System and method for detecting accumulated water in steel box girder based on unmanned aerial vehicle and laser ultrasound
The inside stagnation of water box beams was performed through the drone with laser ultrasonic probes and Mask R-CNN algorithms, which solved the problem of low detection efficiency in the existing technology and unable to conduct non-destructive detection in harsh environments. Staged water detection.
Patent Information
- Application Number
- CN202411506854.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-28
- Publication Date
- 2025-01-24
AI Technical Summary
The existing internal water accumulation detection methods of steel box beams have problems such as large human subjective influence and the inability to detect plate beams without gaps, and it is difficult to conduct non-destructive testing in harsh environments.
Use the drone to carry laser ultrasonic probes for the inside of the steel box beam to detect the water tank beam, and perform non-damaged detection under high temperature, high pressure, wind wind and other bad conditions through non-contact type, and use the Mask R-CNN algorithm to extract the gray degree of obvious gradient changes. Value areas to reduce parameters and improve detection efficiency and accuracy.
It achieves efficient and accurate detection of water accumulation inside steel box beams in harsh environments, improves detection efficiency and accuracy, and avoids human subjective influence.
Smart Images
Figure CN119355796A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of steel box girder detection, and in particular relates to a system and method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound. Background Art
[0002] Steel box girder bridge is a common structural form for long-span bridges, with advantages such as small weight, convenient construction and beautiful appearance. Studies have found that under the influence of continuous natural environmental erosion and long-term operation and use, water will accumulate in the cavity of the steel box girder to varying degrees, which has an adverse effect on the durability of the box girder structure.
[0003] The current method for detecting water accumulation inside steel box girders is to use geological radar to detect the hollow slabs of bridges one span at a time and one piece at a time. Geological radar transmits high-frequency electromagnetic waves, and then performs wave phase analysis, interpretation and judgment based on spectrum processing and analysis. However, judgment is a comprehensive analysis combining theory and practice, which requires a solid theoretical foundation and rich practical experience, and is greatly affected by human subjectivity. Some people have also proposed a non-destructive detection instrument for water accumulation inside box (plate) girders based on virtual instrument technology, but this method cannot detect plate girders without gaps. Summary of the invention
[0004] In view of the above problems, the purpose of the present invention is to provide a system and method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound. The system collects signals for detecting water accumulation inside a steel box girder by carrying a laser ultrasonic probe on an unmanned aerial vehicle. Non-contact measurement is adopted, and non-destructive testing can be performed under adverse conditions such as high temperature, high pressure, and strong winds. The detection efficiency is high. At the same time, the gray value area with obvious gradient changes is extracted through Mask R-CNN, which greatly reduces the number of parameters without affecting feature recognition, thereby improving the recognition speed and accuracy of water accumulation inside the steel box girder.
[0005] The technical solution of the present invention is: a water accumulation detection system inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound, comprising a unmanned aerial vehicle, a regulator is provided at the central position of the top of the unmanned aerial vehicle, a laser ultrasonic probe, a high-definition camera and an optical detector are provided on the regulator, rotor rods are provided around the sides of the unmanned aerial vehicle, rotors are provided on the rotor rods, positioning suction cups are provided on the rotors, support brackets are provided on both sides of the lower part of the unmanned aerial vehicle, a storage platform is provided between the support brackets, and a power supply and a multi-functional integrated component are provided on the storage platform.
[0006] The multifunctional integrated components include a 5G network connector, an anti-interference device and a GPS locator.
[0007] The method for detecting water accumulation inside a steel box girder based on a drone and laser ultrasound uses the above-mentioned system for detecting water accumulation inside a steel box girder based on a drone and laser ultrasound, and includes the following steps: S1: A drone equipped with a laser ultrasonic probe is used to fly along the bottom of the steel box girder to be inspected. There are multiple target stop points at the bottom of the steel box girder to be inspected. The drone is equipped with a laser ultrasonic probe to perform inspection at the target stop points. The laser ultrasonic probe is used to inject a laser beam, and the reflected light signal is received through an optical detector. The reflected light signal is transmitted to the computer through the 5G network connector in the multifunctional integrated component for real-time signal processing; S2: The maximum amplitude image made from the signal is preprocessed, and the gray value areas with obvious gradient changes in different directions are extracted from the maximum amplitude image made from the signal using the Mask R-CNN algorithm; S3: By comparing the maximum amplitude image of the bottom of the steel box girder without internal water accumulation, the collected maximum amplitude image is detected and evaluated, and the internal water accumulation is identified by using the reflection law of ultrasonic waves at different interfaces. When water accumulation is detected inside the cavity, the GPS locator in the multifunctional integrated component transmits the position information to the computer terminal to locate the water accumulation inside the cavity; S4: Repeat the above process and use UAV 1 to go to the next target stop point until all target stop points are detected.
