Indoor power distribution station equipment infrared monitoring method based on camera cruise

Abstract

The invention discloses an indoor power distribution station equipment infrared monitoring method based on camera cruise. The method comprises the following steps: 1) determining shooting positions of an overlooking unmanned aerial vehicle and a side-looking unmanned aerial vehicle; 2) planning flight routes of the overlooking unmanned aerial vehicle and the side-looking unmanned aerial vehicle in a cruising process; 3) constructing a top view image infrared monitoring database and a side view image infrared monitoring database; 4) shooting side orientation and horizontal orientation images of the power distribution station; and 5) comparing the pixel value of the reference pixel point, the pixel value of the calibration pixel point of the corresponding power equipment and the alarm threshold value, and making a response according to the comparison result. According to the invention, the temperature distribution is visually displayed through the infrared image, the faulty equipment can be quickly locked, and the troubleshooting time is shortened. According to the invention, automatic and high-frequency total-station scanning can be realized, a blind area is covered, and the omission ratio is reduced.

Description

Technical Field

[0001] The present invention relates to the field of substation monitoring, and specifically to an infrared monitoring method for indoor substation equipment based on camera cruise. Background Art

[0002] During the operation of indoor substation power equipment, temperature rise faults are likely to occur due to reasons such as poor contact, overload, and short circuit.

[0003] Therefore, by real-time monitoring the temperature distribution of power equipment, local overheating phenomena can be detected in a timely manner, thereby avoiding the expansion of faults leading to unplanned power outages and ensuring the continuity of power supply.

[0004] Traditional temperature monitoring in indoor substations relies on manual regular inspections, which have problems such as high labor intensity, long detection cycles, and difficulty in achieving 24-hour continuous monitoring. Especially in substations with complex equipment layouts or high-density distribution rooms, it is easy to miss hidden fault points during manual inspections.

[0005] Although existing fixedly installed infrared cameras can achieve local temperature monitoring, they are limited by the viewing angle and installation position and are difficult to cover the entire surface of the equipment.

[0006] Therefore, it is necessary to propose a cruise-type infrared monitoring method combined with drones. Summary of the Invention

[0007] The purpose of the present invention is to provide an infrared monitoring method for indoor substation equipment based on camera cruise, including the following steps:

[0008] 1) Obtain the spatial layout of the indoor substation and the distribution of power equipment in the substation;

[0009] 2) According to the spatial layout of the indoor substation and the distribution of power equipment in the substation, determine the horizontal azimuth shooting position matrix B1 of the overhead drone and the side azimuth shooting position matrix B2 of the side-view drone; each horizontal azimuth shooting position corresponds to a horizontal azimuth shooting area; each side azimuth shooting position corresponds to a side azimuth shooting area;

[0010] Among them, the horizontal azimuth shooting position matrix B1 and the side azimuth shooting position matrix B2 are as follows:

[0011]

[0012] In the formula, b n is the nth horizontal azimuth shooting position of the overhead drone; (b nx , b ny , b nz ) are the coordinates of the nth horizontal azimuth shooting position b n ; b nzis the flight altitude of the overhead drone at the nth horizontal azimuth shooting position; b' m is the mth side azimuth shooting position of the side-view drone; (b' mx ,b' my ,b' mz ) is the mth side azimuth shooting position b' m coordinates; b' mz is the flight altitude of the side-view drone at the mth side azimuth shooting position;

[0013] 3) Based on the horizontal azimuth shooting position matrix B1 and the horizontal azimuth shooting position matrix B2, plan the flight routes of the overhead drone and the side-view drone during a single cruise, denoted as the first flight route and the second flight route respectively;

[0014] 4) The control terminal obtains the infrared images D = [D1 D2... D n of each horizontal azimuth shooting area of all power equipment in the indoor substation under normal operating conditions, and the infrared images F = [F1 F2... F m of each side azimuth shooting area;

[0015] 5) Select and extract the power equipment area images in each infrared image of the horizontal azimuth shooting area and the side azimuth shooting area, and select k er calibration pixel points of power equipment in each power equipment area image, record the coordinates, pixel values, power equipment coding, power equipment position coding, and shooting position to which these calibration pixel points of power equipment belong, so as to construct the overhead image infrared monitoring database and the side-view image infrared monitoring database, write the overhead image infrared monitoring database into the chip carried by the overhead drone, and write the side-view image infrared monitoring database into the chip carried by the side-view drone;

[0016] The overhead image infrared monitoring database is as follows:

[0017]

[0018] In the formula, E u is the u-th calibration pixel point of the power equipment; E ux ,E uy are the coordinates of the u-th calibration pixel point of the power equipment; PI' u is the pixel value of the u-th calibration pixel point of the power equipment; PDS u is the power equipment coding to which the u-th calibration pixel point of the power equipment belongs; PDS ucoord is the power equipment position coding to which the u-th calibration pixel point of the power equipment belongs; POS u is the shooting position to which the u-th calibration pixel point of the power equipment belongs;

[0019] The infrared monitoring database of the side view image is as follows:

[0020]

[0021] Where M o is the calibration pixel point of the o-th power equipment; M ox , M oy are the coordinates of the calibration pixel point of the o-th power equipment; PI” o is the pixel value of the calibration pixel point of the u-th power equipment; PDS' o is the power equipment code to which the calibration pixel point of the o-th power equipment belongs; PDS' ocoord is the power equipment position code to which the calibration pixel point of the o-th power equipment belongs; POS' o is the shooting position to which the calibration pixel point of the o-th power equipment belongs;

[0022] The steps of selecting k calibration pixel points of power equipment in each power equipment area image include: er 5.1) Extract the pixel values of each pixel point in the current power equipment area image, and write the pixel values and pixel coordinates into the set X to obtain:

[0023] 5.1) Extract the pixel values of each pixel point in the current power equipment area image, and write the pixel values and pixel coordinates into the set X to obtain:

[0024]

[0025] Where X N is the N-th pixel point in the current power equipment area image; X Nx , X Ny are the coordinates of the pixel point X N ; PI N is the pixel value of the pixel point X N ;

[0026] 5.2) Construct the pixel value set PI = {PI1, PI1,..., PI N};

[0027] 5.3) Randomly select k er pixel values in the pixel value set PI as the clustering centers; k er The initial value is 3; the subscript er represents the r-th power equipment area image in the e-th infrared image;

[0028] 5.4) Calculate the distance matrix DIS of each pixel value in the pixel value set PI to each clustering center, that is:

[0029]

[0030] Where dis ls is the l-th pixel value to the s-th cluster center distance;

[0031] 5.5) With the goal of minimizing the distance to the cluster center, assign each pixel value in the pixel value set PI to the cluster where the nearest cluster center is located;

[0032] 5.6) Calculate the mean of each cluster, and use the mean as the new cluster center. Return to step 5.4) until the cluster center remains unchanged or reaches the maximum number of iterations, and then enter step 5.7)

[0033] 5.7) Calculate the clustering cost L er corresponding to k k cluster centers, that is:

[0034]

[0035] In the formula, C s is the s-th cluster; μ s is the cluster center of the s-th cluster; c is the element of the s-th cluster;

[0036] 5.8) Judge whether it holds. If so, enter step 5.9). Otherwise, let k er = k er + 1, and return to step 5.3);

[0037] 5.9) Select the number of cluster centers with the minimum difference cost as the optimal number of cluster centers k er , and use the cluster centers of each cluster corresponding to the optimal number of cluster centers k er as the optimal cluster centers;

[0038] 5.10) Based on the region division objective function (9), select the image segmentation threshold ε, that is:

[0039] ε = min||PI s1 - PI s2 ||(9)

[0040] In the formula, PI s1 , PI s2 are the s1, s2 optimal cluster centers; s1 = 1, 2,..., k er ; s2 = 1, 2,..., k er ; s2 ≠ s1;

[0041] 5.11) Select the first pixel point in the power equipment area image that has not been assigned to any area as the seed point, mark the seed point as the starting point of the s-th sub-region, and write the seed point into the image sub-region point set; s is initially 1; the image sub-region point set is initially an empty set;

[0042] 5.12) Traverse all adjacent pixel points of the current seed point, and calculate the pixel value difference between each adjacent pixel point and the seed point; if the pixel value difference is greater than or equal to the image segmentation threshold, mark the adjacent pixel point as a region boundary point; if the pixel value difference is less than the image segmentation threshold, mark the adjacent pixel point as a new seed point, and write them into the seed point queue and the image sub-region point set respectively;

[0043] 5.13) Take the first point in the seed point queue as the current seed point, delete the current seed point from the seed point queue, and return to step 5.12) until the seed point queue is empty;

[0044] 5.14) Construct the s-th sub-region that contains the current image sub-region point set and has the smallest area, let s = s + 1, and return to step 5.11) until the image segmentation result consisting of k er sub-regions is obtained;

