A bridge crack detection device using unmanned aerial vehicle and image processing method thereof

Abstract

The present invention relates to a bridge crack detection device by unmanned aerial vehicle and an image processing method thereof. The device of the present invention comprises a frame, a camera fixing platform is arranged at the front end of the frame, a spherical camera is arranged on the camera fixing platform, a laser rangefinder is arranged below the upper lens of the spherical camera, a rotor part is symmetrically arranged on the outer side of the frame, the rotor part comprises an integrated rotor cover and legs symmetrically arranged at the lower part of the integrated rotor cover, the integrated rotor cover comprises a cantilever bracket connected to the frame, a support cover is uprightly arranged at the cantilever end of the cantilever bracket, a propeller motor is arranged in the support cover, a rotor is fixedly connected to the rotating end of the propeller motor, a protective cover is arranged on the top of the support cover, an upper baffle is suspended from the top of the support cover to the outer side of the support cover, the free end of the upper baffle is located below the support cover, the rotor is located on the inner side of the upper baffle, and a side baffle is arranged around the outer side of the free end of the upper baffle.

Description

Technical Field

[0001] The present invention belongs to the technical field of engineering detection, relates to bridge detection, and specifically relates to a bridge crack detection device using an unmanned aerial vehicle and an image processing method thereof. Background Art

[0002] With the development of drone technology, drones have been widely used as a new auxiliary detection method in the fields of construction, roads and bridges. In terms of bridge detection, drones are mainly used to replace traditional manual methods to detect and take pictures of some specific types of diseases (mainly including uneven bridge bottoms, cracks, etc.) and some bridge parts that are difficult for personnel to reach or dangerous, among which cracks are the main detection content.

[0003] To photograph bridge cracks, it is usually necessary to get the bridge crack drone detection device as close to the bridge surface as possible to obtain a clearer image. However, the existing device obstacle avoidance is based on infrared or ultrasonic sensors and lacks physical obstacle avoidance devices. If it is caused by misoperation, communication loss or strong winds, the drone may hit the beam and be damaged. At the same time, due to the lack of reference objects for the captured crack images, the operators of the existing bridge crack drone detection devices cannot know the specific length and width of the cracks. In terms of wireless communication, the existing bridge crack drone detection devices use WiFi transmission between the remote control and the drone, which has a limited transmission distance and is easily blocked by the signal generated by the bridge structure.

[0004] In addition, in the application of existing bridge crack drone detection devices for bridge crack detection, the camera is generally placed directly below the gimbal, and the lens rotation angle is mostly positive 30° to negative 90°, which results in the problem of being unable to collect data on the bottom surface of the bridge, i.e., the area above the bridge crack drone detection device.

[0005] Therefore, at this stage, the bridge crack drone detection device cannot fully meet the needs of bridge crack detection, and there is still a lot of room for development in this professional field. Summary of the invention

[0006] The present invention proposes a bridge crack detection device using a drone and an image processing method thereof. The device of the present invention has physical obstacle avoidance measures that allow the drone to get as close to the bridge area as possible to obtain a more accurate image. The image processing method of the device of the present invention uses a laser rangefinder to measure the distance between the lens and the surface of the beam object being photographed and calculates the conversion relationship between the image pixel size and the actual size of the photographed surface area, uses an image recognition method to identify the shape of the crack and the pixel size occupied, and finally converts the actual length and width of the crack.

[0007] In order to achieve the above purpose, the technical solution adopted by the present invention is: a bridge crack drone detection device, including a frame, a camera fixing platform is arranged at the front end of the frame, a spherical camera is arranged on the camera fixing platform, a laser rangefinder is arranged below the upper lens of the spherical camera, and a rotor part is symmetrically arranged on the outer side of the frame, the rotor part includes an integrated rotor cover and legs symmetrically arranged at the lower part of the integrated rotor cover, the integrated rotor cover includes a cantilever bracket connected to the frame, a support cover is upright at the cantilever end of the cantilever bracket, and a support cover is arranged inside the support cover. A propeller motor is provided, the rotating end of the propeller motor is fixedly connected with a rotor, a protective cover is provided on the top of the support cover, an upper baffle is suspended from the top of the support cover to the outside of the support cover, the free end of the upper baffle is located below the support cover, the rotor is located on the inner side of the upper baffle, and a side baffle is provided on the outer side of the free end of the upper baffle; a signal receiving module is provided in the frame, the signal receiving module is connected to the control processor in the frame, a storage unit is provided on the control processor, a battery is provided on one side of the control processor, and the control processor is electrically connected to the battery.