[0008] In step S1, the interval between target docking points is 3m, and during the detection process, the docking position of the drone is determined by a high-definition camera.
[0009] In step S2, the specific process of image preprocessing is: first, the three-channel maximum amplitude image is converted into a single-channel grayscale image by image grayscale, and then the grayscale image is unified in resolution. The image grayscale adopts one of the component method, maximum value method, average method, and weighted average method.
[0010] The specific formula for image grayscale using the component method is: (1) In the formula R(i, j) Represents the red channel in (i, j) The gray value at g 1 (i, j) , G(i, j) Represents the green channel in (i, j) The gray value at g 2 (i, j) , B(i, j) Represents the blue channel in (i, j) The gray value at g 3 (i, j) ; The specific formula for image grayscale using the maximum value method is: (2) In the formulaR(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) is the grayscale value of the grayscale image, and its value is the maximum value of the brightness of the three channels; The specific formula for image grayscale using the average method is: (3) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) is the grayscale value of the grayscale image, and its value is the average value of the brightness of the three channels; The specific formula for image grayscale using the weighted average method is: (4) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) It is the grayscale value of the grayscale image, and its value is the weighted average of the brightness of the three channels. The weighting coefficient is determined by the sensitivity of the human eye to the three colors.
[0011] In step S2, the specific process of extracting the gray value area with obvious gradient change from the maximum amplitude image based on Mask R-CNN is: using edge detection methods in different directions to extract gradient change features: Horizontal Gradient (5) Vertical Gradient (6) Slope gradient (7) In the formula, is the gradient in the horizontal direction; is the gradient in the horizontal direction; is the gradient in the tilt direction.
[0012] In step S3, the internal water accumulation is identified by utilizing the reflection law of ultrasonic waves at different interfaces. The specific process is as follows: because there is only air and water inside the steel box girder, the maximum amplitude of the ultrasonic echo in the air will be significantly larger than the maximum amplitude of the ultrasonic echo in the water, thereby determining whether there is water accumulation inside the cavity of the steel box girder.
[0013] Relevant terms involved in this invention: (1) Laser ultrasound: Use laser pulses to excite ultrasound in the workpiece being inspected, and use laser beams to detect the propagation of ultrasound to obtain workpiece information, such as workpiece thickness, internal and surface defects, material parameters, etc. (2) Optical detector: used to irradiate the sample surface with continuous detection laser, receive the reflected light from the surface, and derive the laser ultrasonic signal from the changes in the phase, amplitude, frequency, etc. of the reflected light; (3) Mask R-CNN: Mask R-CNN is an instance segmentation algorithm.
[0014] The technical effects of the present invention are as follows: 1. The present invention performs preprocessing of the maximum amplitude image by graying the image, converts the three-channel maximum amplitude image into a single-channel grayscale image by graying, and then unifies the resolution of the grayscale image, effectively simplifying the experiment and improving the detection efficiency; 2. The present invention uses an unmanned aerial vehicle equipped with a laser ultrasonic probe to collect signals for water accumulation detection inside a steel box girder. Compared with the existing geological radar detection method, the detection efficiency is high, and non-contact measurement is adopted, and non-destructive detection can be performed under adverse conditions such as high temperature, high pressure, and strong wind; 3. The present invention uses Mask R-CNN to extract the features of the maximum amplitude image sequence. By extracting the gray value area with obvious gradient changes through Mask R-CNN, the amount of parameters is greatly reduced without affecting feature recognition, the recognition speed is improved, and the accuracy of waveform analysis is effectively improved.
[0015] The following is a further description with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 The present invention is a schematic structural diagram of a system for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound according to an embodiment of the present invention.
[0017] Figure 2 The present invention is a flow chart of water accumulation detection inside a steel box girder according to an embodiment of the present invention.
[0018] Figure 3 The figure is a schematic diagram of the flight route of the UAV according to an embodiment of the present invention.
[0019] Figure 4 Schematic diagram of laser excitation and reflected light collection according to an embodiment of the present invention.
[0020] Figure 5 Schematic diagram of the structure of the regulator according to an embodiment of the present invention.
