An adaptive control system for a patrol unmanned aerial vehicle

Abstract

This invention discloses an adaptive control system for inspection drones, belonging to the field of drone intelligent control technology. This adaptive control system can automatically inspect the furnace, greatly reducing workload and improving inspection efficiency compared to traditional inspection methods. Furthermore, the automatic inspection avoids high-altitude operations, reducing the danger of inspection work. This invention can achieve continuous self-positioning of the drone based on preset reference objects within the furnace. This method can operate independently of GPS navigation, enabling rapid updates of the drone's spatial position even in furnaces with weak GPS signals. The entire process can be performed without signal interaction with other devices or platforms, significantly reducing the impact of signal interference. Additionally, when performing inspection tasks, the drone can operate outside of visual range without affecting its autonomous positioning, greatly enhancing the drone's inspection range and capabilities.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) intelligent control technology, specifically, it relates to an adaptive control system for inspection UAVs. Background Technology

[0002] The boiler furnace of a power plant consists of an extension of the water-cooled wall and an extension of the rear flue, and houses the final reheater and the final superheater. Before inspecting the heating surfaces of the water-cooled wall or horizontal flue, scaffolding or a lifting platform must be erected. Inspectors must wear safety belts, carry instruments, and conduct inspections while standing on the lifting platform or climbing scaffolding tens of meters high, posing risks of working at height. Furthermore, the boiler furnace environment is typically dusty, hot, and poorly lit, with the risk of coke deposits falling from the heating surfaces, placing high demands on the health and mental well-being of the inspectors. Finally, traditional furnace inspections cannot avoid the erection and removal of lifting platforms or scaffolding, a task that generally takes 2-3 days after the furnace has completely cooled, increasing maintenance time.

[0003] With the continuous development of drone technology, new opportunities for innovation and transformation have been brought to this important boiler maintenance work. However, the inside of the furnace is a closed space with low GPS signal strength, making traditional GPS positioning unusable. In order to solve the above problems, this invention provides a method for drone positioning and navigation in a closed space such as the furnace. The invention provides the following technical solution. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive control system for inspection drones, which solves the problems of high difficulty and low accuracy in navigation and positioning of drones in enclosed spaces in the prior art.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] An adaptive control system for an inspection drone includes:

[0007] The vertical positioning unit is used to detect the vertical height of the inspection drone in real time and transmit it to the control unit;

[0008] The power drive unit is used to provide driving power for the inspection drone and adjust the position of the inspection drone;

[0009] BIM units are used to create a 3D model of the area to be inspected.

[0010] The positioning camera unit includes n positioning cameras that can be independently adjusted at 360 degrees, used to acquire image information of the reference object and transmit it to the image processing unit, where n≥2;

[0011] The inspection camera unit is used to acquire image information of the corresponding location of the inspected object and transmit it to the image processing unit.

[0012] The image processing unit is used to analyze and process the image information uploaded by the positioning camera unit and the inspection camera unit;

[0013] The control unit is used to receive the analysis results from the image processing unit and drive the power drive unit to adjust the spatial position of the UAV based on the analysis results from the image processing unit.

[0014] The above-mentioned method for operating an adaptive control system for an inspection drone includes the following steps:

[0015] S1. Establish a three-dimensional model of the area to be inspected using BIM units, and establish a three-dimensional coordinate system using this three-dimensional model as the object;

[0016] Set up reference objects and obtain the position coordinates of each reference object in the three-dimensional coordinate system;

[0017] S2. After the drone takes off and reaches the preset altitude, it maintains its flight altitude and selects n reference objects as positioning reference objects.

[0018] Each positioning camera corresponds to a positioning reference object. The shooting angle of each positioning camera is adjusted so that a positioning reference object is always in the center of the corresponding positioning camera's image.

[0019] Based on the drone's flight altitude, the position coordinates of a positioning reference object, and the deflection angle and direction of a positioning camera relative to a determined direction, a spatial coordinate (a1, b1, c1) of the drone is calculated.

