Unmanned aerial vehicle safety inspection method and system based on image recognition

Abstract

The invention belongs to the field of unmanned aerial vehicle inspection, and particularly relates to an unmanned aerial vehicle safety inspection method based on image recognition. Comprising the following steps that S1, a standard image of an object A is shot at a preset detection position, and the distance between the position where the standard image is shot and power equipment is a safe distance; s2, storing the standard image; s3, the unmanned aerial vehicle gets close to a preset detection position, and when the distance between the unmanned aerial vehicle and the preset detection position is smaller than or equal to a set distance, the unmanned aerial vehicle shoots contrast images of the object A multiple times; s4, comparing the contrast images shot for multiple times with the stored standard images in sequence; s5, when the size of the object A in the contrast image is larger than or equal to that of the object A in the standard image, the distance between the unmanned aerial vehicle and the power equipment is smaller than or equal to the safe distance, and the unmanned aerial vehicle stops moving. The invention provides an unmanned aerial vehicle safety inspection method and system based on image recognition, and aims to solve the problem that an unmanned aerial vehicle is close to power equipment due to the fact that positioning deviation is too large when an RTK module is subjected to electromagnetic interference.

Description

Technical Field

[0001] The present invention belongs to the field of drone inspection, and in particular relates to a drone safety inspection method and system based on image recognition. Background Art

[0002] With the in-depth application of drones in the power grid field, it has become a trend to use drones to inspect power equipment. Drones have the characteristics of high flight altitude and wider field of view. When used in power equipment inspection, drones can inspect places that cannot be inspected by humans or robots, and can promptly discover equipment defects and hidden dangers of power equipment. When equipment defects and hidden dangers are discovered, staff can promptly deal with the defects.

[0003] When using drones to inspect power equipment, the distance between the drone and the power equipment is very important. If the drone is too close to the power equipment, the drone may be affected by electromagnetic interference and lose control, causing damage to the drone and the power system. In the prior art, drones are equipped with RTK modules, which can accurately control the position of the drone to ensure the distance between the drone and the power equipment.

[0004] However, in practice, it is found that since power equipment will generate electromagnetic interference on the RTK module, it often causes excessive positioning errors, leading to problems where the drone and power equipment are too close, and both the drone and power equipment pose safety risks. Summary of the invention

[0005] The present invention provides a method and system for unmanned aerial vehicle safety inspection based on image recognition, which aims to solve the problem that the RTK module is affected by electromagnetic interference and has a large positioning deviation, resulting in the unmanned aerial vehicle being close to power equipment.

[0006] In order to achieve the above object, the present invention provides a UAV safety inspection method based on image recognition, comprising the following steps:

[0007] S1, taking a standard image of object A at a predetermined detection position, where the position where the standard image is taken is a safe distance from the power equipment;

[0008] S2, store the standard image;

[0009] S3, the UAV approaches the predetermined detection position. When the distance between the UAV and the predetermined detection position is less than or equal to the set distance, the UAV takes a comparison image of object A multiple times.

[0010] S4, comparing the reference images captured multiple times with the stored standard images in sequence;

[0011] S5, when the size of object A in the reference image ≥ the size of object A in the standard image, the distance between the UAV and the power equipment ≤ the safety distance, and the UAV stops moving.

[0012] This solution determines the distance between the drone and the power equipment by comparing the reference image with the standard image. Compared with the existing technology, the process of the camera taking images is not easily interfered by electromagnetic interference, thus ensuring that the distance between the drone and the power equipment can be determined more accurately in an electromagnetic environment, avoiding the drone from losing control or colliding with the power equipment, and solving the shortcomings of the existing technology.

[0013] Preferably, in order to determine the detection of the entire inspection path, in S1 of this solution, it is set that the inspection path is provided with a number of predetermined inspection positions, and a standard image is correspondingly captured at each predetermined detection position.

[0014] This solution sets several inspection positions to ensure that all parts to be inspected at each position on the set inspection path can be inspected. At the same time, by taking standard images at each predetermined inspection position, it is ensured that the drone will not collide with the power equipment at each predetermined inspection position.