[0008] Furthermore, the supporting legs are vertically placed at the bottom of the side levers, and the lower ends of the supporting legs are located below the frame.

[0009] Furthermore, a propeller motor is inverted in the support cover, and the rotor includes a centrally arranged rotating rod and symmetrically arranged wings on both sides of the rotating rod. The rotating rod is coaxially connected to the rotating shaft of the propeller motor, and the rotation of the propeller motor drives the rotor to rotate synchronously.

[0010] Furthermore, the rotor is arranged parallel to the frame body, the side lever is arranged at the same height as the rotor, and the lower part of the frame is located below the rotor.

[0011] Furthermore, the cantilever bracket is symmetrically arranged in four groups on both sides of the frame. The cantilever bracket is a high and low rod. The lower end of the cantilever bracket is fixed on the frame, the high end extends to the upper part of the support cover, and the high end of the cantilever bracket is located below the protective cover.

[0012] Furthermore, a wiring hole is provided in the cantilever bracket along its arrangement direction, and the wires of the propeller motor are electrically connected to the control processor along the wiring hole.

[0013] An image processing method for a bridge crack detection device using an unmanned aerial vehicle comprises the following steps:

[0014] 1) The laser rangefinder emits laser light to the surface of the beam being measured. The spherical camera can record the vertical distance between the frame shooting lens and the detection beam structure area in real time during the flight shooting process;

[0015] 2) The conversion relationship between the pixel size of the photo and the size of the actual detection area can be obtained through calculation program;

[0016] 3) Identify and calculate the actual length and width of bridge cracks.

[0017] Furthermore, in the step 2), the proportional conversion relationship between the unit pixel of the photo and the actual shooting area size is calculated based on the distance measured in the step 1) and the parameters of the lens shooting, and the pixel size occupied by the crack is identified.

[0018] Furthermore, in the image processing method of the bridge crack detection device,

[0019] After the laser rangefinder measures the distance, it is based on the following formula:

[0020] f=h×D / H

[0021] in:

[0022] f = focal length

[0023] h = height of the image sensor in the spherical camera

[0024] D = distance from the spherical camera lens to the bridge being measured

[0025] H = actual height of the shooting area

[0026] The actual height H of the shooting detection area is calculated by the above formula; according to the resolution of the photo in the height direction, that is, the pixel size, the conversion relationship between the unit pixel point of the photo and the actual shooting area size is obtained, and the cracks are identified by adjusting the gray value of the photo and performing edge detection, crack connection and linear recognition; the number of pixels occupied by the crack is calculated through the established image processing program; the actual length and width of the crack are calculated based on the actual size corresponding to the unit pixel.

[0027] A further solution: After the laser ranging module measures the real-time distance, it is based on the following formula:

[0028] f=h*D / H

[0029] f=u*D / U

[0030] in:

[0031] f = focal length

[0032] h = height of the image sensor in the spherical camera

[0033] u = width of the image sensor in the spherical camera

[0034] D = distance from the spherical camera lens to the bridge being measured

[0035] U = actual width of the image capture area

[0036] H = actual height of the shooting area

[0037] When the focal length f, the width h of the image sensor in the camera, and the distance D from the lens to the bridge to be measured are known, the above formula can be used to obtain the actual height H or actual width U of the bridge to be measured, that is, the picture. The photo can be opened with an image editing or processing software program and the pixel grid can be adjusted to a 1:1 size. The specific number of pixel grids in the length and width directions of the photo can be determined by checking the basic pixel information of the image, that is, the specific resolution. The calculated H or U is divided by the total number of all pixel grids contained in a single direction of the image to obtain the size of a single pixel grid. Based on the mathematical programming software MATLAB, the image itself is adjusted by adjusting the grayscale value and then performing edge detection, crack connection, and linear recognition to highlight the crack itself. The MATLAB program will automatically identify the crack and the number of pixel grids it occupies. The single pixel grid size obtained through the above steps is the total length and width of the toilet crack.