[0045] 5.15) Select a point from each sub-region as the calibration pixel point of the power equipment;

[0046] 6) Set an alarm threshold for each calibration pixel point of the power equipment, and update the top-down image infrared monitoring database and the side-view image infrared monitoring database to obtain:

[0047]

[0048]

[0049] where, PI umax is the alarm threshold of the u-th calibration pixel point of the power equipment; PI' omax is the alarm threshold of the o-th calibration pixel point of the power equipment;

[0050] 7) Input the infrared monitoring time period [T1, T2] at the control end; the control end selects the drone for taking images according to the infrared monitoring time period [T1, T2]; T1 and T2 are the start and end times of infrared monitoring;

[0051] If and then select the top-down drone I and the side-view drone I to take images, and enter step 8); P1 and P2 are the remaining battery powers of the top-down drone I and the side-view drone I respectively; P 01 、P 02$P_0$ is the power required for the overhead UAV I and the side-view UAV I to fly from the initial position to the starting point of the first flight route; $\Delta P_1$, $\Delta P_2$ are the power required for the overhead UAV I and the side-view UAV I to fly per unit distance; $H_1$, $H_2$ are the lengths of the first flight route and the second flight route; $\Delta t_1$ is the time taken for the overhead UAV I to traverse the first flight route; $\Delta t_2$ is the time taken for the overhead UAV II to traverse the second flight route;

[0052] If and then select the overhead UAV I, the overhead UAV II, and the side-view UAV I to take images, and enter step 9);

[0053] If and then select the overhead UAV I, the side-view UAV I, and the side-view UAV II to take images, and enter step 10);

[0054] If and then select the overhead UAV I, the overhead UAV II, the side-view UAV I, and the side-view UAV II to take images, and enter step 11);

[0055] 8) The control terminal controls the overhead UAV I to fly from the initial position to the starting point of the first flight route and fly repeatedly along the first flight route until the infrared monitoring end time $T_2$ is reached. The overhead UAV flies back to the initial position to charge, and this infrared monitoring ends;

[0056] The control terminal controls the side-view UAV I to fly from the initial position to the starting point of the second flight route and fly repeatedly along the second flight route until the infrared monitoring end time is reached. The side-view UAV flies back to the initial position, and this infrared monitoring ends;

[0057] During the flight, every time the overhead UAV I reaches a shooting position, it takes an overhead image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points;

[0058] During the flight, every time the side-view UAV I reaches a shooting position, it takes a side-view image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points;

[0059] Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the overhead image;

[0060] If the pixel values of all reference pixel points are less than or equal to the pixel values of the corresponding calibrated pixel points of the power equipment, the power equipment is operating normally at the current shooting position;

[0061] If there is a reference pixel point whose pixel value is greater than the pixel value of the calibrated pixel point of the power device and less than the alarm threshold, the top-down UAV I chip will transmit the power device code, the power device position code, and the cooling signal corresponding to the reference pixel point to the control terminal;

[0062] After receiving the cooling signal, the control terminal controls the cooling device to cool the area where the power device is located;

[0063] If there is a reference pixel point whose pixel value is greater than the alarm threshold of the calibrated pixel point of the power device, the top-down UAV I chip will transmit the power device code, the power device position code, and the alarm signal corresponding to the reference pixel point to the control terminal, and control the alarm device carried by the UAV to give an alarm;

[0064] The control terminal transmits the alarm signal to the user terminal;

[0065] 9) The control terminal sets the takeoff times of the top-down UAV I and the top-down UAV II based on the infrared monitoring time period [T1, T2];

[0066] The setting method is as follows: Calculate the takeoff times q1 of the top-down UAV, that is:

[0067]

[0068] In the formula, P'1 is the battery power of the top-down UAV in a fully charged state; represents rounding up;

[0069] Divide the infrared monitoring time period [T1, T2] into q1 time periods, and record the start time point of each time period, denoted as the start time point vector Take the start time point as the takeoff time of the top-down UAV I, and take the start time point as the takeoff time of the top-down UAV II;

[0070] The control terminal controls the top-down UAV I to fly from the initial position to the starting point of the first flight route at the start time point and fly repeatedly along the first flight route until the time difference between the current time and the element in the start time point that is closest is less than or equal to t 01 , and the top-down UAV I flies back to the initial position to charge; t 01 is the handover time of the top-down UAV I and the top-down UAV II;

[0071] The control terminal controls the top-down UAV II at the start time point Fly from the initial position to the starting point of the first flight route and repeat flying along the first flight route until the time difference between the current time and the time of the element closest to the starting time point QI1 is less than or equal to t 01 , looking down, the UAV II flies back to the initial position to charge;

[0072] The control terminal controls the side-looking UAV I to fly from the initial position to the starting point of the second flight route and repeat flying along the second flight route until the infrared monitoring end time is reached. The side-looking UAV flies back to the initial position to end this infrared monitoring;

[0073] During the flight, every time the overhead UAV I or the overhead UAV II reaches a shooting position, it shoots the overhead image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, which are recorded as reference pixel points;

[0074] During the flight, every time the side-looking UAV I reaches a shooting position, it shoots the side view image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, which are recorded as reference pixel points;

[0075] Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the overhead image;

[0076] If the pixel values of all reference pixel points are less than or equal to the pixel values of the calibrated pixel points of the corresponding power equipment, the power equipment at the current shooting position is operating normally;

[0077] If there are reference pixel points whose pixel values are greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip transmits the power equipment code, the power equipment location code, and the cooling signal corresponding to the reference pixel points to the control terminal;

[0078] After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located;

[0079] If there are reference pixel points whose pixel values are greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip transmits the power equipment code, the power equipment location code, and the alarm signal corresponding to the reference pixel points to the control terminal and controls the alarm device carried by the UAV to sound an alarm;

[0080] The control terminal transmits the alarm signal to the user terminal;

[0081] 10) The control terminal sets the takeoff times of the overhead UAV I and the overhead UAV II based on the infrared monitoring time period [T1, T2];

[0082] The setting method is: calculate the takeoff times q2 of the side-looking UAV, that is:

[0083]

[0084] Wherein, P'2 is the power of the side-view unmanned aerial vehicle in a fully charged state; Denotes rounding up;

[0085] Divide the infrared monitoring time period [T1, T2] into q2 time periods, and record the start time point of each time period, denoted as the start time point vector Take the start time point As the take-off time of the side-view unmanned aerial vehicle I, and take the start time point As the take-off time of the side-view unmanned aerial vehicle II;

[0086] The control terminal controls the side-view unmanned aerial vehicle I to fly from the initial position to the starting point of the second flight route at the start time point And fly repeatedly along the second flight route until the time difference from the time of the element closest to the start time point QII2 is less than or equal to t 02 , the side-view unmanned aerial vehicle I flies back to the initial position to charge; t 02 Is the handover time of the side-view unmanned aerial vehicle I and the side-view unmanned aerial vehicle II;

[0087] The control terminal controls the side-view unmanned aerial vehicle II to fly from the initial position to the starting point of the second flight route at the start time point And fly repeatedly along the second flight route until the time difference between the current time and the element closest to the start time point QI2 is less than or equal to t 02 , the side-view unmanned aerial vehicle II flies back to the initial position to charge;

[0088] The control terminal controls the overhead-view unmanned aerial vehicle I to fly from the initial position to the starting point of the first flight route and fly repeatedly along the first flight route until the infrared monitoring end time is reached, and the side-view unmanned aerial vehicle flies back to the initial position to end this infrared monitoring;

[0089] During the flight, every time the overhead-view unmanned aerial vehicle I reaches a shooting position, it shoots the overhead-view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as the reference pixel points;

[0090] During the flight, every time the side-view unmanned aerial vehicle I or the side-view unmanned aerial vehicle II reaches a shooting position, it shoots the side-view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as the reference pixel points;

[0091] Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the overhead-view image;

[0092] If the pixel values of all reference pixels are less than or equal to the pixel value of the calibrated pixel of the corresponding power device, the power device is operating normally at the current shooting position;

[0093] If there are reference pixels with pixel values greater than the pixel value of the calibrated pixel of the power device and less than the alarm threshold, the chip transmits the corresponding power device code, the power device location code, and the cooling signal of the reference pixel to the control terminal;

[0094] After receiving the cooling signal, the control terminal controls the cooling device to cool the area where the power device is located;

[0095] If there are reference pixels with pixel values greater than the alarm threshold of the calibrated pixel of the power device, the chip transmits the corresponding power device code, the power device location code, and the alarm signal to the control terminal, and controls the alarm device carried by the drone to issue an alarm;

[0096] The control terminal transmits the alarm signal to the user terminal;

[0097] 11) The control terminal sets the takeoff times of the overhead drone I, the overhead drone II, the side-view drone I, and the side-view drone II based on the infrared monitoring time period [T1, T2];