[0015] Preferably, in order to improve efficiency, standard images can be captured at detection locations that are rarely visited by people and have complex environments. In S1 of this solution, the drone flies along a set inspection path and captures standard images at predetermined detection locations.

[0016] This solution allows the drone to fly along a set inspection route and then take standard images. Compared with manual photography, it is more efficient and convenient to complete photography at inspection locations with few people and complex environments.

[0017] Preferably, in order to ensure the accuracy of the position for shooting the standard image, in S1 of this solution, the drone is manually operated to fly along the set inspection path.

[0018] When shooting standard images, this solution uses manual operation to fly the drone, and humans can make position judgments and drone flight performance judgments, which is more accurate and safer.

[0019] Preferably, in order to avoid collision between the drone and the power equipment, in S3 of this solution, when the drone approaches the predetermined detection position, the flight path of the drone is recorded.

[0020] The flight path is recorded by the drone, and then the drone retreats along the flight path to avoid the drone's flight path error, thereby preventing the drone from getting out of control due to electromagnetic interference or colliding with power equipment when the drone is too close.

[0021] Preferably, in order to avoid collision between the drone and the power equipment, in S5, when the distance between the drone and the power equipment is less than the safety distance, the drone moves away from the power equipment along the original flight path.

[0022] Preferably, in order to solve the problem that the drone deviates significantly from the predetermined detection position, in S3 of this solution, when the drone approaches the predetermined detection position and the distance between the drone and the predetermined detection position is less than or equal to the initial detection distance, an initial detection image of object A is captured, and the initial detection distance is greater than the set distance;

[0023] By comparing the initial inspection image with the standard image, if object A in the initial inspection image is not identified, the drone will hover and wait for the operator to adjust the flight path so that the drone approaches the predetermined inspection position.

[0024] This solution sets an initial inspection distance. If object A is not identified in the initial inspection distance, it means that the drone has deviated significantly. At this time, through manual intervention, the drone is guided to move accurately in the direction of the predetermined inspection position, further ensuring that the drone can accurately reach the predetermined inspection position.

[0025] In order to solve the problem that the UAV is close to the power equipment, the second aspect of the present invention discloses a UAV safety inspection system based on image recognition, comprising:

[0026] A drone, which can fly along a set inspection path and move to a predetermined inspection location;

[0027] A camera, which is installed on the drone and is used to capture a standard image and a comparison image of object A;

[0028] A control module is connected to the drone and the camera signal.

[0029] This solution obtains a reference image by configuring a camera, and the control module compares the reference image with the standard image to determine the distance between the drone and the power equipment. The distance between the drone and the power equipment can be determined by the camera configured on the drone, which is more accurate than the existing technology.

[0030] Preferably, in order to more accurately determine the distance between the UAV and the power equipment, the present solution further includes an RTK module, which is installed on the UAV, the RTK module is signal-connected to the control module, and the RTK module is used to obtain the spatial position of the UAV.

[0031] This solution uses the RTK module and the camera to cooperate with each other to ensure more accurate distance measurement between the drone and the power equipment.

[0032] The beneficial effect of the present invention is that the present invention determines the distance between the UAV and the power equipment by comparing the reference image with the standard image. Compared with the prior art, the process of the camera taking the image is not easily interfered by electromagnetic interference, thereby ensuring that the distance between the UAV and the power equipment can be determined more accurately in an electromagnetic environment, avoiding collision between the UAV and the power equipment, and solving the shortcomings of the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 The figure is a structural diagram of a UAV safety inspection method based on image recognition. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the embodiments more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are only examples of devices and methods consistent with some aspects of the present disclosure as detailed in the attached claims.

[0035] It should be noted that all actions to obtain signals, information or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where they are located, and with the authorization given by the owner of the corresponding device.

[0036] In the present disclosure, unless otherwise stated, the directional words used, such as "inside" and "outside", are defined according to the corresponding parts' own contours. The terms used in the present disclosure, such as "first" and "second", etc., are used to distinguish one element from another element and do not have order and importance.

[0037] Example 1

[0038] Several predetermined detection positions are set on the set inspection path of the power equipment. The number of detection positions is determined according to the actual situation. For example, 5 detection positions, 10 detection positions or 15 detection positions can be set. Each detection position is installed with power equipment that needs to be inspected. The drone can fly along the set inspection path, and when the drone flies along the set inspection path, the drone can pass through the several detection positions set above in sequence. The drone can detect the power equipment by taking images of the power equipment at this position. Of course, the drone can also be equipped with relevant detection devices to detect the power equipment.