[0038] The technical effect of the present invention is that the device of the present invention has physical obstacle avoidance measures that allow the drone to get as close to the measured bridge area as possible to obtain more accurate images. The image processing method of the device of the present invention uses a laser rangefinder to measure the distance between the lens and the surface of the beam object being photographed and calculates the conversion relationship between the image pixel size and the actual size of the photographed surface area, uses an image recognition method to identify the shape of the crack and the pixel size occupied, and finally converts the actual length and width of the crack. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The present invention will be further described below using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0040] Figure 1 It is a schematic diagram of the top view of the structure of the bridge crack detection drone device of the present invention;

[0041] Figure 2 This is a schematic diagram of the main structure of the bridge crack detection drone device of the present invention;

[0042] Figure 3 A structural diagram of the coordination between the rotor and the propeller motor of the bridge crack detection drone device of the present invention;

[0043] exist Figure 1The invention comprises: 1-frame; 2-camera fixing gimbal; 3-spherical camera; 4-laser rangefinder; 5-overhang bracket; 6-support cover; 7-upper lever; 8-side lever; 9-leg; 10-rotor; 11-5G signal receiving module; 12-control processor; 13-battery; 14-storage device. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The present invention will be further described below in conjunction with the accompanying drawings.

[0046] A bridge crack detection device by unmanned aerial vehicle includes a frame 1, a camera fixing platform 2 is arranged at the front end of the frame 1, a spherical camera 3 is arranged on the camera fixing platform 2, a laser rangefinder 4 is arranged below the upper lens of the spherical camera 3, a rotor part is symmetrically arranged on the outer side of the frame 1, the rotor part includes an integrated rotor cover and legs 9 symmetrically arranged at the lower part of the integrated rotor cover, the integrated rotor cover includes a cantilever bracket 5 connected to the frame 1, a support cover 6 is vertically arranged at the cantilever end of the cantilever bracket 5, a propeller motor 6a is arranged in the support cover 6, and the rotating end of the propeller motor 6a is fixedly connected to The rotor 10 and the support cover 6 are provided with a protective cover piece 6b on the top, an upper baffle 7 is suspended from the top of the support cover 6 to the outside of the support cover 6, the free end of the upper baffle 7 is located below the support cover 6, the rotor 10 is located on the inner side of the upper baffle 7, and the free end of the upper baffle 7 is surrounded by a side baffle 8 on the outside; a 5G signal receiving module 11 is provided in the rack 1, the 5G signal receiving module 11 is connected to the control processor 12 in the rack 1, a storage unit 14 is provided on the control processor 12, a battery 13 is provided on one side of the control processor 12, and the control processor 12 is electrically connected to the battery 13.

[0047] For further information, see Attachment Figure 1 , 2 The legs 9 are symmetrically arranged at the left front, right front, left rear and right rear of the bridge crack UAV detection device.

[0048] Furthermore, the supporting leg 9 is vertically placed at the bottom of the side lever 8 , and the lower end of the supporting leg 9 is located below the frame 1 .

[0049] Furthermore, the propeller motor 6a is inverted in the support cover 6, and the rotor

[0050] 10 includes a centrally arranged rotating rod and winglets symmetrically arranged on both sides of the rotating rod. The rotating rod is coaxially connected to the rotating shaft of the propeller motor 6a. The rotation of the propeller motor 6a drives the rotor 10 to rotate synchronously.

[0051] That is, the rotating shaft of the propeller motor 6a is arranged downward.

[0052] Furthermore, the rotor 10 is arranged parallel to the frame of the frame 1, the side lever 8 is arranged at the same height as the rotor 10, and the lower part of the frame 1 is located below the rotor 10.

[0053] Furthermore, the cantilever bracket 5 is symmetrically arranged in four groups on both sides of the frame 1. The cantilever bracket 5 is a high-low rod. The lower end of the cantilever bracket 5 is fixed on the frame 1 and the high end extends to the upper part of the support cover 6. The high end of the cantilever bracket 5 is located below the protective cover 6b.

[0054] Reference Figure 1 The spherical camera 3 is electrically connected to the control processor 12 , and the video images taken by the spherical camera 3 are processed by the built-in control processor 11 and stored in the storage unit 14 .

[0055] Furthermore, wiring holes are provided in the cantilever bracket 5 along its arrangement direction, and the wires of the propeller motor 6a are electrically connected to the control processor 12 along the wiring holes.