[0098] The control terminal controls the overhead drone I to fly from the initial position to the starting point of the first flight route at the starting time point QI1 = {Q 11 , Q 13 ,..., Q 1(2q1-1)}, and fly repeatedly along the first flight route until reaching the element in QII1 that is closest to the current time. Then the overhead drone I flies back to the initial position for charging;

[0099] The control terminal controls the overhead drone II to fly from the initial position to the starting point of the first flight route at the starting time point QII1 = {Q 12 , Q 14 ,..., Q 1(2q1)}, and fly repeatedly along the first flight route until the time difference between the current time and the element in QI1 that is closest to it is less than or equal to t 02 , and then the overhead drone II flies back to the initial position for charging;

[0100] The control terminal controls the side-view drone I to fly from the initial position to the starting point of the second flight route at the starting time point , and fly repeatedly along the second flight route until reaching the element in QII2 that is closest to the current time. Then the side-view drone I flies back to the initial position for charging;

[0101] The control terminal controls the side-view drone II to fly from the initial position to the starting point of the second flight route at the starting time point Fly from the initial position to the starting point of the second flight path and repeat flying along the second flight path until the time difference between the current time and the time of the element closest to the starting time point QI2 is less than or equal to t 02 The side-view drone II flies back to the initial position to charge;

[0102] During the flight, whenever the overhead drone I or the overhead drone II reaches a shooting position, it takes an overhead image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, which are recorded as reference pixel points;

[0103] During the flight, whenever the side-view drone I or the side-view drone II reaches a shooting position, it takes a side-view image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, which are recorded as reference pixel points;

[0104] Whenever the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the overhead image infrared monitoring database;

[0105] If the pixel values of all reference pixel points are less than or equal to the pixel values of the corresponding calibrated pixel points of the power equipment, the power equipment is operating normally at the current shooting position;

[0106] If there are reference pixel points with pixel values greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip transmits the encoding of the power equipment to which the reference pixel point belongs, the encoding of the position of the power equipment to which it belongs, and the cooling signal to the control terminal;

[0107] After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located;

[0108] If there are reference pixel points with pixel values greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip transmits the encoding of the power equipment to which the reference pixel point belongs, the encoding of the position of the power equipment to which it belongs, and the alarm signal to the control terminal, and controls the alarm device carried by the drone to sound an alarm;

[0109] The control terminal transmits the alarm signal to the user terminal.

[0110] Furthermore, the overhead drone carries at least one infrared camera, which is located below the fuselage of the overhead drone and is used to take an overhead infrared monitoring image of the power equipment;

[0111] The side-view drone carries at least two infrared cameras, which are located on both sides of the fuselage of the side-view drone and are used to take side-view infrared monitoring images of the power equipment.

[0112] Further, in the normal operation state, the infrared image D of each horizontal azimuth shooting area is D = [D1 D2... D n , and the infrared image F of each side azimuth shooting area is F = [F1 F2... F m , the side view image of the substation by the side view UAV, and the top view image of the substation taken by the top view UAV are all single-channel infrared images.

[0113] Further, in step 5.15), the method of selecting a point as the calibration pixel point of the power equipment from each sub-region is: randomly select the calibration pixel point of the power equipment.

[0114] Further, in step 5.15), the method of selecting a point as the calibration pixel point of the power equipment from each sub-region is: use the pixel point corresponding to the median of the pixel values as the calibration pixel point of the power equipment.

[0115] Further, in step 2), when determining the flight altitude a of the top view UAV and the horizontal azimuth shooting position matrix B1, and the side azimuth shooting position matrix B2 of the side view UAV, it is constrained by obtaining unobstructed side views and top views of all power equipment in the substation and minimizing the total number of images.

[0116] Further, when planning the first flight route and the second flight route, it is constrained by the shortest straight-line distance between adjacent shooting positions;

[0117] Among them, the first flight route starts and ends at the horizontal azimuth shooting position closest to the UAV charging pile;

[0118] The second flight route starts and ends at the side azimuth shooting position closest to the UAV charging pile.

[0119] Further, in step 5), the power equipment area in each infrared image is framed manually.

[0120] Further, when the control terminal receives an alarm signal, power off the power equipment higher than the alarm threshold.

[0121] Further, in step 2), the steps of determining the flight altitude a of the top view UAV and the horizontal azimuth shooting position matrix B1, and the side azimuth shooting position matrix B2 of the side view UAV include:

[0122] 2.1) Read the design drawings of the indoor substation, determine the spatial layout of the indoor substation, and construct the adjacency matrix A of the rooms in the indoor substation, that is:

[0123]

[0124] In the formula, Indicates the adjacency relationship between room g1 and room g2; Indicates that room g1 and room g2 are not adjacent; indicates that room g1 and room g2 = 1 indicates that room g1 and room g2 are adjacent; Indicates that room g1 and room g2 are not adjacent; g1 = 1, 2,..., G; g2 = 1, 2,..., G; g2 ≠ g1;

[0125] 2.2) Determine the flight height of the overhead drone based on the maximum height of the electrical equipment in room g1; the initial value of g1 is 1;

[0126] Based on the area U of room g1 g1 , and the single-shot area U of the overhead drone g0 , determine the number of shots of the overhead drone

[0127] Divide the cross-section of room g1 into U rectangular areas on average, and use the center point coordinates of each area as the horizontal and vertical coordinates of each shooting position of the overhead drone;

[0128] Based on the flight height of the overhead drone and the horizontal and vertical coordinates of each shooting position of the overhead drone, determine the horizontal azimuth shooting positions of the overhead drone in room g1;

[0129] 2.3) Divide room g1 vertically into multiple vertical areas according to the layout of the electrical equipment; for each vertical area, with the constraint of obtaining a panoramic image, determine the number of shots of the side-view drone, and then determine the side-azimuth shooting positions of the side-view drone;

[0130] 2.4) Determine whether g1 > G holds. If so, enter step 2.5). Otherwise, set g1 = g1 + 1 and return to step 2.2);

[0131] 2.5) Based on the adjacency matrix A of the indoor power distribution station room, with the constraint of the shortest distance, sort all the horizontal azimuth shooting positions and sort all the side-azimuth shooting positions, so as to determine the flight height a of the overhead drone and the horizontal azimuth shooting position matrix B1, and the side-azimuth shooting position matrix B2 of the side-view drone.

[0132] The technical effect of the present invention is beyond doubt. The present invention can intuitively display the temperature distribution through infrared images, quickly lock the problem equipment, and shorten the fault troubleshooting time. The present invention can achieve automatic and high-frequency full-station scanning, cover blind spots, and reduce the missed inspection rate. Brief Description of the Drawings

[0133] Figure 1 Is the method flow chart. Detailed Embodiment

[0134] The present invention will be further described below in conjunction with embodiments, but it should not be understood that the above-mentioned subject matter scope of the present invention is limited to the following embodiments. Without departing from the above technical idea of the present invention, various substitutions and changes made according to common general knowledge and conventional means in the art should be included within the protection scope of the present invention.

[0135] Embodiment 1:

[0136] Refer to Figure 1 , an infrared monitoring method for indoor substation equipment based on camera cruise, including the following steps:

[0137] 1) Obtain the spatial layout of the indoor substation and the distribution of power equipment in the substation;

[0138] 2) According to the spatial layout of the indoor substation and the distribution of power equipment in the substation, determine the horizontal azimuth shooting position matrix B1 of the overhead drone and the side azimuth shooting position matrix B2 of the side-view drone; each horizontal azimuth shooting position corresponds to a horizontal azimuth shooting area; each side azimuth shooting position corresponds to a side azimuth shooting area;

[0139] Among them, the horizontal azimuth shooting position matrix B1 and the side azimuth shooting position matrix B2 are as follows:

[0140]

[0141] In the formula, b n is the nth horizontal azimuth shooting position of the overhead drone; (b nx , b ny , b nz ) is the coordinate of the nth horizontal azimuth shooting position b n ; b nz is the flight altitude of the overhead drone at the nth horizontal azimuth shooting position; b' m is the mth side azimuth shooting position of the side-view drone; (b' mx , b' my , b' mz ) is the coordinate of the mth side azimuth shooting position b' m ; b' mz is the flight altitude of the side-view drone at the mth side azimuth shooting position;

[0142] 3) Based on the horizontal azimuth shooting position matrix B1 and the horizontal azimuth shooting position matrix B2, plan the flight routes of the overhead drone and the side-view drone during a single cruise, which are respectively recorded as the first flight route and the second flight route;

[0143] 4) The control terminal acquires the infrared images D = [D1 D2... D n of each horizontal azimuth shooting area of all power equipment in the indoor substation under normal operating conditions, and the infrared images F = [F1 F2... F m of each lateral azimuth shooting area;