[0020] By corresponding to other positioning reference objects, obtain the other n-1 spatial coordinates (a2, b2, c2), ..., (an, bn, cn);

[0021] The final calculated coordinates (ap, bp, cp) of the UAV are obtained, where ap = (a1 + a2 + ..., an) / n, bp = (b1 + b2 + ..., bn) / n, and cp = (c1 + c2 + ..., cn) / n;

[0022] S3. When the drone's flight altitude and position change, adjust the shooting angle of the positioning camera so that the corresponding reference object is always in the center of the corresponding positioning camera's image.

[0023] The distance r between the UAV and the reference object and the angle γ between the line connecting the reference object and the UAV and the horizontal direction are calculated based on the position coordinates of a reference object and the final calculated coordinates of the UAV.

[0024] Obtain control samples that satisfy r2≤r≤r1 and γ≤γ1, and mark them as candidate control samples;

[0025] Where r1 and r2 are both preset values, and γ1 is a preset value;

[0026] Select n control substances from the candidate control substances as localization control substances;

[0027] S4. When the position of the UAV changes, causing the corresponding positioning reference to change, mark the replacement positioning reference as A1 and the positioning reference to be replaced as A2.

[0028] Interrupt the spatial coordinate input of positioning reference A2, and use the spatial coordinates of other positioning reference objects to calculate the final calculated coordinates of the UAV;

[0029] The positioning camera corresponding to positioning reference object A2 adjusts its shooting angle based on the final calculated coordinates of the current drone and the spatial coordinates of positioning reference object A1, so that positioning reference object A1 is always in the center of the positioning camera's image.

[0030] The new final calculated coordinates are obtained by following the method in step S2;

[0031] Continue positioning and navigation of the drone according to the new final calculated coordinates;

[0032] S5. Adjust the position of the drone according to the real-time final calculated coordinates of the drone, so that the drone can inspect the inner wall of the furnace according to the planned route, and save the image information collected by the inspection camera unit.

[0033] As a further embodiment of the present invention, the vertical positioning unit is an infrared ranging sensor.

[0034] As a further aspect of the present invention, the adaptive control system also includes a supplementary lighting unit for providing illumination and increasing the illumination intensity at certain locations.

[0035] As a further aspect of the present invention, the method for selecting n reference materials as positioning reference materials is as follows:

[0036] The controller obtains the altitude value h1 of the inspection drone through the vertical positioning unit;

[0037] The controller reads the position coordinates of each reference object to obtain the height value h2 of each reference object;

[0038] Calculate |h2-h1|, and select the n reference objects with the smallest |h2-h1| as positioning reference objects, where n is the number of positioning cameras in the positioning camera unit.

[0039] As a further aspect of the present invention, before obtaining the final calculated coordinates (ap, bp, cp), it is necessary to verify the spatial coordinates. The specific method is as follows: determine whether |ai-a(i+1)| / [ai, a(i+1)]min≤β, |bi-b(i+1)| / [bi, b(i+1)]min≤β, and |ci-c(i+1)| / [ci, c(i+1)]min≤β are always simultaneously true. If they are true, the verification is passed and the final calculated coordinates are calculated. If they are not true, the corresponding spatial coordinates are excluded and the verification is performed again, where 1≤i≤n-1 and β is a preset value.

[0040] As a further aspect of the present invention, the positioning cameras corresponding to the excluded spatial coordinates are marked as abnormal positioning cameras;

[0041] After the inspection is completed, the number of times the positioning camera was marked as an abnormal positioning camera is counted, and the corresponding positioning camera is inspected.

[0042] As a further aspect of the present invention, the method for selecting n control objects as positioning control objects from the candidate control objects is as follows:

[0043] First, determine whether the current positioning control is a candidate control. If it is, do not switch the positioning control. If not, calculate the contrast coefficient G of each candidate control that is not currently a positioning control according to the formula G=α1*|r-r3|+α2*γ1.