[0056] Referring to the accompanying drawings, the present invention includes a frame 1, a spherical camera 3, an integrated rotor cover, a laser rangefinder 4, etc. Further, a 5G signal receiving module, etc. is arranged in the frame 1. The integrated rotor cover includes four rotors 10 arranged symmetrically in front and back and left and right. The rotors 10 are arranged downward, and the rotors 10 are located below the upper block rod 7 and are surrounded by the side block rod 8. Figure 2 , Figure 2The rotor 10 set downward in the middle is blocked by the side lever 8, and the spherical camera 3 is placed at the front of the drone frame 1 in the horizontal direction. Further, the integrated rotor cover includes a cantilever bracket 5 symmetrically extending outward from both sides of the drone frame 1, an upper lever 7 and a side lever 8, and a support cover 6 arranged on the cantilever bracket 5. The laser rangefinder 4 is located below the lens of the spherical camera 3, and the laser rangefinder 4 can rotate synchronously with the lens. The laser irradiation direction is parallel to the direction of the spherical camera 3 and perpendicular to the surface being photographed. When taking a photo, the corresponding distance to the photographed object is recorded at the same time. The 5G signal receiving module 11 is built into the drone gimbal, that is, the frame 1, and a corresponding signal module is provided on the remote control of the drone. Compared with the existing wifi transmission between the remote control and the drone, the transmission distance is longer and is not easily blocked by the signal generated by the bridge structure. And the present invention is connected to the remote control terminal through the 5G signal of the 5G signal receiving module 11, which can greatly improve the speed and stability of the transmission of signals and images between the bridge crack drone detection device and the remote control.

[0057] Furthermore, the integrated rotor cover is made of carbon fiber composite material.

[0058] The present invention adopts a physical obstacle avoidance structure of an integrated rotor cover, the integrated rotor cover and the rotor 10 form an integral structure, and the integrated rotor cover is injection molded with a carbon fiber composite material, which has high strength, is not easy to damage and is light in weight. Figure 1 The drive shaft of the drive unit of the rotor 10 of the present invention is arranged vertically downward. Specifically, a bayonet is provided at the rotating end of the propeller motor 6a, and the connecting end of the rotor 10 is fixedly clamped at the rotating end of the propeller motor 6a, and an upper stopper 7 is provided on the upper part of the propeller motor 6a, that is, on the upper part of the rotor 10. Figure 1 The upper baffle rod 7 and the cantilever bracket 5 form a cross-shaped protective cover for the propeller motor 6a and the rotor 10, and a protective cover sheet 6b is arranged on the top of the propeller motor 6a and a side baffle rod 8 is arranged on the periphery of the cross-shaped protective cover to form a physical anti-collision protection structure.

[0059] In addition, regarding the structure of the spherical camera 3 of the present invention: by placing the spherical camera 3 in front of the frame 1 in the horizontal direction and fixing the spherical camera 3 on both sides with mechanical axis devices, the spherical camera 3 can be rotated to shoot directly above and directly below, and the rotation and shooting angle is +90° to -90° in the up and down direction, and the total rotation angle is 180°. This can solve the problem that the conventional bridge crack drone detection device cannot shoot the top of the drone frame 1, that is, the bottom of the bridge in the forward direction.

[0060] In the present invention, the laser rangefinder 4 emits a laser to the surface of the beam to be measured, and the spherical camera 3 can record the vertical distance between the shooting lens of the frame 1 and the detection beam structure area in real time during the flight shooting process, so that the conversion relationship between the pixel size of the photo and the size of the actual detection area can be obtained through the calculation program, and it is used to further identify and calculate the actual length and width of the bridge crack in the later stage. That is, the laser rangefinder 4 is used to measure the distance between the lens of the spherical camera 3 and the surface of the beam of the object being photographed and calculate the conversion relationship between the pixel size of the image and the actual size of the surface area of ​​the beam of the object being photographed, and use the image recognition method to identify the shape of the crack and the pixel size occupied, and finally convert the actual length and width of the crack.

[0061] That is, further, the detection method of the bridge crack drone detection device of the present invention includes calculating the proportional conversion relationship between the unit pixel of the photo and the actual shooting area size based on the distance measured by the laser rangefinder 4 and the parameters captured by the spherical camera 3, identifying the pixel size occupied by the crack through an image recognition method, and finally calculating the actual length and width of the crack according to the above conversion relationship.

[0062] In summary, in the present invention, when performing flight photography of a specific bridge area, the drone frame 1 performs close-range photography of the surface of the bridge to be tested in the horizontal direction based on the integrated rotor cover as a physical obstacle avoidance and anti-collision rebound device in the horizontal direction. The integrated rotor cover serves as a physical obstacle avoidance and anti-collision rebound device in the vertical direction. The bridge crack drone detection device of the present invention can perform close-range photography of the surface of the bridge to be tested in the vertical direction, while protecting the rotor 10, i.e., the propeller, from damage. When performing the detection task, a close-range cruise photography of the target detection bridge area is performed, especially by adjusting the 0° to positive 90° lens range of the spherical camera 3 to photograph the bottom surface of the bridge. During the shooting process, the laser rangefinder 4 synchronously emits a laser and remains parallel to the direction of the spherical camera 3. The laser rangefinder 4 simultaneously records the vertical distance between the spherical camera 3 and the surface of the bridge to be tested into the screen in real time through the control processor 12 in the drone frame 1, and synchronously displays it as recorded data in the later video and picture.