[0144] 5) Frame and extract the power equipment area images in the infrared images of each horizontal azimuth shooting area and lateral azimuth shooting area, and select k er calibration pixel points of power equipment in each power equipment area image, record the coordinates, pixel values, power equipment codes, power equipment position codes, and shooting positions to which these calibration pixel points of power equipment belong, so as to construct an infrared monitoring database for top-view images and an infrared monitoring database for side-view images, write the infrared monitoring database for top-view images into the chip carried by the top-view UAV, and write the infrared monitoring database for side-view images into the chip carried by the side-view UAV;

[0145] The infrared monitoring database for top-view images is as follows:

[0146]

[0147] In the formula, E u is the u-th calibration pixel point of the power equipment; E ux , E uy is the coordinate of the u-th calibration pixel point of the power equipment; PI' u is the pixel value of the u-th calibration pixel point of the power equipment; PDS u is the power equipment code to which the u-th calibration pixel point of the power equipment belongs; PDS ucoord is the power equipment position code to which the u-th calibration pixel point of the power equipment belongs; POS u is the shooting position to which the u-th calibration pixel point of the power equipment belongs;

[0148] The infrared monitoring database for side-view images is as follows:

[0149]

[0150] In the formula, M o is the o-th calibration pixel point of the power equipment; M ox , M oy is the coordinate of the o-th calibration pixel point of the power equipment; PI” o is the pixel value of the u-th calibration pixel point of the power equipment; PDS' o is the power equipment code to which the o-th calibration pixel point of the power equipment belongs; PDS' ocoord is the power equipment position code to which the o-th calibration pixel point of the power equipment belongs; POS' oThe shooting position to which the pixel point calibrated for the o-th power device belongs;

[0151] The step of selecting k er calibrated pixel points of power devices in each power device area image includes:

[0152] 5.1) Extract the pixel values of each pixel point in the current power device area image, and write the pixel values and pixel coordinates into the set X to obtain:

[0153]

[0154] where X N is the N-th pixel point in the current power device area image; X Nx , X Ny is the coordinate of the pixel point X N ; PI N is the pixel value of the pixel point X N ;

[0155] 5.2) Construct the pixel value set PI = {PI1, PI1,..., PI N};

[0156] 5.3) Randomly select k er pixel values in the pixel value set PI as the clustering centers; the initial value of k er is 3; the subscript er represents the r-th power device area image in the e-th infrared image;

[0157] 5.4) Calculate the distance matrix DIS of each pixel value in the pixel value set PI to each clustering center, that is:

[0158]

[0159] where dis ls is the distance from the l-th pixel value to the s-th clustering center ;

[0160] 5.5) With the goal of the minimum distance to the clustering center, assign each pixel value in the pixel value set PI to the cluster where the nearest clustering center is located;

[0161] 5.6) Calculate the mean value of each clustering cluster, and use the mean value as the new clustering center, return to step 5.4), until the clustering center remains unchanged or reaches the maximum number of iterations, and enter step 5.7)

[0162] 5.7) Calculate the clustering cost L er corresponding to the k k clustering centers, that is:

[0163]

[0164] In the formula, C s is the s-th clustering cluster; μ s is the clustering center of the s-th clustering cluster; c is the element of the s-th clustering cluster;

[0165] 5.8) Judge Whether it holds. If so, go to step 5.9). Otherwise, let k er = k er + 1, and return to step 5.3);

[0166] 5.9) Select the number of clustering centers with the minimum difference cost as the optimal number of clustering centers k er , and use the clustering centers of each clustering cluster corresponding to the optimal number of clustering centers k er as the optimal clustering centers;

[0167] 5.10) Based on the regional division objective function (9), select the image segmentation threshold ε, that is:

[0168] ε = min||PI s1 - PI s2 ||(9)

[0169] In the formula, PI s1 , PI s2 are the s1-th and s2-th optimal clustering centers; s1 = 1, 2,..., k er ; s2 = 1, 2,..., k er ; s2 ≠ s1;

[0170] 5.11) Select the first pixel point in the power equipment area image that has not been divided into any area as the seed point, mark this seed point as the starting point of the s-th sub-area, and write this seed point into the image sub-area point set; the initial value of s is 1; the image sub-area point set is an empty set in the initial state;

[0171] 5.12) Traverse all adjacent pixel points of the current seed point, and calculate the pixel value difference between each adjacent pixel point and the seed point; if the pixel value difference is greater than or equal to the image segmentation threshold, mark this adjacent pixel point as the area boundary point; if the pixel value difference is less than the image segmentation threshold, mark this adjacent pixel point as a new seed point, and write it into the seed point queue and the image sub-area point set respectively;

[0172] 5.13) Use the first point in the seed point queue as the current seed point, delete the current seed point from the seed point queue, and return to step 5.12) until the seed point queue is empty;

[0173] 5.14) Construct the s-th sub-region that contains the point set of the current image sub-region and has the smallest area, let s = s + 1, and return to step 5.11), until the image segmentation result consisting of k er sub-regions is obtained;

[0174] 5.15) Select a point from each sub-region as the calibration pixel point of the power equipment;

[0175] 6) Set the alarm threshold for each calibration pixel point of the power equipment, and update the infrared monitoring databases of the top-view image and the side-view image to obtain:

[0176]

[0177]

[0178] In the formula, PI umax is the alarm threshold of the u-th calibration pixel point of the power equipment; PI' omax is the alarm threshold of the o-th calibration pixel point of the power equipment;

[0179] 7) Input the infrared monitoring time period [T1, T2] at the control end; the control end selects the UAVs used to capture images according to the infrared monitoring time period [T1, T2]; T1 and T2 are the start and end times of the infrared monitoring;

[0180] If and then select the top-view UAV I and the side-view UAV I to capture images, and enter step 8); P1 and P2 are the remaining battery powers of the top-view UAV I and the side-view UAV I; P 01 and P 02 are the battery powers required for the top-view UAV I and the side-view UAV I to fly from the initial position to the starting point of the first flight route; ΔP1 and ΔP2 are the battery powers required for the top-view UAV I and the side-view UAV I to fly per unit distance; H1 and H2 are the lengths of the first flight route and the second flight route; Δt1 is the duration for the top-view UAV I to traverse the first flight route; Δt2 is the duration for the top-view UAV II to traverse the second flight route;

[0181] If and then select the top-view UAV I, the top-view UAV II, and the side-view UAV I to capture images, and enter step 9);

[0182] If and then select the top-view UAV I, the side-view UAV I, and the side-view UAV II to capture images, and enter step 10);

[0183] If and Then, select the top-down UAV I, top-down UAV II, side-view UAV I, and side-view UAV II to capture images, and enter step 11);

[0184] 8) The control terminal controls the top-down UAV I to fly from the initial position to the starting point of the first flight route and fly repeatedly along the first flight route until the infrared monitoring end time T2 is reached. The top-down UAV flies back to the initial position to charge, ending this infrared monitoring;

[0185] The control terminal controls the side-view UAV I to fly from the initial position to the starting point of the second flight route and fly repeatedly along the second flight route until the infrared monitoring end time is reached. The side-view UAV flies back to the initial position, ending this infrared monitoring;

[0186] During the flight, every time the top-down UAV I reaches a shooting position, it captures the top-down image of the substation and retains the pixel points with the same pixel point coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points;

[0187] During the flight, every time the side-view UAV I reaches a shooting position, it captures the side-view image of the substation and retains the pixel points with the same pixel point coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points;

[0188] Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the top-down image infrared monitoring database;

[0189] If the pixel values of all reference pixel points are less than or equal to the pixel values of the calibrated pixel points of the corresponding power equipment, the power equipment is operating normally at the current shooting position;

[0190] If there are reference pixel points with pixel values greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip of the top-down UAV I transmits the encoding of the power equipment to which the reference pixel point belongs, the location encoding of the power equipment to which it belongs, and the cooling signal to the control terminal;

[0191] After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located;

[0192] If there are reference pixel points with pixel values greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip of the top-down UAV I transmits the encoding of the power equipment to which the reference pixel point belongs, the location encoding of the power equipment to which it belongs, and the alarm signal to the control terminal, and controls the alarm device carried by the UAV to issue an alarm;

[0193] The control terminal transmits the alarm signal to the user terminal;

[0194] 9) The control terminal sets the take-off times of the overhead UAV I and the overhead UAV II based on the infrared monitoring time period [T1, T2];

[0195] The setting method is as follows: Calculate the take-off times q1 of the overhead UAVs, that is:

[0196]

[0197] In the formula, P'1 is the power of the overhead UAV in a fully charged state; represents rounding up;

[0198] Divide the infrared monitoring time period [T1, T2] into q1 time periods, and record the starting time point of each time period, denoted as the starting time point vector Take the starting time point as the take-off time of the overhead UAV I, and take the starting time point as the take-off time of the overhead UAV II;

[0199] The control terminal controls the overhead UAV I to fly from the initial position to the starting point of the first flight route at the starting time point and fly repeatedly along the first flight route until the time difference between the current time and the element in the starting time point that is closest is less than or equal to t 01 , and the overhead UAV I flies back to the initial position for charging; t 01 is the handover time between the overhead UAV I and the overhead UAV II;

[0200] The control terminal controls the overhead UAV II to fly from the initial position to the starting point of the first flight route at the starting time point and fly repeatedly along the first flight route until the time difference between the current time and the element in the starting time point QI1 that is closest is less than or equal to t 01 , and the overhead UAV II flies back to the initial position for charging;

[0201] The control terminal controls the side-view UAV I to fly from the initial position to the starting point of the second flight route and fly repeatedly along the second flight route until the infrared monitoring end time is reached, and the side-view UAV flies back to the initial position to end this infrared monitoring;

[0202] During the flight, every time the overhead UAV I or the overhead UAV II reaches a shooting position, it takes a top-down view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as the reference pixel points;

[0203] During the flight, every time the side-view UAV I reaches a shooting position, it takes a side-view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as the reference pixel points;

[0204] Each time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment at the same coordinates in the infrared monitoring database of the top-down image.

