WO2024040945A1 - 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备 - Google Patents

一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备 Download PDF

Info

Publication number
WO2024040945A1
WO2024040945A1 PCT/CN2023/082049 CN2023082049W WO2024040945A1 WO 2024040945 A1 WO2024040945 A1 WO 2024040945A1 CN 2023082049 W CN2023082049 W CN 2023082049W WO 2024040945 A1 WO2024040945 A1 WO 2024040945A1
Authority
WO
WIPO (PCT)
Prior art keywords
corrosion
flood discharge
rod
image
automatic inspection
Prior art date
Application number
PCT/CN2023/082049
Other languages
English (en)
French (fr)
Inventor
庞博慧
肖海斌
曹学兴
陈豪
孟买
陈浩东
Original Assignee
华能澜沧江水电股份有限公司
天津大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华能澜沧江水电股份有限公司, 天津大学 filed Critical 华能澜沧江水电股份有限公司
Priority to US18/329,570 priority Critical patent/US20240183743A1/en
Publication of WO2024040945A1 publication Critical patent/WO2024040945A1/zh

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
    • G01M5/0033—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings by determining damage, crack or wear
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64C—AEROPLANES; HELICOPTERS
    • B64C39/00—Aircraft not otherwise provided for
    • B64C39/02—Aircraft not otherwise provided for characterised by special use
    • B64C39/024—Aircraft not otherwise provided for characterised by special use of the remote controlled vehicle type, i.e. RPV
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64D—EQUIPMENT FOR FITTING IN OR TO AIRCRAFT; FLIGHT SUITS; PARACHUTES; ARRANGEMENT OR MOUNTING OF POWER PLANTS OR PROPULSION TRANSMISSIONS IN AIRCRAFT
    • B64D47/00—Equipment not otherwise provided for
    • B64D47/08—Arrangements of cameras
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64U—UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U50/00—Propulsion; Power supply
    • B64U50/10—Propulsion
    • B64U50/19—Propulsion using electrically powered motors
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
    • G01M5/0075—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings by means of external apparatus, e.g. test benches or portable test systems
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M5/00—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings
    • G01M5/0091—Investigating the elasticity of structures, e.g. deflection of bridges or air-craft wings by using electromagnetic excitation or detection
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/776—Validation; Performance evaluation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64U—UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U2101/00—UAVs specially adapted for particular uses or applications
    • B64U2101/25—UAVs specially adapted for particular uses or applications for manufacturing or servicing
    • B64U2101/26—UAVs specially adapted for particular uses or applications for manufacturing or servicing for manufacturing, inspections or repairs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30181—Earth observation
    • G06T2207/30184—Infrastructure
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping

Definitions

  • the invention belongs to the technical field of inspection and defect detection, and specifically relates to an automatic inspection and intelligent detection method and equipment for corrosion defects of a flood discharge tunnel.
  • the purpose of the present invention is to solve the deficiencies of the existing technology and provide a method and equipment for automatic inspection of flood discharge tunnels and intelligent detection of corrosion defects.
  • the invention can automatically inspect the situation of the flood discharge tunnel, and can identify the type, location, shape, and depth of corrosion defects in an all-round, rapid and high-precision manner, intelligently analyze and judge the degree of corrosion impact, and provide early warning when the corrosion defects exceed the threshold. And it can automatically and quickly give warnings and reminders to avoid dangers to downstream life safety during flood discharges. It is more intelligent, has stronger accident prevention capabilities, greatly improves inspection efficiency and detection efficiency, greatly reduces labor costs, and at the same time avoids personnel A safety accident occurred when entering the flood discharge tunnel to carry out high-risk inspection and measurement operations.
  • An automatic inspection and corrosion defect intelligent detection equipment for flood discharge tunnels including an inspection device body, a protective shell is provided at the top of the inspection device body, and an adjustment mechanism is provided inside the protective shell;
  • the adjustment mechanism includes a connecting pivot block arranged inside the protective shell.
  • One side of the connecting pivot block is connected to a transmission shaft.
  • the other end of the transmission shaft is provided with a servo drive motor for driving.
  • the transmission shaft There are two symmetrically arranged connecting collar rods on the outer sleeve of the shaft rod.
  • a fixed rod is provided inside the connecting collar rod near the other end.
  • a connecting collar block is connected to one side of the middle part of the fixed rod.
  • the connecting sleeve A transmission platform is connected to one side of the ring block.
  • a first drive motor is installed on the transmission platform.
  • a drive rod is connected to the output end of the first drive motor.
  • a holographic three-dimensional scanner is provided on the top of the drive rod.
  • the holographic three-dimensional scanner is installed on the drive platform.
  • a second drive motor is provided inside the scanner.
  • the output end of the second drive motor is connected to a rotating rod, a rear connecting shaft block is sleeved on the outside of the rotating rod, and a measurement camera is fixedly connected to one side of the connecting shaft block.
  • a linkage switching mechanism is set outside one end of the rotating rod.
  • the linkage switching mechanism includes a driving collar rod arranged outside one end of the rotating rod.
  • the driving collar rod is movably connected to the inside near the top.
  • a switching motor is installed on one end of the adjusting rotating rod, the output end of the switching motor is fixedly connected to the adjusting rotating rod, and an adjusting collar is provided outside the adjusting rotating rod and on the top side of the driving collar rod.
  • a first illuminating light is installed on one side of the adjusting collar block
  • a second illuminating light is installed on the other side of the adjusting collar block.
  • a limiting screw block is installed on one side of the driving collar rod.
  • four support frames are symmetrically arranged on the outside of the inspection device body, and a lifting motor is provided near the end of the support frame.
  • the output end of the lifting motor is connected to a driving fan blade, and the driving fan blade
  • a level sensor is installed on one side, a wireless transmitter is provided inside the body of the inspection device, an arc-shaped bracket is provided below two adjacent support frames, a buffer base is provided below the arc-shaped bracket, and the buffer base is The bottom end is connected with a raised support.
  • the output end of the wireless transmitter is connected to a wireless receiver, a processing device is installed at the bottom of the wireless receiver, an audible and visual alarm is installed on one side of the wireless receiver, and the processing device A display screen is embedded on the side, an image storage is provided below the display screen, and an image processor is provided on one side of the image storage.
  • a first battery is provided below the body of the inspection device, and two leveling guide plates are symmetrically provided below the first battery, and guide grooves are provided on opposite sides of the two leveling guide plates.
  • a cross guide block is slidably provided inside the guide groove. One end of the cross guide block is connected to a push rod. One end of the push rod is equipped with an electric cylinder.
  • a second battery is provided between the two cross guide blocks. The bottom end of the second battery is connected to a support plate, and one end of the second battery is provided with a connection port.
  • the invention also provides a method for automatic inspection of flood discharge tunnels and intelligent detection of corrosion defects, using the above-mentioned automatic inspection of flood discharge tunnels and intelligent detection of corrosion defects, including the following steps:
  • Step 1 Establish a database of corrosion defects of flood discharge tunnels: collect data on the operation of flood discharge tunnels at home and abroad, and establish a database of corrosion defects of flood discharge tunnels, including the types of corrosion, damage locations, scale of corrosion, The impact of corrosion damage; then the flood discharge tunnel corrosion defect data in the flood discharge tunnel corrosion defect database is divided into 3 levels.
  • the first level has no impact on flood discharge safety; the second level has a small impact, and needs to be reminded and strengthened observation; Level 3 has an impact, requiring repair measures and early warning prompts;
  • Step 2 Establish analysis, evaluation standards and early warning thresholds for flood discharge tunnel corrosion defects: Based on the impact of erosion, abrasion, and cavitation erosion on flood discharge safety, establish defect type evaluation criteria: No corrosion found on the tunnel surface is ranked No. 1 Level, the corrosion depth less than 3mm is classified as the second level, and the corrosion depth greater than or equal to 3mm is listed as the third level; the early warning threshold is the corrosion depth greater than or equal to 3mm.
  • Step 3 Automatic inspection: Use automatic inspection of the flood discharge tunnel and automatic inspection of corrosion defect intelligent detection equipment. Set the inspection route before takeoff, use laser scanning to inspect the entire tunnel, and perform lighting operations. The collected data Upload images wirelessly;
  • Step 4 Image analysis and processing and intelligent recognition model training: Establish an image database and perform two processes on each image that needs to be trained. One is saved as the original image, and the other is affixed with the corresponding real label. At the same time, the same type of label is Put the pictures into the same type of picture set, build an image recognition neural network, input the original training image into the neural network for image recognition processing, get the output label type, compare it with the real label of the picture, and compare the results Calculate the loss, adjust the parameters in the neural network through the loss calculation, update the neural network model, perform image recognition on the picture again, and finally obtain a more complete image recognition neural network; then use the image recognition neural network to identify corrosion defects. And conduct corresponding evaluation and early warning based on the evaluation standards and early warning thresholds established in step 2.
  • the recognition information includes the type of erosion, location, shape, and depth.
  • the first battery and the second battery can be connected in series by using a wire harness.
  • the battery connection method can be a normal connection method, just to increase the weight to achieve a balanced operation, and the first battery and the second battery are integrated.
  • the equipment provides power, and the connection relationship between the equipment and the electrical structural components can be in the existing manner, and the present invention does not impose too many restrictions.
  • the invention can automatically inspect the situation of the flood discharge tunnel, and can identify the type, location, shape, and depth of corrosion defects in an all-round, rapid and high-precision manner, intelligently analyze and judge the degree of corrosion impact, and provide early warning when the corrosion defects exceed the threshold. And it can automatically and quickly give warnings and reminders to avoid dangers to downstream life safety during flood discharges. It is more intelligent, has stronger accident prevention capabilities, greatly improves inspection efficiency and detection efficiency, greatly reduces labor costs, and at the same time avoids personnel A safety accident occurred when entering a flood discharge tunnel to carry out high-risk inspection and measurement operations;
  • the invention gradually improves the computer image recognition efficiency through training, and performs loss calculation on the image processing results, which can improve the neural network, optimize and update the neural network structure, further improve the accuracy of image recognition, and provide a basis for identifying defect types. , analyze and judge the impact of defects and other assessments to provide technical support;
  • the invention uses an adjustment mechanism to make the rotating rod drive the connecting shaft block to rotate.
  • the connecting shaft block drives the measuring camera to rotate.
  • the measuring camera starts to reciprocate left and right and up and down angles for side illumination.
  • the transmission platform drives the holographic three-dimensional scanner to rotate the measuring camera.
  • Flip upward and clockwise 90 degrees, and the measurement camera begins to scan the top surface of the flood discharge tunnel at a full angle. This can effectively improve the scanning of the side dead corners and the dead corners at the top of the flood discharge tunnel. It not only improves the scanning efficiency, but also improves the scanning speed.
  • the comprehensiveness optimizes the scanning effect and avoids dead spots;
  • the invention adopts a linkage switching mechanism to make the rotating rod drive the driving collar rod to rotate, and the adjusting rotating rod drives the adjusting collar block to deflect downward.
  • the measuring camera can be followed for the light illumination operation, ensuring the accuracy of the light illumination. It provides sufficient light during the scanning process, and can switch to the second illumination light when the first illumination light is damaged, thereby forming a backup light switching operation, ensuring that the light can be used normally, and preventing the drone from returning due to lighting problems, which is effective Improved inspection efficiency;
  • the present invention adopts the process of flipping the measuring camera upward ninety degrees clockwise, and detects the horizontal angle through the horizontal sensor.
  • the electric cylinder drives the support rod to move to the left
  • the second battery drives the support plate to move to the left. In this way, it can be used as a measuring camera.
  • the center of gravity shifts backward
  • the second battery can move to the left to ensure the balance of the center of gravity, which can improve the horizontal stability of the drone during use. The stability is better and the normal operation of the drone is ensured.
  • the efficiency of inspection can be effectively improved, the efficiency and accuracy of intelligent analysis of corrosion defects can be improved, the safety of inspection work can be improved, and safety issues of flood discharge tunnels can be promptly reminded to ensure the safety of flood discharge.
  • Figure 1 is a schematic diagram of the wireless transportation principle of the drone of the present invention.
  • Figure 2 is a schematic diagram of the overall structure of the inspection device body of the present invention.
  • Figure 3 is a schematic structural diagram of the wireless transmitter of the present invention.
  • Figure 4 is a schematic structural diagram of the connection between the transmission shaft and the servo drive motor of the present invention.
  • Figure 5 is a schematic diagram of the disassembled structure of the driving rod and the holographic three-dimensional scanner of the present invention.
  • Figure 6 is a schematic structural diagram of the lower part of the inspection device body of the present invention.
  • Figure 7 is a schematic structural diagram of the connection between the cross guide block and the second battery of the present invention.
  • Figure 8 is an enlarged structural schematic diagram of position A in Figure 5 of the present invention.
  • Figure 9 shows the image recognition training process and its internal neural network structure of the invention.
  • plural means two or more.
  • the orientation or state relationship indicated by the terms “inner”, “upper”, “lower”, etc. is based on the orientation or state relationship shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply what is meant. Devices or elements must be oriented, constructed and operate in a particular orientation and therefore are not to be construed as limitations of the invention.
  • connection should be understood in a broad sense.
  • it can be a fixed connection or a removable connection.
  • Detachable connection, or integral connection it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium.
  • connection can be a fixed connection or a removable connection.
  • Detachable connection, or integral connection it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium.
  • An automatic inspection and defect detection equipment for flood discharge tunnels including an inspection device body 1.
  • a protective shell 2 is provided at the top of the inspection device body 1, and an adjustment mechanism is provided inside the protective shell 2;
  • the adjustment mechanism includes a connecting pivot block 3 arranged inside the protective shell 2.
  • One side of the connecting pivot block 3 is connected to a transmission shaft 4.
  • the other end of the transmission shaft 4 is provided with a servo drive motor 5 for driving.
  • the outer sleeve is provided with two symmetrically arranged connecting collar rods 6.
  • a fixed rod 7 is provided inside the connecting collar rod 6 near the other end.
  • a connecting collar block 8 is connected to one side of the middle part of the fixed rod 7.
  • the connecting collar block 8 is connected to a transmission platform 9 on one side.
  • a first driving motor 10 is installed on the transmission platform 9.
  • a driving rod 11 is connected to the output end of the first driving motor 10.
  • a holographic three-dimensional scanner 12 is provided on the top of the driving rod 11.
  • the holographic three-dimensional scanner 12 is provided with a second drive motor 13 inside.
  • the output end of the second drive motor 13 is connected to a rotating rod 14.
  • the rotating rod 14 is sleeved on the outside and then connected to the shaft block 15.
  • One side of the connecting shaft block 15 is connected to a measurement camera 16.
