along an overhead power line. Its overview sensor arrangement has already detected the pylon 31 with the insulator 7. It has also detected the line section 34, which is attached to the insulator at a mounting height 33. In the current field of view 23 of the overview sensor arrangement, there is another line section 35. Using an evaluation device (not shown), an extrapolation of the path of a sagging overhead power line can be performed based on the now known line sections 34 and 35, which is indicated by the dashed curve 36.Typically, all GPS positions of the pylons are known in advance from the energy network operator's records, so the position of the neighboring pylons 32 in the direction of flight 37 is also known to the aircraft 9. If the extrapolated line route 36 is now extended to an intersection 38 with a line running vertically through pylon 32, a position 39 of the insulator 7 on pylon 32 can be assumed, without the aircraft 9 having already detected this area.

[0059] Additionally or alternatively, position 39 can be assumed to be at an intersection of a horizontal plane, i.e. a plane running parallel to the ground 2, at installation height 33 with the vertical line through the mast 32 and / or with the extrapolated line route.

[0060] This method, in its various versions, offers the advantage that the overview camera can be aligned with the area of ​​interest 39 as soon as it first approaches mast 32, in order to capture and save initial detailed images. The more detailed images taken from different angles as the flying object 9 approaches, the sooner damage to the insulator 7 can be detected, or poor image quality due to unfavorable lighting or weather conditions can be compensated for.

Claims

1. Method for monitoring a condition of an installation (1, 4, 5, 6, 7, 29, 31, 32) with operating means (7), in which overview data are captured by means of a first vehicle (9) having an overview sensor arrangement (17, 18) for optically capturing the installation (1, 4, 5, 6, 7, 29, 31, 32), and detailed images of the operating means (7) are generated by means of a second vehicle (9) having a detail camera (19), wherein a single vehicle (9) which has the overview sensor arrangement (17, 18) and the detail camera (19) is used as the first and second vehicle, wherein the operating means (7) are detected in the overview data by means of an evaluation device (20) and the positions of the operating means (7) are determined taking into account the position of the first vehicle (9), and the detail camera (19) is oriented with respect to the respective positions of the operating means (7), and a laser scanning device (18) for capturing a point cloud data set is used for the overview sensor arrangement (17, 18), wherein an electrical overhead power line having masts (1, 31, 32) and lines (29) is used as the installation (1, 4, 5, 6, 7, 29, 31, 32), wherein electrical operating means (7) are arranged on the masts (1, 31, 32), wherein electrical insulators (7) on the masts (1, 31, 32) are used as operating means, characterized in that the lines (29) and the masts (1, 31, 32) are detected using the evaluation device (20) by means of an automatic classification method trained on the basis of sample data, and a simplified line model of the detected lines (29) and masts (1, 31, 32) is created, and the positions of the insulators (7) are detected at the contact point of the masts (1, 31, 32) with the lines (7).

2. Method according to Claim 1, characterized in that an overview camera (17) for capturing an image data set is used for the overview sensor arrangement (17, 18).

3. Method according to one of the preceding claims, characterized in that the overview data are obtained at a first time and the detailed images are obtained subsequently at a second time and after the positions of the operating means (7) have been determined.

4. Method according to one of the preceding claims, characterized in that the evaluation device (20) is provided in the first vehicle, wherein a single vehicle (9) is used as the first and second vehicle.

5. Method according to one of the preceding claims, characterized in that the evaluation device is provided in a control centre (13) for monitoring the condition.

6. Method according to one of the preceding claims, characterized in that an aerial drone (9) is used as the first and the second vehicle, wherein a single vehicle (9) is used as the first and second vehicle.

7. Method according to one of the preceding claims, characterized in that, as an alternative or in addition to the operating means (7), at least one of the further objects is detected and is captured with detailed images: signs on the masts, anomalies on the lines, attachments on the line and / or the masts, anomalies on a mast, birds' nests.

8. Method according to one of the preceding claims, characterized in that the detected operating means (7) and / or objects are assigned to spatial clusters on the basis of their position by means of the evaluation device (20) in order to reduce the number of required orientation operations of the detail camera (19) by means of the clusters.

9. Arrangement for monitoring a condition of an installation (1, 4, 5, 6, 7, 29, 31, 32) with operating means (7), having a first vehicle (9) which has an overview sensor arrangement (17, 18) for optically capturing overview data relating to the installation (1, 4, 5, 6, 7, 29, 31, 32), and a second vehicle (9) designed to generate detailed images of the operating means (7) using a detail camera (19), wherein a single vehicle (9) which has the overview sensor arrangement (17, 18) and the detail camera (19) is used as the first and second vehicle, wherein there is an evaluation device (20) which is designed to detect the operating means (7) in the overview data and to determine the positions of the operating means (7) taking into account the position of the first vehicle (9), wherein the second vehicle (9) is designed to orient the detail camera (19) with respect to the respective positions of the operating means (7), wherein the overview sensor arrangement (17, 18) has a laser scanning device (18) for capturing a point cloud data set, wherein the arrangement is suitable for monitoring the condition of an installation, wherein the installation (1, 4, 5, 6, 7, 29, 31, 32) has an electrical overhead power line having masts (1, 31, 32) and lines (29), wherein electrical operating means (7) are arranged on the masts (1, 31, 32), wherein the operating means have electrical insulators (7) on the masts (1, 31, 32), characterized in that the evaluation device (20) is designed to detect the lines (29) and the masts (1, 31, 32) by means of an automatic classification method trained on the basis of sample data, and to create a simplified line model of the detected lines (29) and masts (1, 31, 32), and to detect the positions of the insulators (7) at the contact point of the masts (1, 31, 32) with the lines (29).

10. Arrangement according to Claim 9, characterized in that the overview sensor arrangement (17, 18) has an overview camera (17) for capturing an image data set.

11. Arrangement according to either of Claims 9 and 10, characterized in that the first vehicle (9) is designed to obtain the overview data at a first time, and in that the second vehicle (9) is designed to obtain the detailed images at a subsequent, second time and after the positions of the operating means (7) have been determined, wherein a single vehicle (9) is used as the first and second vehicle.

12. Arrangement according to one of Claims 9 to 11, characterized in that the evaluation device (20) is provided in the first vehicle, wherein a single vehicle (9) is used as the first and second vehicle.

13. Arrangement according to one of Claims 9 to 12, characterized in that the first and the second vehicle are an aerial drone (9), wherein a single vehicle (9) is used as the first and second vehicle.

14. Arrangement according to one of Claims 9 to 13, characterized in that the evaluation device (20) is designed to estimate the position of an operating means (7) on a mast by determining, on the basis of a previously known position of the mast (32) and an already captured section (34) of a line to this mast (32), the position of the operating means (7) at an intersection point (38) of a line course (36) extrapolated from the section with a vertical line through the position of the mast (32).