[0205] If the pixel values of all reference pixel points are less than or equal to the pixel values of the calibrated pixel points of the corresponding power equipment, the power equipment is operating normally at the current shooting position.

[0206] If there are reference pixel points whose pixel values are greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip transmits the encoding of the power equipment to which the reference pixel point belongs, the encoding of the position of the power equipment to which it belongs, and the cooling signal to the control terminal.

[0207] After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located.

[0208] If there are reference pixel points whose pixel values are greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip transmits the encoding of the power equipment to which the reference pixel point belongs, the encoding of the position of the power equipment to which it belongs, and the alarm signal to the control terminal, and controls the alarm device carried by the drone to issue an alarm.

[0209] The control terminal transmits the alarm signal to the user terminal.

[0210] 10) The control terminal sets the takeoff times of the top-down drones I and II based on the infrared monitoring time period [T1, T2].

[0211] The setting method is as follows: Calculate the number of takeoff times q2 of the side-looking drone, that is:

[0212]

[0213] In the formula, P'2 is the battery power of the side-looking drone in a fully charged state. represents rounding up;

[0214] Divide the infrared monitoring time period [T1, T2] into q2 time periods, and record the start time point of each time period, denoted as the start time point vector Take the start time point as the takeoff time of the side-looking drone I, and take the start time point as the takeoff time of the side-looking drone II;

[0215] The control terminal controls the side-looking drone I to fly from the initial position to the starting point of the second flight route at the start time point and fly repeatedly along the second flight route until the time difference from the starting time point to the closest element in QII2 is less than or equal to t 02, the side-view drone I flies back to the initial position for charging; t 02 is the handover time between the side-view drone I and the side-view drone II to prevent the drones from colliding.

[0216] The control terminal controls the side-view drone II at the starting time point to fly from the initial position to the starting point of the second flight route and repeat flying along the second flight route until the time difference between the current time and the time of the element closest to the starting time point QI2 is less than or equal to t 02 , the side-view drone II flies back to the initial position for charging;

[0217] The control terminal controls the overhead drone I to fly from the initial position to the starting point of the first flight route and repeat flying along the first flight route until the infrared monitoring end time is reached. The side-view drones fly back to the initial position, ending this infrared monitoring;

[0218] During the flight, every time the overhead drone I reaches a shooting position, it shoots an overhead image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points;

[0219] During the flight, every time the side-view drone I or the side-view drone II reaches a shooting position, it shoots a side-view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points;

[0220] Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the overhead image;

[0221] If the pixel values of all reference pixel points are less than or equal to the pixel values of the corresponding calibrated pixel points of the power equipment, the power equipment is operating normally at the current shooting position;

[0222] If there are reference pixel points whose pixel values are greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip transmits the encoding of the power equipment to which the reference pixel points belong, the encoding of the position of the power equipment to which the reference pixel points belong, and the cooling signal to the control terminal;

[0223] After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located;

[0224] If there are reference pixel points whose pixel values are greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip transmits the encoding of the power equipment to which the reference pixel points belong, the encoding of the position of the power equipment to which the reference pixel points belong, and the alarm signal to the control terminal, and controls the alarm device carried by the drone to give an alarm;

[0225] The control terminal transmits the alarm signal to the user terminal;

[0226] 11) The control terminal sets the take-off times of the overhead UAV I, overhead UAV II, side-view UAV I, and side-view UAV II based on the infrared monitoring time period [T1, T2].

[0227] The control terminal controls the overhead UAV I to fly from the initial position to the starting point of the first flight route at the starting time point and fly repeatedly along the first flight route until reaching the element in the starting time point QII1 that is closest, and then the overhead UAV I flies back to the initial position to charge.

[0228] The control terminal controls the overhead UAV II to fly from the initial position to the starting point of the first flight route at the starting time point and fly repeatedly along the first flight route until the time difference between the current time and the time of the element in the starting time point QI1 that is closest is less than or equal to t 02 , and then the overhead UAV II flies back to the initial position to charge.

[0229] The control terminal controls the side-view UAV I to fly from the initial position to the starting point of the second flight route at the starting time point and fly repeatedly along the second flight route until reaching the element in the starting time point QII2 that is closest, and then the side-view UAV I flies back to the initial position to charge.

[0230] The control terminal controls the side-view UAV II to fly from the initial position to the starting point of the second flight route at the starting time point and fly repeatedly along the second flight route until the time difference between the current time and the time of the element in the starting time point QI2 that is closest is less than or equal to t 02 , and then the side-view UAV II flies back to the initial position to charge.

[0231] During the flight, every time the overhead UAV I or the overhead UAV II reaches a shooting position, it takes an overhead view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, which are recorded as reference pixel points.

[0232] During the flight, every time the side-view UAV I or the side-view UAV II reaches a shooting position, it takes a side-view image of the substation and retains the pixel points with the same coordinates as the calibrated pixel points of the power equipment at the current shooting position, which are recorded as reference pixel points.

[0233] Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the overhead view image.

[0234] If the pixel values of all the reference pixel points are less than or equal to the pixel values of the corresponding calibrated pixel points of the power equipment, the power equipment at the current shooting position is operating normally.

[0235] If there are reference pixel points whose pixel values are greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip will transmit the power equipment code, the power equipment location code, and the cooling signal corresponding to the reference pixel points to the control end;

[0236] After receiving the cooling signal, the control end controls the cooling equipment to cool the area where the power equipment is located;

[0237] If there are reference pixel points whose pixel values are greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip will transmit the power equipment code, the power equipment location code, and the alarm signal corresponding to the reference pixel points to the control end, and control the alarm device carried by the drone to give an alarm;

[0238] The control end transmits the alarm signal to the user end.

[0239] Embodiment 2:

[0240] An infrared monitoring method for indoor substation equipment based on camera cruising, the technical content is the same as that of any one of Embodiment 1. Further, the overhead drone carries at least one infrared camera, and the infrared camera is located below the fuselage of the overhead drone and is used to capture the overhead infrared monitoring image of the power equipment;

[0241] The side-view drone carries at least two infrared cameras, and the two infrared cameras are located on both sides of the fuselage of the side-view drone and are used to capture the side-view infrared monitoring image of the power equipment.

[0242] Embodiment 3:

[0243] An infrared monitoring method for indoor substation equipment based on camera cruising, the technical content is the same as that of any one of Embodiments 1-2. Further, in the normal operation state, the infrared images D = [D1 D2...D n of each horizontal azimuth shooting area, and the infrared images F = [F1 F2...F m of each side azimuth shooting area, the side-view image of the substation by the side-view drone, and the overhead image of the substation captured by the overhead drone are all single-channel infrared images.

[0244] Embodiment 4:

[0245] An infrared monitoring method for indoor substation equipment based on camera cruising, the technical content is the same as that of any one of Embodiments 1-3. Further, in step 5.15), the method of selecting a point from each sub-region as the calibrated pixel point of the power equipment is: randomly select the calibrated pixel point of the power equipment.

[0246] Embodiment 5:

[0247] An infrared monitoring method for indoor substation equipment based on camera cruise, the technical content is the same as any one of Embodiments 1-4. Further, in step 5.15), the method of selecting a point from each sub-region as the calibration pixel point of the power equipment is: using the pixel point corresponding to the median of the pixel values as the calibration pixel point of the power equipment.

[0248] Embodiment 6:

[0249] An infrared monitoring method for indoor substation equipment based on camera cruise, the technical content is the same as any one of Embodiments 1-5. Further, in step 2), when determining the flight height a of the overhead drone and the horizontal azimuth shooting position matrix B1, and the side azimuth shooting position matrix B2 of the side-view drone, it is constrained by obtaining unobstructed side views, top views of all power equipment in the substation, and the minimum total number of images.