  • the connecting shaft block 15 It is fixedly connected to the measuring camera 16 so that the second drive motor 13 drives the rotating rod 14 to rotate counterclockwise, and the rotating rod 14 drives the connecting shaft block 15 to rotate to realize the angle adjustment of the measuring camera 16 in the vertical direction.
  • the upper and lower ends of the drive rod 11 are fixedly connected to the first drive motor 10 and the holographic three-dimensional scanner 12 respectively, and the transmission shaft 4 and the servo drive motor 5 are fixedly connected so that , the first drive motor 10 can drive the drive rod 11 to drive the holographic three-dimensional scanner 12 for lateral angle deflection, and the servo drive motor 5 can drive the transmission shaft 4 for drive.
  • one end of the rotating rod 14 is equipped with a linkage switching mechanism.
  • the linkage switching mechanism includes a driving collar rod 17 provided on the outside of one end of the rotating rod 14.
  • the driving collar rod 17 is movably connected with an adjustment knob near the top.
  • Rod 18, and a switching motor 19 is installed at one end of the adjusting rotating rod 18.
  • the output end of the switching motor 19 is fixedly connected to the adjusting rotating rod 18.
  • An adjusting collar block is provided outside the adjusting rotating rod 18 and at the top side of the driving collar rod 17. 20.
  • the first illuminating lamp 21 is installed on one side of the adjusting collar block 20, and the second illuminating lamp 22 is installed on the other side of the adjusting collar block 20, so that the rotating rod 14 drives the driving collar rod 17 to rotate, and the adjusting collar block 20 is deflected at a downward angle.
  • the switching motor 19 can be started to rotate, and the second illumination lamp 22 and the first illumination lamp 21 can be switched to facilitate the later startup of the backup light illumination.
  • the two sides of the adjusting collar block 20 are detachably connected to the second illuminating lamp 22 and the first illuminating lamp 21 through bolts, and the driving collar rod 17 is connected to the adjusting rotating rod 18 and the rotating rod. 14 is fixedly connected, and a limiting screw block 23 is installed on one side of the driving collar rod 17 so that the limiting screw block 23 can limit the position of the driving collar rod 17 .
  • a lifting motor 25 is provided near the end of the support frame 24.
  • the output end of the lifting motor 25 is connected to a driving fan blade 26.
  • the driving fan A level sensor 46 is installed on one side of the leaf 26, a wireless transmitter 27 is provided inside the inspection device body 1, an arc bracket 29 is provided below the two adjacent support frames 24, and a buffer base 30 is provided below the arc bracket 29.
  • a raised support 31 is connected to the bottom of the base 30 so that the horizontal sensor 46 can detect the inclination angle of the support frame 24.
  • the lift motor 25 drives the drive blade 26 to rotate, thus generating lift for flight.
  • the raised support 31 is connected to the lower end of the base 30.
  • the support 31 plays a buffering role on the buffering base 30
  • the buffering base 30 plays a buffering role on the two arc-shaped brackets 29 .
  • a first battery 28 is provided below the inspection device body 1.
  • Two leveling guide plates 32 are symmetrically provided below the first battery 28.
  • the two leveling guide plates 32 are provided on opposite sides.
  • One end of the cross guide block 34 is connected to a push rod 35, and one end of the push rod 35 is installed with an electric cylinder 36.
  • the second battery 37 has a support plate 38 connected to the bottom end of the second battery 37.
  • One end of the second battery 37 is provided with a connection port 39 so as to realize the leveling operation and ensure the level of the entire drone during the flipping process of the measurement camera 16. state.
  • the output end of the wireless transmitter 27 is connected to a wireless receiver 40, a processing device 41 is installed at the bottom of the wireless receiver 40, an audible and visual alarm 42 is installed on one side of the wireless receiver 40, and an audible and visual alarm 42 is installed on one side of the wireless receiver 40.
  • a display screen 45 is embedded and installed.
  • An image storage 43 is provided below the display screen 45.
  • An image processor 44 is provided on one side of the image storage 43 so that the wireless receiver 40 can receive the incoming image into the image storage 43 for image comparison and recognition. This allows alarm operations to be performed based on the warning level.
  • a flood discharge tunnel automatic inspection and corrosion defect intelligent detection method using the flood discharge tunnel automatic inspection and corrosion defect intelligent detection equipment described in any one of claims 1 to 7, characterized in that it includes the following steps:
  • Step 1 Establish a database of corrosion defects of flood discharge tunnels: collect data on the operation of flood discharge tunnels at home and abroad, and establish a database of corrosion defects of flood discharge tunnels, including the types of corrosion, damage locations, scale of corrosion, The impact of corrosion damage; then the flood discharge tunnel corrosion defect data in the flood discharge tunnel corrosion defect database is divided into 3 levels.
  • the first level has no impact on flood discharge safety; the second level has a small impact, and needs to be reminded and strengthened observation; Level 3 has an impact, requiring repair measures and early warning prompts;
  • Step 2 Establish an analysis, evaluation standard and early warning threshold index system for flood discharge tunnel corrosion defects: Based on the impact of erosion, abrasion, and cavitation erosion on flood discharge safety, establish a defect type evaluation standard: no corrosion found on the tunnel surface is classified as For the first level, the corrosion depth less than 3mm is classified as the second level, and the corrosion depth greater than or equal to 3mm is classified as the third level; the early warning threshold is the corrosion depth greater than or equal to 3mm.
  • Step 3 Automatic inspection: Use automatic inspection of the flood discharge tunnel and intelligent detection equipment for corrosion defects or automatic inspection of drones. Set the inspection route before takeoff, and use laser scanning to achieve high-precision inspection of the entire tunnel without blind spots. Inspect and perform lighting operations, and the collected images are uploaded wirelessly;
  • Step 4 Image analysis and processing and intelligent recognition model training: Establish an image database and perform two processes on each image that needs to be trained. One is saved as the original image, and the other is affixed with the corresponding real label. At the same time, the same type of label is Put the pictures into the same type of picture set, build an image recognition neural network, input the original training image into the neural network for image recognition processing, get the output label type, compare it with the real label of the picture, and compare the results Calculate the loss, adjust the parameters in the neural network through the loss calculation, update the neural network model, perform image recognition on the picture again, and finally obtain a more complete image recognition neural network; then use the image recognition neural network to identify corrosion defects. And conduct corresponding evaluation and early warning based on the evaluation standards and early warning thresholds established in step 2.
  • the recognition information includes the type of erosion, location, shape, and depth.
  • the computer's ability to recognize images can be gradually improved, and the efficiency and accuracy of recognition can be improved.
  • the neural network model can be upgraded, inspecting once before the flood discharge, and inspecting again after experiencing a flood discharge, and comparing it with the evaluation results before the flood discharge.
  • the original Pictures of various damage types in the database strengthen computer in-depth learning of image recognition during inspections and flood discharges, improve the neural network framework for updated calculations, improve the accuracy of subsequent prediction results, and continue to update data level evaluation. Inspection requires two or more people to conduct inspection at the same time, one person to inspect and one person to keep records, and the number of reciprocal inspections should be no less than five times.
  • training requires a large picture database.
  • the main source of the database is the long-term accumulation of pictures in the project.
  • the algorithm model and corresponding functions involved in the neural network in the training are not unique. Different algorithm models and functions can be replaced. Different algorithm models and functions can be replaced.
  • the image recognition accuracy of algorithm models is different, but through repeated deep learning and continuous expansion of existing databases, the accuracy of image recognition can be improved.
  • the computer first converts the transmitted image into data and loads it into the established neural network for image recognition.
  • the convolutional neural network is most commonly used for image recognition.
  • the present invention uses a convolutional neural network in its design.
  • the neural network structure includes an input layer, a hidden layer and an output layer.
  • the input layer takes the image converted data as input.
  • the job of the hidden layer is to convert the input into something that can be used by the output layer.
  • the output layer obtains the probability that the image is a specific category.
  • the hidden layer of the convolutional neural network includes convolution layer, activation layer, pooling layer and fully connected layer.
  • the convolutional layer functions like a filter, searching for each part and pixel in the image, extracting key features of the image, and finding the recognition pattern.
  • the generation process of the convolutional layer Hi can be described as , where Wi represents the weight vector of the i-th layer convolution kernel, the operation symbol Represents the convolution operation between the convolution kernel and the i-1th layer image or feature map, and bi is the selected offset vector; since the convolution operation is linear, in order to increase the complexity of the model, an activation function is often added through the activation layer. Increase the nonlinearity of the model to simulate any function and enhance the expressive ability of the network.
  • the activation function often used in convolutional neural networks is the ReLU function.
  • the pooling layer process is divided into maximum pooling and average pooling.
  • the pooling process is to convert the input image The image is evenly divided into several rectangular areas.
  • the maximum pooling is to take the maximum value of the pixels in the pooling area.
  • the average pooling is to average the images in the pooling area.
  • the function of the fully connected layer is to combine multiple pixels in the pooling area.
  • the features in the image feature maps of the convolutional layer and the pooling layer are integrated to obtain the high-level meaning of the image features, which are then used for image classification. After obtaining the output result, the loss is calculated.
  • the value calculated by the loss function is used to express the quantitative expression of the gap between the probability model in the neural network and the ideal probability model.
  • the mean square error function is used. ,in represents the output of the neural network, represents the true value of the supervised data, Represents the dimensions of the data.
  • Back propagation is a strategy for adjusting parameters. Back propagation reverses the deviations generated by the training data after the neural network training. Transfer, assign the deviation value to each parameter to adjust, optimize the model, and reduce losses.
  • the working principle of the present invention when taking off the drone, the GPS positioning is used to follow the prescribed path to start the lifting motor 25 on the support frame 24.
  • the lifting motor 25 drives the driving fan blade 26 to rotate, thereby generating lift to drive the inspection device body 1.
  • the first drive motor 10 inside the transmission platform 9 is started.
  • the first drive motor 10 drives the drive rod 11 to drive clockwise. At the same time, it drives counterclockwise within three seconds to achieve reciprocating drive.
  • the drive rod 11 drives the holographic three-dimensional scan.
  • the instrument 12 drives left and right reciprocally, and at the same time starts the second drive motor 13 to drive the rotating rod 14 to rotate counterclockwise.
  • the rotating rod 14 drives the connecting shaft block 15 to rotate, and the connecting shaft block 15 drives the measuring camera 16 to rotate. In this way, the measuring camera 16 starts Carry out left and right reciprocating and up and down angle reciprocating driving for side irradiation;
  • the rotating rod 14 drives the driving collar rod 17 to rotate, the driving collar rod 17 drives the adjusting rotating rod 18 to deflect downward, and the adjusting rotating rod 18 drives the adjusting collar.
  • the block 20 is deflected downward, and when the first illumination lamp 21 is damaged, it can be rotated by starting the switching motor 19.
  • the switching motor 19 drives the adjusting rotating rod 18 to rotate, and the adjusting rotating rod 18 drives the adjusting collar block 20 to rotate. Rotate one hundred and eighty degrees clockwise to switch the second illumination lamp 22 and the first illumination lamp 21 to prevent the first illumination lamp 21 from being damaged and unable to be used normally;
  • the servo drive motor 5 can be started to drive the transmission shaft 4 to drive clockwise, and the transmission shaft 4 drives the two connecting collar rods 6 in The connecting shaft block 3 is driven, and the connecting collar rod 6 drives the fixed rod 7 to drive the connecting collar block 8 clockwise.
  • the connecting collar block 8 drives the transmission platform 9 to flip upward, and the transmission platform 9 drives the holographic three-dimensional scanner 12 Flip the measuring camera 16 up and clockwise ninety degrees, and the measuring camera 16 starts to face the top of the flood discharge tunnel. This can start the first driving motor 10 to drive the driving rod 11 for reciprocating driving, so that the top of the tunnel can be scanned without blind spots;
  • the transmitted image and scan data are transmitted to the wireless receiver 40 through the wireless transmitter 27, and are received by the wireless receiver 40 and transmitted to the image storage 43, and are processed by the image processor 44 for image comparison. , select different levels, and the sound and light alarm 42 can play the role of alarm.
  • the measurement camera 16 is used to achieve measurements facing the top surface of the flood discharge tunnel and measurements inside.
  • the holographic three-dimensional scanner 12 is used to scan the tunnel roof without blind spots and the measurement camera 16 is constructed to perform dimensional measurement during the scanning process, thereby achieving simultaneous three-dimensional panoramic scanning and data measurement.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • Multimedia (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Electromagnetism (AREA)
  • Quality & Reliability (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Signal Processing (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