[0039] A UAV safety inspection method based on image recognition is basically as follows: Figure 1 As shown, the specific steps include:

[0040] S1, the drone is manually operated to fly along the set inspection path and pass through different inspection positions in sequence. When the drone arrives at different inspection positions, the camera configured by the drone takes a picture of object A at the position to obtain a standard image of object A. Object A can be the power equipment to be inspected, or object A can also be a corresponding identification mark, etc.

[0041] When the drone takes a standard image of object A, the drone is at a safe distance from object A. The safe distance can be 1 meter, 1.5 meters, or 2 meters, etc.

[0042] It is understandable that: at different detection positions, object A is also different. At different detection positions, the drone captures a standard image of object A at the position.

[0043] S2, the captured standard image is stored in the control module, and the control module is in a signal connection state with the camera. The control module is installed on the drone, and the staff can remotely control the control module.

[0044] S3, the UAV flies along the predetermined inspection path by itself, and moves towards the predetermined detection position under the positioning of the RTK module. At this time, the predetermined detection position under the positioning of the RTK module may be offset due to electromagnetic interference. However, even if this position may have been offset, the distance between the UAV and the predetermined detection position is also detected by the RTK module. When the RTK module detects that the distance between the UAV and the predetermined detection position is less than or equal to the set distance (for example, 1 meter), the camera starts working. The camera configured on the UAV continuously captures the comparison image of object A. The captured images are several, and they are continuously captured at the same time.

[0045] When the drone moves toward the predetermined detection position, due to the existence of electromagnetic interference, the problem of the drone offsetting too far may occur. Therefore, when the drone approaches the predetermined detection position under the guidance of the RTK module, when the distance between the drone and the predetermined detection position is ≤ the initial inspection distance (for example, 2 meters), the camera configured on the drone takes the initial inspection image of object A and compares the initial inspection image with the standard image. If object A cannot be identified in the initial inspection image, it means that the drone has a large offset direction and has deviated from the predetermined detection position. Therefore, at this time, manual intervention can be performed to manually operate the drone to move in the direction of the predetermined detection position, further ensuring that the drone moves in the direction of the predetermined detection position.

[0046] S4, comparing the captured control image with the stored standard image, wherein the comparison target is the size of object A in different images.

[0047] S5, when the size of object A in the control image ≥ the size of object A in the standard image, it means that the distance between the drone and the power equipment is ≤ the safe distance. For example: object A in the standard image occupies 1 / 3 of the entire image, and object A in the control image occupies 2 / 5 of the entire image, which means that the distance between the drone and the power equipment is less than the safe distance. At this time, in order to avoid electromagnetic interference loss of control or collision with power equipment after the drone is too close, the drone stops moving, and the drone takes an image of the power equipment at the stop position, thereby completing the detection of the power equipment.

[0048] S6, after completing the inspection of the power equipment, the drone moves away from the power equipment.

[0049] After that, the drone continues to fly along the set inspection path and arrives at the next predetermined inspection location to carry out inspection of the power equipment.

[0050] Example 2

[0051] This embodiment is improved on the basis of embodiment 1. When the drone is too close to the power equipment, the drone's autonomous flight may pose a safety risk due to electromagnetic interference.

[0052] For example, if object A in the standard image occupies 1 / 3 of the entire image, and object A in the control image occupies 1 / 2 of the entire image, it means that the distance between the drone and the power equipment is significantly smaller than the safe distance and smaller than the threshold (for example, the distance between the drone and the power equipment is less than 0.3 meters). At this time, because the drone is too close to the power equipment, the drone may collide with the power equipment due to electromagnetic interference.

[0053] In order to solve the above problem, in S3, when the drone approaches the predetermined detection position, the flight path of the drone is recorded. Then, in S5, when the distance between the drone and the power equipment is less than the safe distance and the distance between the drone and the power equipment is less than the threshold, the drone returns along the original flight path, and the drone stays away from the power equipment to avoid electromagnetic interference loss of control or collision with the power equipment when the drone is too close. By returning along the original route, the drone is prevented from colliding with the power equipment.