[0250] Embodiment 7:

[0251] An infrared monitoring method for indoor substation equipment based on camera cruise, the technical content is the same as any one of Embodiments 1-6. Further, when planning the first flight route and the second flight route, it is constrained by the shortest straight-line distance between adjacent two shooting positions;

[0252] Among them, the first flight route starts and ends at the horizontal azimuth shooting position closest to the drone charging pile;

[0253] The second flight route starts and ends at the side azimuth shooting position closest to the drone charging pile.

[0254] Embodiment 8:

[0255] An infrared monitoring method for indoor substation equipment based on camera cruise, the technical content is the same as any one of Embodiments 1-7. Further, in step 5), the power equipment area in each infrared image is framed manually.

[0256] Embodiment 9:

[0257] An infrared monitoring method for indoor substation equipment based on camera cruise, the technical content is the same as any one of Embodiments 1-8. Further, when the control terminal receives an alarm signal, power-off processing is performed on the power equipment higher than the alarm threshold.

[0258] Embodiment 10:

[0259] An infrared monitoring method for indoor substation equipment based on camera cruise, the technical content is the same as any one of Embodiments 1-9. Further, the steps of determining the flight height a of the overhead drone and the horizontal azimuth shooting position matrix B1, and the side azimuth shooting position matrix B2 of the side-view drone in step 2) include:

[0260] 2.1) Read the design drawings of the indoor distribution substation, determine the spatial layout of the indoor distribution substation, and construct the adjacency matrix A of the rooms in the indoor distribution substation, that is:

[0261]

[0262] In the formula, represents the adjacency relationship between room g1 and room g2; represents that room g1 and room g2 are not adjacent; represents that room g1 and room g2 = 1 means room g1 and room g2 are adjacent; represents that room g1 and room g2 are not adjacent; g1 = 1, 2,..., G; g2 = 1, 2,..., G; g2 ≠ g1;

[0263] 2.2) Based on the maximum height of the electrical equipment in room g1, determine the flight height of the overhead drone; the initial value of g1 is 1;

[0264] Based on the area U of room g1 g1 , and the single-shot area U of the overhead drone g0 , determine the number of shots of the overhead drone

[0265] Divide the cross-section of room g1 into U rectangular areas on average, and use the center point coordinates of each area as the horizontal and vertical coordinates of each shooting position of the overhead drone;

[0266] Based on the flight height of the overhead drone and the horizontal and vertical coordinates of each shooting position of the overhead drone, determine the horizontal azimuth shooting positions in room g1;

[0267] 2.3) According to the layout of the electrical equipment, divide room g1 vertically into multiple vertical areas; for each vertical area, with the constraint of obtaining a panoramic image, determine the number of shots of the side-view drone, and then determine the side-azimuth shooting positions of the side-view drone;

[0268] 2.4) Determine whether g1 > G holds. If so, enter step 2.5). Otherwise, let g1 = g1 + 1 and return to step 2.2);

[0269] 2.5) Based on the adjacency matrix A of the rooms in the indoor distribution substation, with the constraint of the shortest distance, sort all the horizontal azimuth shooting positions and sort all the side-azimuth shooting positions, so as to determine the flight height a of the overhead drone and the horizontal azimuth shooting position matrix B1, as well as the side-azimuth shooting position matrix B2 of the side-view drone.

[0270] Example 11:

[0271] An infrared monitoring method for indoor substation equipment based on camera cruise, with the technical content being the same as any one of Embodiments 1-10. When shooting top-down images and side-view images, the top-down images and side-view images of adjacent areas can overlap.

[0272] Since the shooting times of adjacent area images are different, the top-down images and side-view images are not cropped, and the independent top-down images and side-view images are used as the input to the chip.

[0273] Embodiment 12:

[0274] An infrared monitoring method for indoor substation equipment based on camera cruise, with the technical content being the same as any one of Embodiments 1-11. Every time the drone takes a picture, it is input into the chip for infrared temperature judgment.

[0275] Embodiment 13:

[0276] An infrared monitoring method for indoor substation equipment based on camera cruise, with the technical content being the same as any one of Embodiments 1-12. The location of the power equipment at least includes the room number of the indoor substation.

[0277] Embodiment 14:

[0278] An infrared monitoring method for indoor substation equipment based on camera cruise, with the technical content being the same as any one of Embodiments 1-13. After the cooling equipment is started, if the pixel value of the power equipment returns to normal when the next cruise reaches the area where the power equipment to be cooled is located, the cooling equipment is turned off.

[0279] Embodiment 15:

[0280] An infrared monitoring method for indoor substation equipment based on camera cruise, with the technical content being the same as any one of Embodiments 1-14. Q 11 = Q 21 = T1. In the starting time point vector Among them, In addition, the difference between other adjacent starting time points is equal to

[0281] In the starting time point vector Among them, In addition, the difference between other adjacent starting time points is equal to

[0282] After the drone takes off at the starting time point and reaches the termination time T2, the drone directly flies back to the initial position for charging.

[0283] After the drone takes off at the starting time point and reaches the termination time T2, the drone directly flies back to the initial position for charging.