本发明涉及一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备，属于巡检及缺陷检测技术领域。该方法包括如下步骤：建立泄洪隧洞蚀损缺陷数据库；建立泄洪隧洞蚀损缺陷分析、评价标准和预警阈值；自动巡检；图像分析处理与智能识别模型训练。本发明能够自动巡检泄洪隧洞情况，并能快速识别蚀损类型、分析判断蚀损影响程度，蚀损超出阈值时予以预警报警提示，并且可以自动、快速预警，提醒处理，避免泄洪时造成下游生命安全发生危险，智能化程度更高，事故预防能力更强，巡检效率以及检测效率大幅度提高，人工成本大幅降低，同时避免人员进入泄洪隧道发生安全事故。

Description

一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备 技术领域
本发明属于巡检及缺陷检测技术领域，具体涉及一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备。
 背景技术
泄水消能建筑物运行期发生破坏的几率很高。据不完全统计，近1/3已建泄水消能建筑物发生不同程度的破坏。泄水消能建筑物发生破坏几率较高的原因，主要在于影响泄水消能建筑物安全性的因素繁多且复杂。根据调研结果，超标来水条件、设计水平不足、施工质量存在薄弱环节、运行调度不合理、长期服役过程中出现病害或损伤等均是造成泄水消能建筑物破坏的主要原因，且很多工程发生破坏是由于其中多种因素共同引起的，如不及时发现并采取有效修复措施，一旦发生破坏，轻则影响泄水消能建筑物本身安全，更甚者影响相邻的坝体、边坡安全，重则影响下游生命安全，造成严重的社会经济损失，安全性极大。
 发明内容
本发明的目的是为了解决现有技术的不足，提供一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备。本发明能够自动巡检泄洪隧洞情况，并能全方位、快速、高精度识别蚀损缺陷类型、部位、形状、深度，智能分析判断蚀损影响程度，蚀损缺陷超出阈值时予以预警报警提示，并且可以自动、快速预警，提醒处理，避免泄洪时造成下游生命安全发生危险，智能化程度更高，事故预防能力更强，巡检效率以及检测效率大幅度提高，人工成本大幅降低，同时避免人员进入泄洪隧道开展高危巡检测量作业而发生安全事故。
为实现上述目的，本发明采用的技术方案如下：
一种泄洪隧洞自动巡检和蚀损缺陷智能检测设备，包括巡检装置本体，所述巡检装置本体顶端设置有防护壳，所述防护壳内部设置有调节机构；
所述调节机构包括设置在防护壳内部的连接支轴块，所述连接支轴块一侧连接有传动轴杆，所述传动轴杆的另一端设置有用于驱动的伺服驱动电机，所述传动轴杆外部套设有两个对称设置的连接轴环杆，所述连接轴环杆靠近另一端内部贯穿设置有固定杆，所述固定杆中部一侧连接有连接套环块，所述连接套环块一侧连接有传动平台，所述传动平台上安装有第一驱动电机，所述第一驱动电机输出端连接有驱动杆，所述驱动杆顶端设置有全息三维扫描仪，所述全息三维扫描仪内部设置有第二驱动电机。
进一步，优选的是，所述第二驱动电机输出端连接有转动杆，所述转动杆外部套设后连接轴块，所述连接轴块一侧固定连接有测量摄像头。
进一步，优选的是，所述转动杆一端外部均套设有联动切换机构，所述联动切换机构包括设置在转动杆一端外部的驱动套环杆，所述驱动套环杆靠近顶端内部活动连接有调节转杆，所述调节转杆一端安装有切换电机，所述切换电机输出端与调节转杆之间固定连接，所述调节转杆外部且位于驱动套环杆顶端一侧设置有调节轴环块，所述调节轴环块一侧安装有第一照射灯，所述调节轴环块另一侧安装有第二照射灯。
进一步，优选的是，所述驱动套环杆一侧安装有限位螺块。
进一步，优选的是，所述巡检装置本体外部对称设置有四个支撑架，所述支撑架靠近端部设置有升降电机，所述升降电机输出端连接有驱动扇叶，所述驱动扇叶一侧安装有水平传感器，所述巡检装置本体内部设置有无线发射器，两个相邻所述支撑架下方设置有弧形支架，所述弧形支架下方设置有缓冲底座，所述缓冲底座底端连接有凸起支座。
进一步，优选的是，所述无线发射器输出端连接有无线接收器，所述无线接收器底端安装有处理装置，所述无线接收器一侧安装有声光报警器，所述处理装置一侧嵌入安装有显示屏，所述显示屏下方设置有图像储存器，所述图像储存器一侧设置有图像处理器。
进一步，优选的是，所述巡检装置本体下方设置有第一电池，所述第一电池下方对称设置有两个调平导向板，两个调平导向板相对一侧均开设有导向槽，所述导向槽内部滑动设置有十字导向块，所述十字导向块一端连接有推动支杆，所述推动支杆一端安装有电动气缸，两个所述十字导向块之间设置有第二电池，所述第二电池底端连接有支撑板，所述第二电池一端设置有连接口。
本发明同时提供一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法，采用上述泄洪隧洞自动巡检和蚀损缺陷智能检测设备，包括如下步骤：
步骤一、建立泄洪隧洞蚀损缺陷数据库：采集国内外泄洪隧洞运行情况数据，建立泄洪隧洞蚀损缺陷数据库，包括国内外大量泄洪隧洞运行中出现的蚀损类型、蚀损部位、蚀损规模、蚀损影响；然后将泄洪隧洞蚀损缺陷数据库中的泄洪隧洞蚀损缺陷数据划分为3个等级，第1等级对泄洪安全无影响；第2等级有较小影响，需提醒注意，加强观测；第3等级有影响，需采取修补措施，预警提示；
步骤二、建立泄洪隧洞蚀损缺陷分析、评价标准和预警阈值：根据冲蚀、磨蚀、空蚀类型对泄洪安全的影响程度，建立缺陷类型评价标准：将隧洞表面未发现蚀损列为第1等级，将蚀损深度小于3mm列为第2等级，将蚀损深度大于等于3mm列为第3等级；预警阈值即蚀损深度大于等于3mm。
步骤三、自动巡检：采用泄洪隧洞自动巡检和蚀损缺陷智能检测设备自动巡检，按照起飞前设置好巡检路线，利用激光扫描实现对全隧洞巡检，并且进行照明操作，所采集的图像进行无线上传；
步骤四、图像分析处理与智能识别模型训练：建立图像数据库，将需要训练的每张图像进行两种处理，一份保存为原图，另一份贴上对应的真实标签，同时将同一类型标签的图片放入到同一类型的图片集中，搭建图像识别神经网络，将训练原图输入到神经网络中进行图像识别处理，得出输出的标签类型，与该图片的真实标签进行对比，并对结果进行损失计算，通过损失计算调整神经网络中的参数，更新神经网络模型，再次对图片进行图像识别，最终得到较为完善的图像识别神经网络；之后采用该图像识别神经网络对蚀损缺陷进行识别，并基于步骤二建立的评价标准和预警阈值进行相应的评价与预警。
进一步，优选的是，步骤四中，对图像进行识别时，识别信息包括蚀损类型、位置、形状、深度。
本发明中，第一电池与第二电池连接方式可以为利用线束进行串联操作，电池连接方式采用正常的连接方式即可，只是为了增加重量实现平衡操作，并且第一电池与第二电池为整台设备提供电源，其与用电结构部件的连接关系采用现有方式即可，本发明不做过多限制。
本发明与现有技术相比，其有益效果为：
本发明能够自动巡检泄洪隧洞情况，并能全方位、快速、高精度识别蚀损缺陷类型、部位、形状、深度，智能分析判断蚀损影响程度，蚀损缺陷超出阈值时予以预警报警提示，并且可以自动、快速预警，提醒处理，避免泄洪时造成下游生命安全发生危险，智能化程度更高，事故预防能力更强，巡检效率以及检测效率大幅度提高，人工成本大幅降低，同时避免人员进入泄洪隧道开展高危巡检测量作业而发生安全事故；
本发明通过建立的图像识别神经网络，通过训练逐渐提高计算机图像识别效率，对图像处理结果进行损失计算，可以完善神经网络，优化更新神经网络结构，进一步提高图像识别的准确性，为识别缺陷类型、分析判断缺陷影响程度等评估提供技术保障；
本发明采用调节机构使转动杆带动连接轴块进行转动，连接轴块带动测量摄像头进行转动，测量摄像头开始进行左右往复以及上下角度往复驱动，进行侧面照射，传动平台带动全息三维扫描仪使测量摄像头向上顺时针翻转九十度，测量摄像头开始面朝泄洪隧道顶面进行全角度扫描，这样可以有效提高对侧面死角位置的扫描，与泄洪隧洞顶端死角的扫描，不仅提高了扫描效率，而且提高扫描的全面性，优化了扫描效果，避免存在死角；
本发明采用联动切换机构使转动杆带动驱动套环杆进行转动，调节转杆带动调节轴环块进行向下角度偏转，这样可以跟随测量摄像头进行灯光照射操作，保证了灯光照射的准确性，为扫描过程中提供充足的灯光作用，并且可以当第一照射灯损坏后，切换第二照射灯，从而可以形成切换备用灯光操作，确保灯光能够正常使用，避免无人机因灯光问题进行返航，有效提高了巡检效率；