[0044] The candidate control with the smallest contrast coefficient G was selected as the new positioning control to replace the corresponding positioning control.

[0045] As a further aspect of the present invention, the processing content of the image processing unit includes:

[0046] Identify the reference object and its position in the image captured by the positioning camera unit, and determine whether there is any heat deformation of the burner nozzle, ash accumulation and coking, deformation of the heated surface pipe wall, corrosion, or ash accumulation and coking in the image captured by the inspection camera unit.

[0047] The beneficial effects of this invention are:

[0048] (1) The present invention can automatically perform inspection of the furnace, which greatly reduces the workload and improves the inspection efficiency compared with the traditional inspection method. Furthermore, due to the realization of automatic inspection, high-altitude operation is avoided, reducing the danger of inspection work.

[0049] (2) The present invention can achieve continuous self-positioning of the UAV based on the preset reference objects in the furnace. This method can be separated from the GPS navigation method and can quickly update the spatial position of the UAV in the furnace where the GPS signal is weak. The whole process can be done without signal interaction with other devices or platforms, which greatly reduces the impact of signal interference. In addition, when the UAV performs inspection tasks, it can leave the visual range without affecting its autonomous positioning, which greatly improves the inspection range and inspection capability of the UAV.

[0050] (3) The present invention changes the shooting angle of the positioning camera so that the corresponding reference object is always in the center of the positioning camera's image. The coordinate position of the drone in the three-dimensional coordinate system is calculated by reading the rotation angle of the positioning camera, the height of the drone, and the coordinates of the reference object. The whole is positioned by vision, and the image recognition calculation amount during positioning is significantly reduced.

[0051] (4) The present invention reduces the problem of large calculation error caused by a single control by collecting two or more control objects at the same time and averaging them. Attached Figure Description

[0052] The invention will now be further described with reference to the accompanying drawings.

[0053] Figure 1 This is a schematic diagram of the framework structure of an adaptive control system for an inspection drone according to the present invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] An adaptive control system for inspection drones, such as Figure 1 As shown, it includes:

[0056] The vertical positioning unit is used to detect the vertical height of the inspection drone in real time and transmit it to the control unit;

[0057] In one embodiment of the present invention, the vertical positioning unit is an infrared ranging sensor;

[0058] The power drive unit is used to provide driving power for the inspection drone and adjust the position of the inspection drone;

[0059] The supplementary lighting unit is used to provide illumination and increase the light intensity in certain areas;

[0060] BIM units are used to create a 3D model of the area to be inspected.

[0061] The positioning camera unit includes n positioning cameras that can be independently adjusted in 360 degrees, used to acquire image information of the reference object and transmit it to the image processing unit;

[0062] Where n≥2;

[0063] The inspection camera unit is used to acquire image information of the corresponding location of the inspected object and transmit it to the image processing unit.

[0064] The image processing unit is used to analyze and process the image information uploaded by the positioning camera unit and the inspection camera unit, identify the reference object and its position in the image captured by the positioning camera unit, and determine whether there are any issues such as heat deformation of the burner nozzle, ash accumulation and coking, deformation of the heated surface pipe wall, corrosion, etc. in the image captured by the inspection camera unit.

[0065] The control unit is used to receive the analysis results from the image processing unit and drive the power drive unit to adjust the spatial position of the UAV based on the analysis results from the image processing unit.

[0066] The above-mentioned method for operating an adaptive control system for an inspection drone includes the following steps:

[0067] S1. Establish a three-dimensional model of the area to be inspected using BIM units, and establish a three-dimensional coordinate system using this three-dimensional model as the object;

[0068] Set up reference objects and obtain the position coordinates of each reference object in the three-dimensional coordinate system;

[0069] The reference object can be a structure that is already inside the power plant furnace or a structure that is temporarily set up during inspection work. The reference object should be unique and conspicuous within the power plant furnace. Uniqueness means that the reference object is different from other structures in the power plant furnace, and conspicuousness means that there is a significant difference between the reference object and the background, as well as between the reference object and other structures.