[0063] Further preferably, a lighting device is provided on the spherical camera 3, so that when the picture is dark, the lighting device can supplement the light for the picture.

[0064] After the laser rangefinder 4 measures the distance, it is based on the following formula:

[0065] f=h×D / H

[0066] in:

[0067] f = focal length

[0068] h = the height of the image sensor in the camera (i.e., spherical camera)

[0069] D = distance from the (spherical camera) lens to the bridge being measured

[0070] H = actual height of the shooting area

[0071] In this way, when the focal length f, the width h of the image sensor in the camera, and the distance D from the lens to the bridge to be measured are known, the actual height H of the shooting detection area can be calculated by the above formula. According to the resolution of the photo in the height direction, that is, the pixel size, the conversion relationship between the unit pixel point of the photo and the actual shooting area size can be obtained, and the photo can be identified by edge detection, crack connection and linear recognition after adjusting the gray value. The number of pixels occupied by the crack is calculated through the established image processing program. The actual length and width of the crack are further calculated by the actual size corresponding to the unit pixel obtained in the above steps.

[0072] Further, after the laser ranging module 4 measures the real-time distance, based on the following formula:

[0073] f=h*D / H

[0074] f=u*D / U

[0075] in:

[0076] f = focal length

[0077] h = the height of the image sensor in the camera (i.e., spherical camera)

[0078] u = width of the image sensor in the (spherical camera)

[0079] D = distance from the (spherical camera) lens to the bridge being measured

[0080] U = actual width of the image capture area

[0081] H = actual height of the shooting area

[0082] When the focal length f, the width h of the image sensor in the camera, and the distance D from the lens to the bridge to be measured are known, the actual height H or actual width U of the bridge to be measured, i.e., the picture, can be obtained by the above formula. Further, the photo is opened with an image editing or processing software program and the pixel grid is adjusted to a 1:1 size, i.e., a square. By checking the basic pixel information of the image, i.e., the specific resolution, the specific number of pixel grids in the length and width directions of the photo can be determined. Further, based on the calculated H or U divided by the total number of all pixel grids contained in a single direction of the image, the size of a single pixel grid can be obtained.

[0083] Specifically, the calculation program (i.e., the established image processing program) is based on the mathematical programming software MATLAB, and the cracks themselves are highlighted by edge detection, crack connection, and linear recognition after adjusting the gray value of the image itself. Further, the MATLAB program will automatically identify the cracks and the number of pixels they occupy, and further the total length and width of the toilet cracks obtained by the single pixel size of the above steps.

[0084] The beneficial effects of all the above schemes are as follows: the spherical camera 3 has a wide shooting range, especially for shooting below the bridge deck; the integrated rotor cover as a physical anti-collision structure allows the bridge crack drone detection device of the present invention to perform safe close-range shooting; and the laser rangefinder 4 and the detection method adopted do not require a reference object but use a laser ranging method to obtain the size of the actual shooting area, and can identify cracks based on image recognition methods and calculate the actual length and width of the cracks, which greatly improves the scope of application, effectiveness and efficiency of drones for bridge detection.