[0021] Figure numerals: 1-UAV; 2-laser ultrasonic probe; 3-high-definition camera; 4-support bracket; 5-adjuster; 6-positioning suction cup; 7-storage platform; 8-rotor; 9-power supply; 10-multi-function integrated component; 11-optical detector. DETAILED DESCRIPTION Example 1
[0022] like Figure 1 , Figure 5 As shown, a system for detecting water accumulation inside a steel box girder based on a UAV and laser ultrasound comprises a UAV 1, a regulator 5 is provided at a central position on the top of the UAV 1, a laser ultrasonic probe 2, a high-definition camera 3 and an optical detector 11 are provided on the regulator 5, rotor rods are provided around the sides of the UAV 1, rotors 8 are provided on the rotor rods, positioning suction cups 6 are provided on the rotors 8, supporting brackets 4 are provided on both sides of the lower part of the UAV 1, storage platforms 7 are provided between the supporting brackets 4, and a power supply 9 and a multifunctional integrated component 10 are provided on the storage platform 7.
[0023] The multifunctional integrated component 10 includes a 5G network connector, an anti-interference device and a GPS locator.
[0024] like Figure 2 As shown, the method for detecting water accumulation inside a steel box girder based on a drone and laser ultrasound uses the above-mentioned system for detecting water accumulation inside a steel box girder based on a drone and laser ultrasound, and includes the following steps: S1: Plan the flight route of the drone, and use the drone 1 to carry the laser ultrasonic probe 2 to fly along the bottom of the steel box girder to be inspected, such as Figure 3 As shown in FIG. 1 , a plurality of target stop points are set at the bottom of the steel box girder to be inspected, and the UAV 1 is equipped with a laser ultrasonic probe 2 to perform inspection at the target stop points, such as Figure 4 As shown, a laser ultrasound probe 2 is used to incident a laser beam, and a reflected light signal is received by an optical detector 11, and the reflected light signal is transmitted to a computer through a 5G network connector in a multifunctional integrated component 10 for real-time signal processing; S2: The maximum amplitude image made from the signal is preprocessed, and the gray value areas with obvious gradient changes in different directions are extracted from the maximum amplitude image made from the signal using the Mask R-CNN algorithm; S3: By comparing the maximum amplitude image of the bottom of the steel box girder without internal water accumulation, the collected maximum amplitude image is detected and evaluated, and the internal water accumulation is identified by using the reflection law of ultrasonic waves at different interfaces. When water accumulation is detected inside the cavity, the GPS locator in the multifunctional integrated component 10 transmits the position information to the computer terminal to locate the water accumulation inside the cavity; S4: Repeat the above process and use UAV 1 to go to the next target stop point until all target stop points are detected.
[0025] In step S1, the interval between target docking points is 3m, and during the detection process, the docking position of the drone 1 is determined by the high-definition camera 3.
[0026] In step S2, the specific process of image preprocessing is: first, the three-channel maximum amplitude image is converted into a single-channel grayscale image by image grayscale, and then the grayscale image is unified in resolution. The image grayscale adopts one of the component method, maximum value method, average method, and weighted average method.
[0027] The specific formula for image grayscale using the component method is: (1) In the formula R(i, j) Represents the red channel in (i, j) The gray value at g 1 (i, j) , G(i, j) Represents the green channel in (i, j) The gray value at g 2 (i, j) , B(i, j) Represents the blue channel in (i, j) The gray value at g 3 (i, j) ; The specific formula for image grayscale using the maximum value method is: (2) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) is the grayscale value of the grayscale image, and its value is the maximum value of the brightness of the three channels; The specific formula for image grayscale using the average method is: (3) In the formula R(i, j) Represents the red channel in(i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) is the grayscale value of the grayscale image, and its value is the average value of the brightness of the three channels; The specific formula for image grayscale using the weighted average method is: (4) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) It is the grayscale value of the grayscale image, and its value is the weighted average of the brightness of the three channels. The weighting coefficient is determined by the sensitivity of the human eye to the three colors.
[0028] In step S2, the specific process of extracting the gray value area with obvious gradient change from the maximum amplitude image based on Mask R-CNN is: using edge detection methods in different directions to extract gradient change features: Horizontal Gradient (5) Vertical Gradient (6) Slope gradient (7) In the formula, is the gradient in the horizontal direction; is the gradient in the horizontal direction; is the gradient in the tilt direction.
[0029] In step S3, the internal water accumulation is identified by utilizing the reflection law of ultrasonic waves at different interfaces. The specific process is as follows: because there is only air and water inside the steel box girder, the maximum amplitude of the ultrasonic echo in the air will be significantly larger than the maximum amplitude of the ultrasonic echo in the water, thereby determining whether there is water accumulation inside the cavity of the steel box girder.
[0030] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.