[0070] It should be noted that the size of the reference object should not be too large. If it is too large, the reference object will occupy a large area in the image when the positioning camera unit and the inspection camera unit acquire images of the reference object, which will result in a large error in the processing results of the image processing unit.

[0071] S2. After the drone takes off and reaches the preset altitude, it maintains its flight altitude. The controller obtains the altitude value h1 of the inspection drone through the vertical positioning unit.

[0072] The controller reads the position coordinates of each reference object to obtain the height value h2 of each reference object;

[0073] Calculate |h2-h1|, and select the n reference objects with the smallest |h2-h1| as positioning reference objects, where n is the number of positioning cameras in the positioning camera unit;

[0074] Each positioning camera corresponds to a positioning reference object. The shooting angle of each positioning camera is adjusted so that a positioning reference object is always in the center of the corresponding positioning camera's image.

[0075] Based on the drone's flight altitude, the position coordinates of a positioning reference object, and the deflection angle and direction of a positioning camera relative to a determined direction, a spatial coordinate (a1, b1, c1) of the drone is calculated.

[0076] By corresponding to other positioning reference objects, obtain the other n-1 spatial coordinates (a2, b2, c2), ..., (an, bn, cn);

[0077] The final calculated coordinates (ap, bp, cp) of the UAV are obtained, where ap = (a1 + a2 + ..., an) / n, bp = (b1 + b2 + ..., bn) / n, and cp = (c1 + c2 + ..., cn) / n;

[0078] This step involves changing the shooting angle of the positioning camera to ensure that the corresponding reference object is always in the center of the positioning camera's image. The coordinate position of the drone in the three-dimensional coordinate system is calculated by reading the rotation angle of the positioning camera, the altitude of the drone, and the coordinates of the reference object.

[0079] This step also reduces the problem of large calculation errors caused by a single control by collecting two or more control samples at the same time and averaging the results.

[0080] Before obtaining the final calculated coordinates (ap, bp, cp), the spatial coordinates need to be verified. The specific method is as follows: determine whether |ai-a(i+1)| / [ai, a(i+1)]min≤β, |bi-b(i+1)| / [bi, b(i+1)]min≤β, and |ci-c(i+1)| / [ci, c(i+1)]min≤β are always true at the same time. If they are true, the verification is passed and the final calculated coordinates are calculated. If they are not true, the corresponding spatial coordinates are excluded and the verification is performed again. The positioning cameras corresponding to the excluded spatial coordinates are marked as abnormal positioning cameras.

[0081] Where 1≤i≤n-1, and β is a preset value;

[0082] After the inspection is completed, the number of times the positioning camera was marked as an abnormal positioning camera is counted, and the corresponding positioning camera is inspected.

[0083] S3. When the drone's flight altitude and position change, adjust the shooting angle of the positioning camera so that the corresponding reference object is always in the center of the corresponding positioning camera's image.

[0084] The distance r between the UAV and the reference object and the angle γ between the line connecting the reference object and the UAV and the horizontal direction are calculated based on the position coordinates of a reference object and the final calculated coordinates of the UAV.

[0085] Obtain control samples that satisfy r2≤r≤r1 and γ≤γ1, and mark them as candidate control samples;

[0086] r1 and r2 are both preset values. The values ​​of r1 and r2 should be able to ensure that the distance between the drone and the corresponding reference object is appropriate. γ1 is a preset value. The value of γ1 should be able to ensure that the angle between the drone and the corresponding reference object is appropriate.

[0087] To avoid the image processing unit's processing results being inaccurate due to the distance between the reference object and the drone being too close, too far, or the angle being too large;

[0088] Select n control substances from the candidate control substances as localization control substances;

[0089] In one embodiment of the present invention, the method for selecting n control objects as positioning control objects from candidate control objects is as follows:

[0090] First, determine whether the current positioning control is a candidate control. If it is, do not switch the positioning control. If not, calculate the contrast coefficient G of each candidate control that is not currently a positioning control according to the formula G=α1*|r-r3|+α2*γ1.