[0085] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An image processing method for a bridge crack detection device using an unmanned aerial vehicle, the bridge crack detection device using an unmanned aerial vehicle comprises a frame (1), a camera fixing platform (2) is arranged at the front end of the frame (1), a spherical camera (3) is arranged on the camera fixing platform (2), a laser rangefinder (4) is arranged below the upper lens of the spherical camera (3), and a rotor portion is symmetrically arranged on the outer side of the frame (1), characterized in that: The rotor portion comprises an integrated rotor cover and legs (9) symmetrically arranged at the bottom of the integrated rotor cover, the integrated rotor cover comprises a cantilever bracket (5) connected to the frame (1), a support cover (6) is vertically arranged at the cantilever end of the cantilever bracket (5), a propeller motor (6a) is arranged in the support cover (6), a rotor (10) is fixedly connected to the rotating end of the propeller motor (6a), a protective cover sheet (6b) is arranged at the top of the support cover (6), an upper stop rod (7) is suspended from the top of the support cover (6) to the outside of the support cover (6), and the free end of the upper stop rod (7) is located at the support cover (6). Below the support cover (6), the rotor (10) is located on the inner side of the upper blocking rod (7), and the free end of the upper blocking rod (7) is surrounded by a side blocking rod (8); a 5G signal receiving module (11) is arranged in the frame (1), and the 5G signal receiving module (11) is connected to a control processor (12) in the frame (1), a storage unit (14) is arranged on the control processor (12), a battery (13) is arranged on one side of the control processor (12), and the control processor (12) is electrically connected to the battery (13); the image processing method comprises the following steps: 1) The laser rangefinder (4) emits laser light to the surface of the beam to be measured, and the spherical camera (3) can record the vertical distance between the shooting lens of the frame (1) and the structural area of ​​the beam to be measured in real time during the flight shooting process; 2) The conversion relationship between the pixel size of the photo and the size of the actual detection area can be obtained through the calculation program; 3) Identify and calculate the actual length and width of the bridge crack; After the laser rangefinder (4) measures the distance, it is based on the following formula: f=h×D / H in: f = focal length h = height of the image sensor in the spherical camera (3) D = distance from the spherical camera (3) lens to the bridge being measured H = Actual height of the captured area The actual height H of the shooting detection area is calculated by the above formula; according to the resolution of the photo in the height direction, that is, the pixel size, the conversion relationship between the unit pixel point of the photo and the actual shooting area size is obtained, and the cracks are identified by adjusting the gray value of the photo and performing edge detection, crack connection and linear recognition; the number of pixels occupied by the crack is calculated through the established image processing program; the actual length and width of the crack are calculated based on the actual size corresponding to the unit pixel.

2. The image processing method of the bridge crack detection device by drone according to claim 1 is characterized by: The supporting leg (9) is vertically placed at the bottom of the side lever (8), and the lower end of the supporting leg (9) is located below the frame (1).

3. The image processing method of the bridge crack detection device by drone according to claim 2 is characterized in that: A propeller motor (6a) is inverted in the support cover (6), and the rotor (10) comprises a centrally arranged rotating rod and symmetrically arranged wings on both sides of the rotating rod. The rotating rod is coaxially connected to the rotating shaft of the propeller motor (6a), and the rotation of the propeller motor (6a) drives the rotor (10) to rotate synchronously.

4. The image processing method of the bridge crack detection device using an unmanned aerial vehicle according to claim 3 is characterized by: The rotor (10) is arranged parallel to the frame (1), the side lever (8) is arranged at the same height as the rotor (10), and the lower part of the frame (1) is located below the rotor (10).

5. The image processing method of the bridge crack detection device by drone according to claim 1, 2, 3 or 4, characterized in that: The overhang brackets (5) are symmetrically arranged in four groups on both sides of the frame (1), and the overhang brackets (5) are high-low rods, the lower ends of the overhang brackets (5) are fixed to the frame (1), the higher ends extend to the upper part of the support cover (6), and the higher ends of the overhang brackets (5) are located below the protective cover sheet (6b).

6. The image processing method of the bridge crack detection device using an unmanned aerial vehicle according to claim 1 is characterized in that: A wiring hole is provided in the cantilever bracket (5) along its arrangement direction, and the electric wires of the propeller motor (6a) are electrically connected to the control processor (12) along the wiring hole.

7. The image processing method of the bridge crack detection device using an unmanned aerial vehicle according to claim 1 is characterized in that: In the step 2), the ratio conversion relationship between the unit pixel of the photo and the actual shooting area size is calculated based on the distance measured in the step 1) and the parameters of the lens shooting, and the pixel size occupied by the crack is identified.

8. The image processing method of the bridge crack detection device using an unmanned aerial vehicle according to claim 1 is characterized in that: The laser rangefinder (4) measures the real-time distance based on the following formula: At a known focal length , the height of the image sensor in the camera , width u and the distance from the lens to the bridge being measured In the case of or actual width , open the photo with an image editing or processing software program and adjust the pixel grid to 1:1 size. By checking the basic pixel information of the image, that is, the specific resolution, the specific number of pixels in the length and width of the photo can be determined. Based on the calculated or Divide by the total number of pixels contained in a single direction of the image to obtain the size of a single pixel. Based on the mathematical programming software MATLAB, the image itself is adjusted by adjusting the grayscale value and then performing edge detection, crack connection and linear identification to highlight the cracks themselves. The MATLAB program will automatically identify the cracks and the number of pixels they occupy. The total length and width of the cracks are calculated using the size of a single pixel obtained through the above steps.

Images