Claims

1. An infrared monitoring method for indoor substation equipment based on camera cruising, characterized in that, Including the following steps: 1) Obtain the spatial layout of the indoor distribution substation and the distribution of power equipment within the substation; 2) Determine the horizontal azimuth shooting position matrix B1 of the overhead drone and the side azimuth shooting position matrix B2 of the side-view drone according to the spatial layout of the indoor distribution substation and the distribution of power equipment within the substation; Each horizontal azimuth shooting position corresponds to a horizontal azimuth shooting area; Each side azimuth shooting position corresponds to a side azimuth shooting area; Among them, the horizontal azimuth shooting position matrix B1 and the side azimuth shooting position matrix B2 are as follows: Where, b n is the nth horizontal azimuth shooting position of the overhead drone; (b nx , b ny , b nz ) are the coordinates of the nth horizontal azimuth shooting position b n ; b nz is the flight altitude of the overhead drone at the nth horizontal azimuth shooting position; b' m is the mth side azimuth shooting position of the side-view drone; (b' mx , b' my , b' mz ) are the coordinates of the mth side azimuth shooting position b' m ; b' mz is the flight altitude of the side-view drone at the mth side azimuth shooting position; 3) Based on the horizontal azimuth shooting position matrix B1 and the horizontal azimuth shooting position matrix B2, plan the flight routes of the overhead drone and the side-view drone during a single cruise, denoted as the first flight route and the second flight route respectively; 4) The control terminal acquires the infrared images D = [D1 D2... D n of each horizontally oriented shooting area of all power equipment in the indoor substation under normal operating conditions, and the infrared images F = [F1 F2... F m of each laterally oriented shooting area; 5) Select and extract the power equipment area images in the infrared images of each horizontally photographed area and laterally photographed area, and select k er calibrated pixel points of power equipment in each power equipment area image, record the coordinates, pixel values, power equipment code numbers, power equipment location code numbers, and the location of the photographed area to which these calibrated pixel points of power equipment belong, so as to construct an infrared monitoring database for the top-view image and an infrared monitoring database for the side-view image, write the infrared monitoring database for the top-view image into the chip carried by the top-view UAV, and write the infrared monitoring database for the side-view image into the chip carried by the side-view UAV; The overhead image infrared monitoring database is as follows: where E u is the calibrated pixel point of the u-th power device; E ux , E uy are the coordinates of the calibrated pixel point of the u-th power device; PI u ' is the pixel value of the calibrated pixel point of the u-th power device; PDS u is the power device code to which the calibrated pixel point of the u-th power device belongs; PDS ucoord is the power device location code to which the calibrated pixel point of the u-th power device belongs; POS u is the shooting location to which the calibrated pixel point of the u-th power device belongs; The side-view image infrared monitoring database is as follows: Where, M o is the calibration pixel point of the o-th power device; M ox , M oy are the coordinates of the calibration pixel point of the o-th power device; PI' o ' is the pixel value of the calibration pixel point of the u-th power device; PDS' o is the power device code to which the calibration pixel point of the o-th power device belongs; PDS' ocoord is the power device position code to which the calibration pixel point of the o-th power device belongs; POS' o is the shooting position to which the calibration pixel point of the o-th power device belongs; Select k calibration pixel points of power equipment in each image of the power equipment area er The steps are as follows: 5.1) Extract the pixel values of each pixel point in the current power equipment area image, and write the pixel values and pixel coordinates into the set X to obtain: Wherein, X N is the Nth pixel point in the current power equipment area image; X Nx , X Ny are the coordinates of the pixel point X N ; PI N is the pixel value of the pixel point X N . 5.2) Construct a set of pixel values PI = {PI1, PI1,..., PI N}; 5.3) Randomly select k pixel values from the set PI of pixel values er as the clustering centers; k er The initial value is 3; the subscript er represents the r-th power equipment area image in the e-th infrared image; 5.4) Calculate the distance matrix DIS of each pixel value in the pixel value set PI to each cluster center, that is: where dis ls is the distance from the l-th pixel value to the s-th cluster center ; 5.5) With the goal of minimizing the distance to the cluster center, assign each pixel value in the pixel value set PI to the cluster where the nearest cluster center is located; 5.6) Calculate the mean value of each cluster, and use the mean value as the new cluster center, return to step 5.4), until the cluster center remains unchanged or reaches the maximum number of iterations, and enter step 5.7) 5.7) Calculate k er The clustering cost L corresponding to the cluster centers k , that is: where C s is the s-th clustering cluster; μ s is the clustering center of the s-th clustering cluster; c is an element of the s-th clustering cluster; 5.8) Judgment is established. If so, proceed to step 5.9). Otherwise, let k er = k er + 1, and return to step 5.3); 5.9) Select the number of clustering centers with the smallest difference cost as the optimal number of clustering centers k er , and use the clustering centers of each clustering cluster corresponding to the optimal number of clustering centers k er as the optimal clustering centers; 5.10) Based on the region division objective function (9), select the image segmentation threshold ε, that is: ε=min||PI s1 -PI s2 ||(9) wherein, PI s1 , PI s2 are the s1-th and s2-th optimal clustering centers; s1 = 1, 2,..., k er ; s2 = 1, 2,..., k er ; s2 ≠ s1; 5.11) Select the first pixel point in the power equipment area image that has not been divided into any region as the seed point, mark the seed point as the starting point of the s-th sub-region, and write the seed point into the image sub-region point set; the initial value of s is 1; the image sub-region point set is an empty set in the initial state; 5.12) Traverse all adjacent pixel points of the current seed point, and calculate the pixel value difference between each adjacent pixel point and the seed point; if the pixel value difference is greater than or equal to the image segmentation threshold, mark the adjacent pixel point as a region boundary point; if the pixel value difference is less than the image segmentation threshold, mark the adjacent pixel point as a new seed point, and write them into the seed point queue and the image sub-region point set respectively; 5.13) Take the first point in the seed point queue as the current seed point, delete the current seed point from the seed point queue, and return to step 5.12), until the seed point queue is empty; 5.14) Construct the s-th sub-region that contains the point set of the current image sub-region and has the smallest area, let s = s + 1, and return to step 5.11) until the image segmentation result consisting of k er sub-regions is obtained; 5.15) Select a point from each sub-region as the power equipment calibration pixel point; 6) Set an alarm threshold for each power equipment calibration pixel point, and update the overhead image infrared monitoring database and the side-view image infrared monitoring database to obtain: where, PI umax is the alarm threshold of the calibration pixel point of the u-th power device; PI' omax is the alarm threshold of the calibration pixel point of the o-th power device; 7) Input the infrared monitoring time period [T1, T2] at the control end; the control end selects the drone for taking pictures according to the infrared monitoring time period [T1, T2]; T1 and T2 are the start and end times of infrared monitoring; If and then select the images taken by the top-view UAV I and the side-view UAV I, and enter step 8); P1 and P2 are the current remaining battery levels of the top-view UAV I and the side-view UAV I; P 01 , P 02 are the battery levels required for the top-view UAV I and the side-view UAV I to fly from the initial position to the starting point of the first flight route; ΔP1 and ΔP2 are the battery levels required for the top-view UAV I and the side-view UAV I to fly per unit distance; H1 and H2 are the lengths of the first flight route and the second flight route; Δt1 is the duration for the top-view UAV I to traverse the first flight route; Δt2 is the duration for the top-view UAV II to traverse the second flight route; If and then select the images taken by the top-view UAV I, the top-view UAV II, and the side-view UAV I, and proceed to step 9); If and then select the images taken by the top-view UAV I, the side-view UAV I, and the side-view UAV II, and enter step 10); If and then select the images taken by the top-view UAV I, top-view UAV II, side-view UAV I, and side-view UAV II, and enter step 11); 8) The control terminal controls the overhead drone I to fly from the initial position to the starting point of the first flight route and repeat the flight along the first flight route until the infrared monitoring termination time T2 is reached. Then, the overhead drone flies back to the initial position to charge, ending this infrared monitoring; The control terminal controls the side-view drone I to fly from the initial position to the starting point of the second flight route and repeat the flight along the second flight route until the infrared monitoring termination time is reached. Then, the side-view drone flies back to the initial position, ending this infrared monitoring; During the flight, every time the overhead drone I reaches a shooting position, it takes a top-down image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points; During the flight, every time the side-view drone I reaches a shooting position, it takes a side-view image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points; Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the top-down image; If the pixel values of all reference pixel points are less than or equal to the pixel values of the calibrated pixel points of the corresponding power equipment, the power equipment at the current shooting position is operating normally; If there are reference pixel points with pixel values greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip of the overhead drone I transmits the encoding of the power equipment to which the reference pixel point belongs, the encoding of the location of the power equipment to which it belongs, and the cooling signal to the control terminal; After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located; If there are reference pixel points with pixel values greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip of the overhead drone I transmits the encoding of the power equipment to which the reference pixel point belongs, the encoding of the location of the power equipment to which it belongs, and the alarm signal to the control terminal, and controls the alarm device carried by the drone to sound an alarm; The control terminal transmits the alarm signal to the user terminal; 9) The control terminal sets the takeoff times of the overhead drone I and the overhead drone II based on the infrared monitoring time period [T1, T2]; The setting method is: calculate the takeoff times q1 of the overhead drone, that is: Wherein, P'1 is the power of the drone in a fully charged state when viewed from above; represents rounding up; Divide the infrared monitoring time period [T1, T2] into q1 time periods, and record the starting time point of each time period, denoted as the starting time point vector Take the starting time point as the takeoff time of the overhead drone I, and take the starting time point as the takeoff time of the overhead drone II; The control terminal controls the overhead drone I at the starting time point to fly from the initial position to the starting point of the first flight route and repeatedly fly along the first flight route until the time difference between the current time and the time of the element closest to the starting time point is less than or equal to t 01 , and the overhead drone I flies back to the initial position to charge; t 01 is the handover time between the overhead drone I and the overhead drone II; The control terminal controls the overhead UAV II at the starting time point to fly from the initial position to the starting point of the first flight route and repeatedly fly along the first flight route until the time difference between the current time and the time of the element closest to the starting time point QI1 is less than or equal to t 01 , and the overhead UAV II flies back to the initial position for charging; The control terminal controls the side-view drone I to fly from the initial position to the starting point of the second flight route and repeat the flight along the second flight route until the infrared monitoring termination time is reached. Then, the side-view drone flies back to the initial position, ending this infrared monitoring; During the flight, every time the overhead drone I or the overhead drone II reaches a shooting position, it takes a top-down image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points; During the flight, every time the side-view drone I reaches a shooting position, it takes a side-view image of the substation and retains the pixel points with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixel points; Every time the chip obtains a set of reference pixel points, it compares the reference pixel points with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment with the same coordinates in the infrared monitoring database of the top-down image; If the pixel values of all reference pixels are less than or equal to the pixel values of the calibrated pixels of the corresponding power equipment, the power equipment is operating normally at the current shooting position; If there are reference pixels whose pixel values are greater than the pixel values of the calibrated pixels of the power equipment and less than the alarm threshold, the chip transmits the power equipment code, the power equipment location code, and the cooling signal corresponding to the reference pixel to the control terminal; After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located; If there are reference pixels whose pixel values are greater than the alarm threshold of the calibrated pixels of the power equipment, the chip transmits the power equipment code, the power equipment location code, and the alarm signal corresponding to the reference pixel to the control terminal, and controls the alarm device carried by the drone to issue an alarm; The control terminal transmits the alarm signal to the user terminal; 10) The control terminal sets the takeoff times of the overhead drones I and II based on the infrared monitoring time period [T1, T2]; The setting method is: calculate the takeoff times q2 of the side-view drone, that is: Wherein, P'2 is the power of the side-view UAV in a fully charged state; represents rounding up; Divide the infrared monitoring time period [T1, T2] into q2 time periods, and record the start time point of each time period, denoted as the start time point vector Take the start time point as the takeoff time of the side-looking drone I, and take the start time point as the takeoff time of the side-looking drone II; The control end controls the side-view UAV I at the starting time point to fly from the initial position to the starting point of the second flight route and repeat flying along the second flight route until the time difference of the element closest to the starting time point QII2 is less than or equal to t 02 , and the side-view UAV I flies back to the initial position for charging; t 02 is the handover time between the side-view UAV I and the side-view UAV II; The control end controls the side-looking UAV II at the starting time point to fly from the initial position to the starting point of the second flight route and repeatedly fly along the second flight route until the time difference between the current time and the time of the element closest to the starting time point QI2 is less than or equal to t 02 , and the side-looking UAV II flies back to the initial position to charge; The control terminal controls the overhead drone I to fly from the initial position to the starting point of the first flight route and fly repeatedly along the first flight route until the infrared monitoring end time is reached, and the side-view drone flies back to the initial position to end this infrared monitoring; During the flight, every time the overhead drone I reaches a shooting position, it takes a top-down image of the substation and retains the pixels with the same coordinates as the calibrated pixels of the power equipment at the current shooting position, which are recorded as reference pixels; During the flight, every time the side-view drone I or the side-view drone II reaches a shooting position, it takes a side-view image of the substation and retains the pixels with the same coordinates as the calibrated pixels of the power equipment at the current shooting position, which are recorded as reference pixels; Every time the chip obtains a set of reference pixels, it compares the reference pixels with the pixel values and alarm thresholds of the calibrated pixels of the power equipment at the same coordinates in the infrared monitoring database of the top-down image; If the pixel values of all reference pixels are less than or equal to the pixel values of the calibrated pixels of the corresponding power equipment, the power equipment is operating normally at the current shooting position; If there are reference pixels whose pixel values are greater than the pixel values of the calibrated pixels of the power equipment and less than the alarm threshold, the chip transmits the power equipment code, the power equipment location code, and the cooling signal corresponding to the reference pixel to the control terminal; After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located; If there are reference pixels whose pixel values are greater than the alarm threshold of the calibrated pixels of the power equipment, the chip transmits the power equipment code, the power equipment location code, and the alarm signal corresponding to the reference pixel to the control terminal, and controls the alarm device carried by the drone to issue an alarm; The control terminal transmits the alarm signal to the user terminal; 11) The control terminal sets the takeoff times of the overhead drones I and II, the side-view drones I and II based on the infrared monitoring time period [T1, T2]; The control terminal controls the overhead UAV I at the starting time point to fly from the initial position to the starting point of the first flight route and repeatedly fly along the first flight route until reaching the closest element in the starting time point QII1, and then the overhead UAV I flies back to the initial position for charging; The control terminal controls the overhead UAV II at the starting time point to fly from the initial position to the starting point of the first flight route and repeatedly fly along the first flight route until the time difference between the current time and the time of the element closest to the starting time point QI1 is less than or equal to t 02 , and the overhead UAV II flies back to the initial position to charge; The control terminal controls the side-view UAV I to fly from the starting position to the starting point of the second flight route at the starting time point and fly along the second flight route repeatedly until reaching the element closest to the starting time point QII2, then the side-view UAV I flies back to the initial position for charging; The control terminal controls the side-looking UAV II at the starting time point to fly from the initial position to the starting point of the second flight route and repeat flying along the second flight route until the time difference between the current time and the time of the element closest to the starting time point QI2 is less than or equal to t 02 , and the side-looking UAV II flies back to the initial position to charge; During flight, when the overhead drone I or overhead drone II reaches each shooting position, it takes a top-down image of the substation and retains the pixels with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixels; During flight, when the side-view drone I or side-view drone II reaches each shooting position, it takes a side-view image of the substation and retains the pixels with the same pixel coordinates as the calibrated pixel points of the power equipment at the current shooting position, denoted as reference pixels; Each time the chip obtains a set of reference pixels, it compares the reference pixels with the pixel values and alarm thresholds of the calibrated pixel points of the power equipment at the same coordinates in the infrared monitoring database of the top-down image; If the pixel values of all reference pixels are less than or equal to the pixel values of the calibrated pixel points of the corresponding power equipment, the power equipment is operating normally at the current shooting position; If there are reference pixels with pixel values greater than the pixel values of the calibrated pixel points of the power equipment and less than the alarm threshold, the chip transmits the power equipment code, power equipment location code, and cooling signal corresponding to the reference pixel to the control terminal; After receiving the cooling signal, the control terminal controls the cooling equipment to cool the area where the power equipment is located; If there are reference pixels with pixel values greater than the alarm threshold of the calibrated pixel points of the power equipment, the chip transmits the power equipment code, power equipment location code, and alarm signal corresponding to the reference pixel to the control terminal and controls the alarm device carried by the drone to issue an alarm; The control terminal transmits the alarm signal to the user terminal.