本发明采用测量摄像头向上顺时针翻转九十度的过程中，通过水平传感器进行水平角度探测，电动气缸带动推动支杆向左移动，第二电池带动支撑板向左移动，这样可以当测量摄像头进行切换角度时重心向后偏移，而第二电池可以向左移动保证重心平衡，可以提高无人机在使用过程中的水平稳定性，稳定性更好，保证无人机正常运行。
综上，通过上述多个作用的相互影响，自动、快速、稳定进行扫描，高精度、智能分析蚀损深度，及时自动预警，避免造成泄洪时对洞身及下游生命安全造成破坏和危险，风险预防能力更强，智能化水平更高，不仅提高了扫描效率，而且可以提高扫描的全面性，优化扫描效果，避免存在死角，切换备用灯光操作，确保灯光能够正常使用，避免无人机因灯光问题进行返航，有效提高巡检效率，提高无人机正常使用过程中的水平稳定性，同时采用智能算法，能够高效、快速、高精度、自动识别蚀损位置、形状、深度，并与阈值自动对比分析，智能预警和报警。综上可以有效提高巡检效率，提高蚀损缺陷智能分析效率和精度，提高巡检工作安全性，及时提醒处理泄洪隧洞安全问题，保证泄洪的安全性。
 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为本发明的无人机无线输送原理示意图。
图2为本发明的巡检装置本体整体结构示意图。
图3为本发明的无线发射器结构示意图。
图4为本发明的传动轴杆与伺服驱动电机连接处结构示意图。
图5为本发明的驱动杆与全息三维扫描仪拆分结构示意图。
图6为本发明的巡检装置本体下方结构示意图。
图7为本发明的十字导向块与第二电池连接处结构示意图。
图8为本发明的图5中A处放大结构示意图。
图9为发明的图像识别训练流程与其内部神经网络结构。
其中，1、巡检装置本体；2、防护壳；3、连接支轴块；4、传动轴杆；5、伺服驱动电机；6、连接轴环杆；7、固定杆；8、连接套环块；9、传动平台；10、第一驱动电机；11、驱动杆；12、全息三维扫描仪；13、第二驱动电机；14、转动杆；15、连接轴块；16、测量摄像头；17、驱动套环杆；18、调节转杆；19、切换电机；20、调节轴环块；21、第一照射灯；22、第二照射灯；23、限位螺块；24、支撑架；25、升降电机；26、驱动扇叶；27、无线发射器；28、第一电池；29、弧形支架；30、缓冲底座；31、凸起支座；32、调平导向板；33、导向槽；34、十字导向块；35、推动支杆；36、电动气缸；37、第二电池；38、支撑板；39、连接口；40、无线接收器；41、处理装置；42、声光报警器；43、图像储存器；44、图像处理器；45、显示屏；46、水平传感器。
 具体实施方式
下面结合实施例对本发明作进一步的详细描述。
本领域技术人员将会理解，下列实施例仅用于说明本发明，而不应视为限定本发明的范围。实施例中未注明具体技术或条件者，按照本领域内的文献所描述的技术或条件或者按照产品说明书进行。所用材料或设备未注明生产厂商者，均为可以通过购买获得的常规产品。本技术领域技术人员可以理解，除非特意声明，这里使用的单数形式“一”、“一个”、“所述”和“该”也可包括复数形式。应该进一步理解的是，本发明的说明书中使用的措辞“包括”是指存在所述特征、整数、步骤、操作、元件和/或组件，但是并不排除存在或添加一个或多个其他特征、整数、步骤、操作、元件、组件和/或它们的组。应该理解，当我们称元件被“连接”到另一元件时，它可以直接连接到其他元件，或者也可以存在中间元件。此外，这里使用的“连接”可以包括无线连接。这里使用的措辞“和/或”包括一个或更多个相关联的列出项的任一单元和全部组合。
在本发明的描述中，除非另有说明，“多个”的含义是两个或两个以上。术语“内”、“上”、“下”等指示的方位或状态关系为基于附图所示的方位或状态关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。
在本发明的描述中，需要说明的是，除非另有明确的规定和限定，术语“安装”、“连接”、“设有”应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连。对于本领域的普通技术人员而言，根据具体情况理解上述术语在本发明中的具体含义。
本技术领域技术人员可以理解，除非另外定义，这里使用的所有术语（包括技术术语和科学术语）具有与本发明所属领域中的普通技术人员的一般理解相同的意义。还应该理解的是，诸如通用字典中定义的那些术语应该被理解为具有与现有技术的上下文中的意义一致的意义，并且除非像这里一样定义，不会用理想化或过于正式的含义来解释。
一种泄洪隧洞自动巡检及缺陷检测设备，包括巡检装置本体1，巡检装置本体1顶端设置有防护壳2，防护壳2内部设置有调节机构；
调节机构包括设置在防护壳2内部的连接支轴块3，连接支轴块3一侧连接有传动轴杆4，传动轴杆4的另一端设置有用于驱动的伺服驱动电机5，传动轴杆4外部套设有两个对称设置的连接轴环杆6，连接轴环杆6靠近另一端内部贯穿设置有固定杆7，固定杆7中部一侧连接有连接套环块8，连接套环块8一侧连接有传动平台9，传动平台9上安装有第一驱动电机10，第一驱动电机10输出端连接有驱动杆11，驱动杆11顶端设置有全息三维扫描仪12，全息三维扫描仪12内部设置有第二驱动电机13。
如图5和图8所示，第二驱动电机13输出端连接有转动杆14，转动杆14外部套设后连接轴块15，连接轴块15一侧连接有测量摄像头16，连接轴块15与测量摄像头16之间固定连接，以便第二驱动电机13带动转动杆14进行逆时针转动，转动杆14带动连接轴块15进行转动，实现测量摄像头16的竖直方向的角度调节。
如图4和图5所示，驱动杆11上下两端分别与第一驱动电机10和全息三维扫描仪12两两之间固定连接，传动轴杆4与伺服驱动电机5之间固定连接，以便，可以有第一驱动电机10带动驱动杆11驱动全息三维扫描仪12进行横向角度偏转，伺服驱动电机5可以带动传动轴杆4进行驱动。
如图5所示，转动杆14一端外部均套设有联动切换机构，联动切换机构包括设置在转动杆14一端外部的驱动套环杆17，驱动套环杆17靠近顶端内部活动连接有调节转杆18，调节转杆18一端安装有切换电机19，切换电机19输出端与调节转杆18之间固定连接，调节转杆18外部且位于驱动套环杆17顶端一侧设置有调节轴环块20，调节轴环块20一侧安装有第一照射灯21，调节轴环块20另一侧安装有第二照射灯22，以便转动杆14带动驱动套环杆17进行转动，调节轴环块20进行向下角度偏转，第一照射灯21损坏后，可以通过启动切换电机19进行转动，第二照射灯22与第一照射灯21进行切换，方便后期启动备用灯光照射。
如图5所示，调节轴环块20两侧分别与第二照射灯22和第一照射灯21两两之间通过螺栓可拆卸式连接，驱动套环杆17与调节转杆18和转动杆14固定连接，驱动套环杆17一侧安装有限位螺块23，以便限位螺块23对驱动套环杆17起到限位的作用。
如图2和图3所示，巡检装置本体1外部对称设置有四个支撑架24，支撑架24靠近端部设置有升降电机25，升降电机25输出端连接有驱动扇叶26，驱动扇叶26一侧安装有水平传感器46，巡检装置本体1内部设置有无线发射器27，两个相邻支撑架24下方设置有弧形支架29，弧形支架29下方设置有缓冲底座30，缓冲底座30底端连接有凸起支座31，以便由水平传感器46探测支撑架24倾斜角度，升降电机25带动驱动扇叶26进行转动，从而产生升力进行飞行，且无人机落地时，凸起支座31对缓冲底座30起到缓冲的作用，缓冲底座30对两个弧形支架29起到缓冲的作用。
如图6和图7所示，巡检装置本体1下方设置有第一电池28，第一电池28下方对称设置有两个调平导向板32，两个调平导向板32相对一侧均开设有导向槽33，导向槽33内部滑动设置有十字导向块34，十字导向块34一端连接有推动支杆35，推动支杆35一端安装有电动气缸36，两个十字导向块34之间设置有第二电池37，第二电池37底端连接有支撑板38，第二电池37一端设置有连接口39，以便可以实现调平操作，保证测量摄像头16在进行翻转过程中整个无人机的水平状态。