[0054] Alternatively, in order to solve the above problem, in some embodiments, when the distance between the drone and the power equipment is less than the safe distance and the distance between the drone and the power equipment is less than a threshold, the drone can remain in a hovering state. Then, the remote staff can manually operate the drone to keep the drone away from the power equipment to prevent the drone from colliding with the power equipment.

[0055] Example 3

[0056] This embodiment provides a drone safety inspection system based on image recognition, comprising:

[0057] The drone may be a drone in the prior art. The drone may fly along a set inspection path and may move toward a predetermined inspection position with the help of an RTK module.

[0058] The camera is installed on the drone. When the drone approaches the predetermined detection position, and the distance between the drone and the predetermined detection position is greater than or equal to a standard value (e.g., 1 meter), the camera continuously captures a comparison image of object A. The comparison image is sent to the control module. At the same time, when the drone hovers, the camera can capture images of the power equipment to be detected to achieve detection of the power equipment.

[0059] RTK module, RTK module is also installed on the UAV, and RTK module adopts the RTK module in the existing technology. RTK module continuously obtains the spatial position of the UAV.

[0060] The control module is in signal connection with the RTK module, the camera and the drone. The control module can store the standard image and compare the reference image with the standard image through an algorithm. Then, based on the measurement results of the RTK module and the camera, the control module controls the drone to perform operations such as hovering or flying. The control module can be an industrial control computer or a drone flight control module in the prior art.

[0061] It should be noted that: this embodiment determines the distance between the drone and the power equipment by configuring the RTK module and the camera at the same time. When the RTK module is interfered by electromagnetic interference and the positioning error is large, the distance between the drone and the power equipment can be determined by the camera with higher accuracy.

[0062] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A drone safety inspection method based on image recognition, characterized in that: The following steps are included: S1, taking a standard image of object A at a predetermined detection position, where the position where the standard image is taken is a safe distance from the power equipment; S2, store the standard image; S3, the UAV approaches the predetermined detection position. When the distance between the UAV and the predetermined detection position is less than or equal to the set distance, the UAV takes a comparison image of object A multiple times. S4, comparing the reference images captured multiple times with the stored standard images in sequence; S5, when the size of object A in the reference image ≥ the size of object A in the standard image, the distance between the UAV and the power equipment ≤ the safety distance, and the UAV stops moving.

2. The drone safety inspection method according to claim 1, characterized in that: In S1, the inspection path is set to have a number of predetermined inspection positions, and each predetermined inspection position corresponds to a photographed standard image.

3. The drone safety inspection method according to claim 1 or 2, characterized in that: In S1, the drone flies along a set inspection path and captures a standard image at a predetermined inspection position.

4. The drone safety inspection method according to claim 3, characterized in that: In S1, the drone is manually operated to fly along a set inspection path.

5. The drone safety inspection method according to claim 1, characterized in that: In S3, when the drone approaches the predetermined detection position, the flight path of the drone is recorded.

6. The drone safety inspection method according to claim 5, characterized in that: In S5, when the distance between the UAV and the power equipment is less than the safety distance, the UAV moves away from the power equipment along the original flight path.

7. The drone safety inspection method according to claim 1, characterized in that: In S3, when the drone approaches the predetermined detection position and the distance between the drone and the predetermined detection position is less than or equal to the initial detection distance, a preliminary inspection image of object A is captured, and the initial inspection distance is greater than the set distance; Compare the initial inspection image with the standard image. If object A in the initial inspection image is not identified, the drone will hover and wait for the operator to adjust the flight path so that the drone approaches the predetermined inspection position.

8. An unmanned aerial vehicle safety inspection system based on image recognition, characterized in that: include A drone, which can fly along a set inspection path and move to a predetermined inspection location; A camera, which is installed on the drone and is used to capture a standard image and a comparison image of object A; A control module is connected to the drone and the camera signal.

9. The drone safety inspection system according to claim 8, characterized in that: Also includes An RTK module is installed on the drone, the RTK module is connected to the control module signal, and the RTK module is used to obtain the spatial position of the drone.

Images