Claims
1. The water accumulation detection system inside the steel box girder based on drone and laser ultrasound is characterized by: The invention comprises an unmanned aerial vehicle (1), wherein an adjuster (5) is provided at the center of the top of the unmanned aerial vehicle (1), and a laser ultrasonic probe (2), a high-definition camera (3) and an optical detector (11) are provided on the adjuster (5); rotor rods are provided around the sides of the unmanned aerial vehicle (1), and rotors (8) are provided on the rotor rods; and positioning suction cups (6) are provided on the rotors (8); and support brackets (4) are provided on both sides of the lower part of the unmanned aerial vehicle (1); a storage platform (7) is provided between the support brackets (4); and a power supply (9) and a multifunctional integrated component (10) are provided on the storage platform (7).
2. The steel box girder internal water accumulation detection system based on drone and laser ultrasound according to claim 1 is characterized in that: The multifunctional integrated component (10) comprises a 5G network connector, an anti-interference device and a GPS locator.
3. A method for detecting water accumulation inside a steel box girder based on drones and laser ultrasound, using the system for detecting water accumulation inside a steel box girder based on drones and laser ultrasound as claimed in claim 1, characterized in that: The following steps are involved: S1: Using a drone (1) equipped with a laser ultrasonic probe (2) to fly along the bottom of a steel box girder to be inspected, the bottom of the steel box girder to be inspected is provided with a plurality of target stop points, the drone (1) equipped with a laser ultrasonic probe (2) performs inspection at the target stop points, the laser ultrasonic probe (2) is used to project a laser beam, and a reflected light signal is received through an optical detector (11), and the reflected light signal is transmitted to a computer through a 5G network connector in a multifunctional integrated component (10) for real-time signal processing; S2: The maximum amplitude image made from the signal is preprocessed, and the gray value areas with obvious gradient changes in different directions are extracted from the maximum amplitude image made from the signal using the Mask R-CNN algorithm; S3: By comparing the maximum amplitude image of the bottom of the steel box girder without internal water accumulation, the collected maximum amplitude image is detected and evaluated, and the internal water accumulation is identified by using the reflection law of ultrasonic waves at different interfaces. When water accumulation is detected inside the cavity, the position information is transmitted to the computer terminal through the GPS locator in the multifunctional integrated component (10), so as to locate the water accumulation inside the cavity; S4: Repeat the above process and use UAV 1 to go to the next target stop point until all target stop points are detected.
4. The method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound according to claim 3 is characterized in that: In step S1, the distance between target docking points is 3 m, and during the detection process, the docking position of the drone (1) is determined by a high-definition camera (3).
5. The method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound according to claim 3 is characterized in that: In step S2, the specific process of image preprocessing is: first, the three-channel maximum amplitude image is converted into a single-channel grayscale image by image grayscale, and then the grayscale image is unified in resolution. The image grayscale adopts one of the component method, maximum value method, average method, and weighted average method.
6. The method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound according to claim 5 is characterized in that: The specific formula for image grayscale using the component method is: (1) In the formula R(i, j) Represents the red channel in (i, j) The gray value at g 1 (i, j) , G(i, j) Represents the green channel in (i, j) The gray value at g 2 (i, j) , B(i, j) Represents the blue channel in (i, j) The gray value at g 3 (i, j) ; The specific formula for image grayscale using the maximum value method is: (2) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) is the grayscale value of the grayscale image, and its value is the maximum value of the brightness of the three channels; The specific formula for image grayscale using the average method is: (3) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) is the grayscale value of the grayscale image, and its value is the average value of the brightness of the three channels; The specific formula for image grayscale using the weighted average method is: (4) In the formula R(i, j) Represents the red channel in (i, j) The gray value at G(i, j) Represents the green channel in (i, j) The gray value at B(i, j) Represents the blue channel in (i, j) The gray value at g(i, j) It is the grayscale value of the grayscale image, and its value is the weighted average of the brightness of the three channels. The weighting coefficient is determined by the sensitivity of the human eye to the three colors.
7. The method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound according to claim 3 is characterized in that: In step S2, the specific process of extracting the gray value area with obvious gradient change from the maximum amplitude image based on Mask R-CNN is: using edge detection methods in different directions to extract gradient change features: Horizontal Gradient (5) Vertical Gradient (6) Slope gradient (7) In the formula, is the gradient in the horizontal direction; is the gradient in the horizontal direction; is the gradient in the tilt direction.
8. The method for detecting water accumulation inside a steel box girder based on an unmanned aerial vehicle and laser ultrasound according to claim 3 is characterized in that: In step S3, the internal water accumulation is identified by utilizing the reflection law of ultrasonic waves at different interfaces. The specific process is as follows: because there is only air and water inside the steel box girder, the maximum amplitude of the ultrasonic echo in the air will be significantly larger than the maximum amplitude of the ultrasonic echo in the water, thereby determining whether there is water accumulation inside the cavity of the steel box girder.