如图1所示，无线发射器27输出端连接有无线接收器40，无线接收器40底端安装有处理装置41，无线接收器40一侧安装有声光报警器42，处理装置41一侧嵌入安装有显示屏45，显示屏45下方设置有图像储存器43，图像储存器43一侧设置有图像处理器44，以便无线接收器40接收传入到图像储存器43内经过图像对比识别，从而可以根据预警等级进行报警操作。
一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法，采用权利要求1~7任意一项所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于，包括如下步骤：
步骤一、建立泄洪隧洞蚀损缺陷数据库：采集国内外泄洪隧洞运行情况数据，建立泄洪隧洞蚀损缺陷数据库，包括国内外大量泄洪隧洞运行中出现的蚀损类型、蚀损部位、蚀损规模、蚀损影响；然后将泄洪隧洞蚀损缺陷数据库中的泄洪隧洞蚀损缺陷数据划分为3个等级，第1等级对泄洪安全无影响；第2等级有较小影响，需提醒注意，加强观测；第3等级有影响，需采取修补措施，预警提示；
步骤二、建立泄洪隧洞蚀损缺陷分析、评价标准和预警阈值指标体系：根据冲蚀、磨蚀、空蚀类型对泄洪安全的影响程度，建立缺陷类型评价标准：将隧洞表面未发现蚀损列为第1等级，将蚀损深度小于3mm列为第2等级，将蚀损深度大于等于3mm列为第3等级；预警阈值即蚀损深度大于等于3mm。
步骤三、自动巡检：采用泄洪隧洞自动巡检和蚀损缺陷智能检测设备或无人机自动巡检，按照起飞前设置好巡检路线，利用激光扫描实现对全隧洞无死角、高精度巡检，并且进行照明操作，所采集的图像进行无线上传；
步骤四、图像分析处理与智能识别模型训练：建立图像数据库，将需要训练的每张图像进行两种处理，一份保存为原图，另一份贴上对应的真实标签，同时将同一类型标签的图片放入到同一类型的图片集中，搭建图像识别神经网络，将训练原图输入到神经网络中进行图像识别处理，得出输出的标签类型，与该图片的真实标签进行对比，并对结果进行损失计算，通过损失计算调整神经网络中的参数，更新神经网络模型，再次对图片进行图像识别，最终得到较为完善的图像识别神经网络；之后采用该图像识别神经网络对蚀损缺陷进行识别，并基于步骤二建立的评价标准和预警阈值进行相应的评价与预警。
具体地，步骤四中，对图像进行识别时，识别信息包括蚀损类型、位置、形状、深度。
随着数据库的不断增大和计算机反复进行图像识别的深度学习，可逐渐提高计算机对图像的识别能力，提高识别的效率和准确率。
同时，可以对于神经网络模型进行升级，在泄洪前巡检一次，经历一次泄洪后再巡检一次，与泄洪之前的评估结果进行对比，如此往复通过多次巡检、多次泄洪，增加原有的数据库中各破坏类型的图片，加强计算机对巡检和泄洪过程对图像识别的深度学习，完善更新计算的神经网络框架，提高对后续预测结果的精度，继续更新数据等级评判。巡检需要两人以上进行同时巡检，一人检查一人做好记录，往复巡检次数不低于五次。
本发明中，训练需要较大的图片数据库，数据库的主要来源在于工程长期积累图片，训练中神经网络所涉及的算法模型与相应的函数并不唯一，可更换不同的算法模型与函数，不同的算法模型对图像的识别准确率有所不同，但通过反复的深度学习，不断扩大已有数据库，都可提高图像识别的准确率。
如图9所示，计算机将传输的图像先转化为数据，加载到已搭建的神经网络进行图像识别。神经网络模型有很多，但图像识别最常采用的是卷积神经网络，本发明在设计中采用卷积神经网络。神经网络结构包括输入层、隐层和输出层，输入层将图像转化的数据作为输入，隐层的工作是将输入转换成输出层可以使用的东西，输出层得出图像为特定类别的概率。其中卷积神经网络的隐层又包括卷积层、激活层、池化层与全连接层。卷积层的作用类似于过滤器，搜索图像中的每个部分和像素，提取图像关键的特征，找出识别模式，其中卷积层Hi的产生过程可描述为 ，其中Wi表示第i层卷积核的权值向量，运算符号 代表卷积核与第i-1层图像或者特征图进行卷积操作，bi为选用的偏移向量；由于卷积操作是线性的，为了增加模型的复杂程度，常通过激活层增加一个激活函数增加模型的非线性，用以模拟任何函数，增强网络的表达能力，卷积神经网络中常采用的激活函数为ReLU函数， ；由于卷积产生的特征图会存在大量的冗余信息，需要池化层进行缩小，降低特征图的维度，池化层过程分为最大池化和平均池化，池化过程是将输入图像平均划分成若干个矩形区域，最大的池化则是将池化区域的像素点取最大值，平均池化则是对池化区域内的图像取平均值；全连接层的作用是将经过多个卷积层和池化层的图像特征图中的特征进行整合，获取图像特征具有的高层含义，之后用于图像分类。得到输出结果后进行损失计算，损失函数计算的值用于表示神经网络里的概率模型和理想的概率模型之间的差距的定量表达，采用均方误差函数， ，其中 表示神经网络的输出， 表示监督数据的真实值， 表示数据的维度。为优化损失函数，减小神经网络模型与理想模型的差距，需调整神经网络参数，反向传播则是调整参数的策略，反向传播将训练数据经过神经网络训练后产生的的偏差进行反向传递，将偏差值分配给各个参数来调整，优化模型，减小损失。通过这种深度学习，在大量的图片数据基础上，计算机不断优化神经网络，提高预测的准确度，实现精确的图像识别功能。
本发明工作原理：起飞无人机时，利用GPS定位按照规定的路径，启动支撑架24上的升降电机25，升降电机25带动驱动扇叶26进行转动，从而产生升力带动巡检装置本体1进行飞行，当飞行到泄洪隧洞内部后，可以打开第一照射灯21进行灯光照射；
侧面扫描时，启动传动平台9内部的第一驱动电机10，第一驱动电机10带动驱动杆11进行顺时针驱动，同时在三秒内进行逆时针驱动实现往复驱动，驱动杆11带动全息三维扫描仪12进行左右往复驱动，同时启动第二驱动电机13带动转动杆14进行逆时针转动，转动杆14带动连接轴块15进行转动，连接轴块15带动测量摄像头16进行转动，这样测量摄像头16开始进行左右往复以及上下角度往复驱动，进行侧面照射；
灯光切换跟随照射时，转动杆14在进行转动的过程中，转动杆14带动驱动套环杆17进行转动，驱动套环杆17带动调节转杆18向下偏转，调节转杆18带动调节轴环块20进行向下角度偏转，并且当第一照射灯21损坏后，可以通过启动切换电机19进行转动，切换电机19带动调节转杆18进行进行转动，调节转杆18带动调节轴环块20进行顺时针旋转一百八十度，从而第二照射灯22与第一照射灯21进行切换，可以避免第一照射灯21损坏后无法正常使用；
顶端面扫描操作，当顺着规定的路径已经对泄洪隧洞侧面进行扫描完毕后，可以启动伺服驱动电机5带动传动轴杆4进行顺时针驱动，传动轴杆4带动两个连接轴环杆6在连接支轴块3上进行驱动，连接轴环杆6带动固定杆7使连接套环块8顺时针驱动，这样连接套环块8带动传动平台9向上翻转，传动平台9带动全息三维扫描仪12使测量摄像头16向上顺时针翻转九十度，测量摄像头16开始面朝泄洪隧道顶面，这样可以启动第一驱动电机10带动驱动杆11进行往复驱动，这样可以对隧道顶面无死角扫描；
稳定校平，当测量摄像头16向上顺时针翻转九十度的过程中，支撑架24产生角度偏移，同时可以通过水平传感器46进行水平探测，并且可以启动电动气缸36带动推动支杆35向左移动，推动支杆35带动十字导向块34向左移动，十字导向块34沿着导向槽33向左移动，并且十字导向块34顺着调平导向板32向左移动，十字导向块34带动两个第二电池37向左移动，第二电池37带动支撑板38向左移动，支撑板38带动连接口39向左移动，这样可以第一驱动电机10支撑板38的重量向左移动，从而与测量摄像头16翻转后的重量持平，这样可以保证在翻转过程中巡检装置本体1的稳定性；
数据回传报警时，当传输回来的图像和扫描数据通过无线发射器27输送到无线接收器40，由无线接收器40接收传入到图像储存器43内，经过图像处理器44进行图像对比处理，选择不同等级，可以有声光报警器42起到报警的作用。
测量摄像头16用于实现面朝泄洪隧道顶面的测量以及内部的测量。
全息三维扫描仪12用于隧道顶面无死角扫描的扫描与测量摄像头16构建扫描过程中进行尺寸测量，实现三维全景扫描以及数据测量同步进行。
以上显示和描述了本发明的基本原理、主要特征和本发明的优点。本行业的技术人员应该了解，本发明不受上述实施例的限制，上述实施例和说明书中描述的只是说明本发明的原理，在不脱离本发明精神和范围的前提下，本发明还会有各种变化和改进，这些变化和改进都落入要求保护的本发明范围内。本发明要求保护范围由所附的权利要求书及其等效物界定。