[0091] Select the candidate control with the smallest contrast coefficient G as the new positioning control to replace the corresponding positioning control;

[0092] S4. When the position of the UAV changes, causing the corresponding positioning reference to change, mark the replacement positioning reference as A1 and the positioning reference to be replaced as A2.

[0093] Interrupt the spatial coordinate input of positioning reference A2, and use the spatial coordinates of other positioning reference objects to calculate the final calculated coordinates of the UAV;

[0094] The positioning camera corresponding to positioning reference object A2 adjusts its shooting angle based on the final calculated coordinates of the current drone and the spatial coordinates of positioning reference object A1, so that positioning reference object A1 is always in the center of the positioning camera's image.

[0095] The new final calculated coordinates are obtained by following the method in step S2;

[0096] Continue positioning and navigation of the drone according to the new final calculated coordinates;

[0097] This invention enables continuous self-positioning of drones based on preset reference objects inside the furnace. This method can be independent of GPS navigation and can quickly update the spatial position of drones in furnaces where GPS signals are weak. The entire process can be completed without signal interaction with other devices or platforms, greatly reducing the impact of signal interference. In addition, when the drone is performing inspection tasks, it can move out of the visual range without affecting its autonomous positioning, greatly improving the inspection range and inspection capabilities of the drone.

[0098] S5. Adjust the position of the drone according to the real-time final calculated coordinates of the drone, so that the drone can inspect the inner wall of the furnace according to the planned route, and save the image information collected by the inspection camera unit.

[0099] This invention can automatically perform inspections of the furnace, which greatly reduces the workload and improves inspection efficiency compared to traditional inspection methods. Furthermore, due to the realization of automatic inspection, high-altitude operations are avoided, reducing the danger of inspection work.