2. The infrared monitoring method for indoor substation equipment based on camera cruising according to claim 1, wherein: The overhead drone carries at least one infrared camera, which is located below the fuselage of the overhead drone and is used to take a top-down infrared monitoring image of the power equipment; The side-view drone carries at least two infrared cameras, which are located on both sides of the fuselage of the side-view drone and are used to take side-view infrared monitoring images of the power equipment.

3. The infrared monitoring method for indoor substation equipment based on camera cruising according to claim 1, characterized in that: In the normal operating state, the infrared images D = [D1 D2... D n of each horizontal azimuth shooting area, and the infrared images F = [F1 F2... F m of each lateral azimuth shooting area, the side view image of the substation captured by the side view UAV, and the top view image of the substation captured by the top view UAV are all single-channel infrared images.

4. A method for infrared monitoring of indoor substation equipment based on camera cruise according to claim 1, characterized in that: In step 5.15), the method of selecting a point as the calibrated pixel point of the power equipment from each sub-region is: randomly select the calibrated pixel point of the power equipment.

5. The infrared monitoring method for indoor substation equipment based on camera cruising according to claim 1, characterized in that: In step 5.15), the method of selecting a point as the calibrated pixel point of the power equipment from each sub-region is: use the pixel point corresponding to the median of the pixel values as the calibrated pixel point of the power equipment.

6. The infrared monitoring method for indoor substation equipment based on camera cruise according to claim 1, characterized in that, In step 2), when determining the flight height a of the overhead drone and the horizontal azimuth shooting position matrix B1, and the side-view shooting position matrix B2 of the side-view drone, the constraint is to obtain unobstructed side views, top views of all power equipment in the substation, and the minimum total number of images.

7. A method for infrared monitoring of indoor substation equipment based on camera cruising according to claim 1, characterized in that, When planning the first flight route and the second flight route, the constraint is that the straight-line distance between adjacent shooting positions is the shortest; Among them, the first flight route starts and ends at the horizontal azimuth shooting position closest to the drone charging pile; The second flight route starts and ends at the side-view shooting position closest to the drone charging pile.

8. The infrared monitoring method for indoor substation equipment based on camera cruising according to claim 1, characterized in that In step 5), the power equipment area in each infrared image is framed manually.

9. The infrared monitoring method for indoor substation equipment based on camera cruise according to claim 1, characterized in that, When the control terminal receives the alarm signal, it cuts off the power supply of the power equipment above the alarm threshold.

10. The infrared monitoring method for indoor substation equipment based on camera cruise according to claim 1, wherein, In step 2), the steps of determining the flight height a of the overhead-view drone, the horizontal azimuth shooting position matrix B1, and the side azimuth shooting position matrix B2 of the side-view drone include: 2.1) Read the design drawings of the indoor substation, determine the spatial layout of the indoor substation, and construct the adjacency matrix A of the rooms in the indoor substation, that is: In the formula, represents the adjacency relationship between room g1 and room g2; represents that room g1 and room g2 are not adjacent; represents that room g1 and room g2 = 1 represents that room g1 and room g2 are adjacent; represents that room g1 and room g2 are not adjacent; g1 = 1, 2,..., G; g2 = 1, 2,..., G; g2 ≠ g1; 2.2) Based on the maximum height of the electrical equipment in room g1, determine the flight height of the overhead-view drone; the initial value of g1 is 1; Based on the area U of room g1 g1 , and the area U captured by the overhead drone in a single shot g0 , determine the number of shots of the overhead drone Divide the cross-section of room g1 into U rectangular areas on average, and use the center point coordinates of each area as the horizontal and vertical coordinates of each shooting position of the overhead-view drone; Based on the flight height of the overhead-view drone and the horizontal and vertical coordinates of each shooting position of the overhead-view drone, determine the horizontal azimuth shooting positions of the overhead-view drone in room g1; 2.3) Divide room g1 vertically into multiple vertical areas according to the layout of the electrical equipment; for each vertical area, with the constraint of obtaining a panoramic image, determine the shooting times of the side-view overhead-view drone, and then determine the side azimuth shooting positions of the side-view drone; 2.4) Judge whether g1 > G holds. If so, enter step 2.5). Otherwise, let g1 = g1 + 1 and return to step 2.2); 2.5) Based on the adjacency matrix A of the rooms in the indoor substation, with the shortest distance as the constraint, sort all the horizontal azimuth shooting positions and sort all the side azimuth shooting positions, so as to determine the flight height a of the overhead-view drone, the horizontal azimuth shooting position matrix B1, and the side azimuth shooting position matrix B2 of the side-view drone.

Images