Claims (9)

  1. 一种泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    包括巡检装置本体（1），所述巡检装置本体（1）顶端设置有防护壳（2），所述防护壳（2）内部设置有调节机构；
    所述调节机构包括设置在防护壳（2）内部的连接支轴块（3），所述连接支轴块（3）一侧连接有传动轴杆（4），所述传动轴杆（4）的另一端设置有用于驱动的伺服驱动电机（5），所述传动轴杆（4）外部套设有两个对称设置的连接轴环杆（6），所述连接轴环杆（6）靠近另一端内部贯穿设置有固定杆（7），所述固定杆（7）中部一侧连接有连接套环块（8），所述连接套环块（8）一侧连接有传动平台（9），所述传动平台（9）上安装有第一驱动电机（10），所述第一驱动电机（10）输出端连接有驱动杆（11），所述驱动杆（11）顶端设置有全息三维扫描仪（12），所述全息三维扫描仪（12）内部设置有第二驱动电机（13）。
  2. 根据权利要求1所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    所述第二驱动电机（13）输出端连接有转动杆（14），所述转动杆（14）外部套设后连接轴块（15），所述连接轴块（15）一侧固定连接有测量摄像头（16）。
  3. 根据权利要求2所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    所述转动杆（14）一端外部均套设有联动切换机构，所述联动切换机构包括设置在转动杆（14）一端外部的驱动套环杆（17），所述驱动套环杆（17）靠近顶端内部活动连接有调节转杆（18），所述调节转杆（18）一端安装有切换电机（19），所述切换电机（19）输出端与调节转杆（18）之间固定连接，所述调节转杆（18）外部且位于驱动套环杆（17）顶端一侧设置有调节轴环块（20），所述调节轴环块（20）一侧安装有第一照射灯（21），所述调节轴环块（20）另一侧安装有第二照射灯（22）。
  4. 根据权利要求3所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    所述驱动套环杆（17）一侧安装有限位螺块（23）。
  5. 根据权利要求1所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    所述巡检装置本体（1）外部对称设置有四个支撑架（24），所述支撑架（24）靠近端部设置有升降电机（25），所述升降电机（25）输出端连接有驱动扇叶（26），所述驱动扇叶（26）一侧安装有水平传感器（46），所述巡检装置本体（1）内部设置有无线发射器（27），两个相邻所述支撑架（24）下方设置有弧形支架（29），所述弧形支架（29）下方设置有缓冲底座（30），所述缓冲底座（30）底端连接有凸起支座（31）。
  6. 根据权利要求5所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    所述无线发射器（27）输出端连接有无线接收器（40），所述无线接收器（40）底端安装有处理装置（41），所述无线接收器（40）一侧安装有声光报警器（42），所述处理装置（41）一侧嵌入安装有显示屏（45），所述显示屏（45）下方设置有图像储存器（43），所述图像储存器（43）一侧设置有图像处理器（44）。
  7. 根据权利要求1所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于：
    所述巡检装置本体（1）下方设置有第一电池（28），所述第一电池（28）下方对称设置有两个调平导向板（32），两个调平导向板（32）相对一侧均开设有导向槽（33），所述导向槽（33）内部滑动设置有十字导向块（34），所述十字导向块（34）一端连接有推动支杆（35），所述推动支杆（35）一端安装有电动气缸（36），两个所述十字导向块（34）之间设置有第二电池（37），所述第二电池（37）底端连接有支撑板（38），所述第二电池（37）一端设置有连接口（39）。
  8. 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法，采用权利要求1~7任意一项所述的泄洪隧洞自动巡检和蚀损缺陷智能检测设备，其特征在于，
    包括如下步骤：
    步骤一、建立泄洪隧洞蚀损缺陷数据库：采集国内外泄洪隧洞运行情况数据，建立泄洪隧洞蚀损缺陷数据库，包括国内外大量泄洪隧洞运行中出现的蚀损类型、蚀损部位、蚀损规模、蚀损影响；然后将泄洪隧洞蚀损缺陷数据库中的泄洪隧洞蚀损缺陷数据划分为3个等级，第1等级对泄洪安全无影响；第2等级有较小影响，需提醒注意，加强观测；第3等级有影响，需采取修补措施，预警提示；
    步骤二、建立泄洪隧洞蚀损缺陷分析、评价标准和预警阈值：根据冲蚀、磨蚀、空蚀类型对泄洪安全的影响程度，建立缺陷类型评价标准：将隧洞表面未发现蚀损列为第1等级，将蚀损深度小于3mm列为第2等级，将蚀损深度大于等于3mm列为第3等级；
    步骤三、自动巡检：采用泄洪隧洞自动巡检和蚀损缺陷智能检测设备自动巡检，按照起飞前设置好巡检路线，利用激光扫描实现对全隧洞巡检，并且进行照明操作，所采集的图像进行无线上传；
    步骤四、图像分析处理与智能识别模型训练：建立图像数据库，将需要训练的每张图像进行两种处理，一份保存为原图，另一份贴上对应的真实标签，同时将同一类型标签的图片放入到同一类型的图片集中，搭建图像识别神经网络，将训练原图输入到神经网络中进行图像识别处理，得出输出的标签类型，与该图片的真实标签进行对比，并对结果进行损失计算，通过损失计算调整神经网络中的参数，更新神经网络模型，再次对图片进行图像识别，最终得到较为完善的图像识别神经网络；之后采用该图像识别神经网络对蚀损缺陷进行识别，并基于步骤二建立的评价标准和预警阈值进行相应的评价与预警。
  9. 根据权利要求8所述的泄洪隧洞自动巡检和蚀损缺陷智能检测方法，其特征在于，
    步骤四中，对图像进行识别时，识别信息包括蚀损类型、位置、形状、深度。
PCT/CN2023/082049 2022-08-26 2023-03-17 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备 WO2024040945A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US18/329,570 US20240183743A1 (en) 2022-08-26 2023-06-05 Automatic patrol inspection and intelligent erosion defect detection method and apparatus for flood discharge tunnel