[0100] In the description of this specification, the references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0101] The above description is merely an example and illustration of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. An adaptive control system for a patrol unmanned aerial vehicle, characterized in that, include: The vertical positioning unit is used to detect the vertical height of the inspection drone in real time and transmit it to the control unit; The power drive unit is used to provide driving power for the inspection drone and adjust the position of the inspection drone; BIM units are used to create a 3D model of the area to be inspected. The positioning camera unit includes n positioning cameras that can be independently adjusted at 360 degrees, used to acquire image information of the reference object and transmit it to the image processing unit, where n≥2; The inspection camera unit is used to acquire image information of the corresponding location of the inspected object and transmit it to the image processing unit. The image processing unit is used to analyze and process the image information uploaded by the positioning camera unit and the inspection camera unit; The control unit receives the analysis results from the image processing unit and drives the power drive unit to adjust the spatial position of the UAV based on the analysis results. The above-mentioned method for operating an adaptive control system for an inspection drone includes the following steps: S1. Establish a three-dimensional model of the area to be inspected using BIM units, and establish a three-dimensional coordinate system using this three-dimensional model as the object; Set up reference objects and obtain the position coordinates of each reference object in the three-dimensional coordinate system; S2. After the drone takes off and reaches the preset altitude, it maintains its flight altitude and selects n reference objects as positioning reference objects. Each positioning camera corresponds to a positioning reference object. The shooting angle of each positioning camera is adjusted so that a positioning reference object is always in the center of the corresponding positioning camera's image. Based on the drone's flight altitude, the position coordinates of a positioning reference object, and the deflection angle and direction of a positioning camera relative to a determined direction, a spatial coordinate (a1, b1, c1) of the drone is calculated. By corresponding to other positioning reference objects, obtain the other n-1 spatial coordinates (a2, b2, c2), ..., (an, bn, cn); The final calculated coordinates (ap, bp, cp) of the UAV are obtained, where ap = (a1 + a2 + ..., an) / n, bp = (b1 + b2 + ..., bn) / n, and cp = (c1 + c2 + ..., cn) / n; S3. When the drone's flight altitude and position change, adjust the shooting angle of the positioning camera so that the corresponding reference object is always in the center of the corresponding positioning camera's image. The distance r between the UAV and the reference object and the angle γ between the line connecting the reference object and the UAV and the horizontal direction are calculated based on the position coordinates of a reference object and the final calculated coordinates of the UAV. Obtain control samples that satisfy r2≤r≤r1 and γ≤γ1, and mark them as candidate control samples; Where r1 and r2 are both preset values, and γ1 is a preset value; Select n control substances from the candidate control substances as localization control substances; S4. When the position of the UAV changes, causing the corresponding positioning reference to change, mark the replacement positioning reference as A1 and the positioning reference to be replaced as A2. Interrupt the spatial coordinate input of positioning reference A2, and use the spatial coordinates of other positioning reference objects to calculate the final calculated coordinates of the UAV; The positioning camera corresponding to positioning reference object A2 adjusts its shooting angle based on the final calculated coordinates of the current drone and the spatial coordinates of positioning reference object A1, so that positioning reference object A1 is always in the center of the positioning camera's image. The new final calculated coordinates are obtained; Continue positioning and navigation of the drone according to the new final calculated coordinates; S5. Adjust the position of the drone according to the real-time final calculated coordinates of the drone, so that the drone can inspect the inner wall of the furnace according to the planned route, and save the image information collected by the inspection camera unit. Before obtaining the final calculated coordinates (ap, bp, cp), the spatial coordinates need to be verified. The specific method is as follows: determine whether |ai-a(i+1)| / [ai, a(i+1)]min≤β, |bi-b(i+1)| / [bi, b(i+1)]min≤β, and |ci-c(i+1)| / [ci, c(i+1)]min≤β are always true at the same time. If they are true, the verification is passed and the final calculated coordinates are calculated. If they are not true, the corresponding spatial coordinates are excluded and the verification is performed again. Here, 1≤i≤n-1 and β is a preset value. The method for selecting n control subjects as localization control subjects from the candidate control subjects is as follows: First, determine whether the current positioning control is a candidate control. If it is, do not switch the positioning control. If not, calculate the contrast coefficient G of each candidate control that is not currently a positioning control according to the formula G=α1*|r-r3|+α2*γ1. The candidate control with the smallest contrast coefficient G was selected as the new positioning control to replace the corresponding positioning control.

2. The self-adaptive control system for the inspection unmanned aerial vehicle according to claim 1, wherein The vertical positioning unit is an infrared ranging sensor.

3. The self-adaptive control system for the inspection unmanned aerial vehicle according to claim 1, wherein The adaptive control system also includes a supplemental lighting unit to provide illumination and enhance the light intensity in certain areas.

4. The adaptive control system for an inspection drone according to claim 1, characterized in that, The method for selecting n control objects as localization control objects is as follows: The controller obtains the altitude value h1 of the inspection drone through the vertical positioning unit; The controller reads the position coordinates of each reference object to obtain the height value h2 of each reference object; Calculate |h2-h1|, and select the n reference objects with the smallest |h2-h1| as positioning reference objects, where n is the number of positioning cameras in the positioning camera unit.

5. The adaptive control system for an inspection drone according to claim 1, characterized in that, And mark the positioning cameras corresponding to the excluded spatial coordinates as abnormal positioning cameras; After the inspection is completed, the number of times the positioning camera was marked as an abnormal positioning camera is counted, and the corresponding positioning camera is inspected.

6. The self-adaptive control system for the inspection drone according to claim 1, wherein The image processing unit processes the following: Identify the reference object and its position in the image captured by the positioning camera unit, and determine whether there is any heat deformation of the burner nozzle, ash accumulation and coking, deformation of the heated surface pipe wall, corrosion, or ash accumulation and coking in the image captured by the inspection camera unit.

Images