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211034688.6 2022-08-26
CN202211034688.6A CN115616000A (zh) 2022-08-26 2022-08-26 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/329,570 Continuation US20240183743A1 (en) 2022-08-26 2023-06-05 Automatic patrol inspection and intelligent erosion defect detection method and apparatus for flood discharge tunnel

Publications (1)

Publication Number Publication Date
WO2024040945A1 true WO2024040945A1 (zh) 2024-02-29

Family

ID=84856498

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/082049 WO2024040945A1 (zh) 2022-08-26 2023-03-17 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备

Country Status (3)

Country Link
US (1) US20240183743A1 (zh)
CN (1) CN115616000A (zh)
WO (1) WO2024040945A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117963195A (zh) * 2024-03-28 2024-05-03 建潘鲲鹭物联网技术研究院(厦门)有限公司 一种基于无人机的输电网智能巡检装置及其方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115616000A (zh) * 2022-08-26 2023-01-17 华能澜沧江水电股份有限公司 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备
CN117719716B (zh) * 2024-02-07 2024-05-28 华能澜沧江新能源有限公司 一种分布式光伏三维建模的无人机测绘系统及设备

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN203224453U (zh) * 2013-05-17 2013-10-02 深圳市施罗德工业测控设备有限公司 多视角管道检测爬行器
CN207067019U (zh) * 2017-08-07 2018-03-02 三峡大学 一种工程建筑裂缝检测无人机
CN207539634U (zh) * 2017-12-02 2018-06-26 浙江景迈环境工程有限公司 一种具有旋转摄像组件的管道检测小车
CN108318499A (zh) * 2018-01-05 2018-07-24 株洲时代电子技术有限公司 一种桥梁巡检方法
CN110398390A (zh) * 2019-07-04 2019-11-01 湖南城市学院 一种用于地下天然岩洞的探测取样设备
US20210396842A1 (en) * 2019-10-09 2021-12-23 Shandong University Multi-scale inspection and intelligent diagnosis system and method for tunnel structural defects
CN215640904U (zh) * 2021-09-18 2022-01-25 四川省凡灯环保科技有限责任公司 一种可移动建筑材料检测装置
CN115616000A (zh) * 2022-08-26 2023-01-17 华能澜沧江水电股份有限公司 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN203224453U (zh) * 2013-05-17 2013-10-02 深圳市施罗德工业测控设备有限公司 多视角管道检测爬行器
CN207067019U (zh) * 2017-08-07 2018-03-02 三峡大学 一种工程建筑裂缝检测无人机
CN207539634U (zh) * 2017-12-02 2018-06-26 浙江景迈环境工程有限公司 一种具有旋转摄像组件的管道检测小车
CN108318499A (zh) * 2018-01-05 2018-07-24 株洲时代电子技术有限公司 一种桥梁巡检方法
CN110398390A (zh) * 2019-07-04 2019-11-01 湖南城市学院 一种用于地下天然岩洞的探测取样设备
US20210396842A1 (en) * 2019-10-09 2021-12-23 Shandong University Multi-scale inspection and intelligent diagnosis system and method for tunnel structural defects
CN215640904U (zh) * 2021-09-18 2022-01-25 四川省凡灯环保科技有限责任公司 一种可移动建筑材料检测装置
CN115616000A (zh) * 2022-08-26 2023-01-17 华能澜沧江水电股份有限公司 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117963195A (zh) * 2024-03-28 2024-05-03 建潘鲲鹭物联网技术研究院(厦门)有限公司 一种基于无人机的输电网智能巡检装置及其方法

Also Published As

Publication number Publication date
CN115616000A (zh) 2023-01-17
US20240183743A1 (en) 2024-06-06

Similar Documents

Publication Publication Date Title
WO2024040945A1 (zh) 一种泄洪隧洞自动巡检和蚀损缺陷智能检测方法及设备
CN112700444B (zh) 基于自注意力与中心点回归模型的桥梁螺栓检测方法
CN110503112B (zh) 一种增强特征学习的小目标检测及识别方法
CN103543161A (zh) 一种连铸坯表面质量在线检测方法
CN110991466A (zh) 一种基于新型视觉传感设备的高速公路路面状况检测系统
CN109060826B (zh) 一种不停机的风电叶片检测装置
WO2021027366A1 (zh) 一种光伏电站的电池片检测方法、装置及系统
CN109374043A (zh) 一种基于bim的运营桥梁多源检测系统及检测方法
CN204142649U (zh) 一种容器缺口的在线视觉检测装置
CN111414807A (zh) 一种基于yolo技术的潮水识别与危机预警方法
CN109087305A (zh) 一种基于深度卷积神经网络的裂缝图像分割方法
CN111157537A (zh) 一种基于仿生机器视觉的带钢缺陷检测装置及系统
CN107421963A (zh) 一种核燃料生坯裂纹检测装置及裂纹检测方法
CN1519143A (zh) 电气化铁路接触网视觉检测系统及其方法
CN114689600A (zh) 一种桥梁混凝土结构表面裂缝检测方法及系统
CN111046936A (zh) 基于级联宽度神经网络桥梁裂缝分类识别方法
Liu et al. Improved YOLOX-S abnormal condition detection for power transmission line corridors
CN112345531B (zh) 一种基于仿生机器鱼的变压器故障检测方法
CN117197412B (zh) 一种基于ar的公路病害智能巡检系统及方法
Özer et al. An approach based on deep learning methods to detect the condition of solar panels in solar power plants
CN116433580A (zh) 一种基于智能无人机的大坝裂缝检测方法
CN115761466A (zh) 一种基于深度学习的混合驱动水下小型监测系统
CN208544411U (zh) 应用于受限空间安全探测的带有下置式旋翼的无人机
CN110525595A (zh) 一种水面垃圾监测系统及方法
CN202141434U (zh) 角码宽度视觉测量装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23856042

Country of ref document: EP

Kind code of ref document: A1