Steel box girder bridge apparent disease assessment method and system based on unmanned aerial vehicle inspection

Abstract

The invention discloses a steel box girder bridge apparent disease assessment method and system based on unmanned aerial vehicle inspection. The method comprises the following steps that an unmanned aerial vehicle airport and an unmanned aerial vehicle are deployed on a steel box girder bridge site; the method comprises the following steps: planning a photographing route in an unmanned aerial vehicle flight control system, acquiring a steel box girder picture by adopting an unmanned aerial vehicle matched with an aerial survey camera, and performing three-dimensional reconstruction on the acquired picture to form a three-dimensional'coarse model 'of the steel box girder; calculating the relative flight height, the camera view width and height, and scale factors in the width and height direction; planning a patrol route; the patrol route is imported into an unmanned aerial vehicle flight control system, and an unmanned aerial vehicle is matched with an aerial survey camera to collect apparent coating pictures of steel box girder webs and bottom plates; coating disease photos of various steel box girders are collected, a target detection network model is constructed, and coating disease types and areas are identified based on a target detection algorithm; according to the invention, automatic identification of the apparent coating diseases of the steel box girder is realized, the deterioration degree of the coating can be automatically evaluated, a detection report can be automatically generated, the detection quality is ensured, and the detection efficiency is improved.

Description

Technical Field

[0001] The present application relates to the technical field of bridge defect assessment, and specifically to a method and system for assessing the apparent defects of steel box girder bridges based on drone inspection. Background Art

[0002] Painting is the main measure for corrosion protection of steel box girder bridges. During the operation of steel box girder bridges, the coating is subjected to the effects of moisture, corrosive media and ultraviolet rays in the atmosphere, and will have apparent defects such as powdering, bubbles, cracks, rust, and peeling. These defects will cause premature failure of the coating, making the steel box girder bridge directly exposed to the air, in contact with moisture, oxygen and other corrosive substances, and the corrosion rate will accelerate, thereby gradually weakening the bearing capacity and durability of the steel box girder bridge. Therefore, the detection and evaluation of apparent coating defects during the operation stage of steel box girder bridges should be strengthened to provide timely support for the maintenance and repair decisions of steel box girder bridges.

[0003] At present, the detection of apparent defects of anti-corrosion coatings on steel box girder bridges mainly relies on manual work, including daily inspections, regular inspections, and periodic inspections. However, these inspection methods face many challenges in practical applications. Daily inspections are mostly carried out in a "car inspection" mode, that is, staff conduct visual inspections by car. This method can only detect significant defects and abnormalities. Due to the limited viewing angle and observation conditions, it is difficult to touch the coating defects in hidden areas such as the web and bottom plate of the steel box girder, and there are blind spots for detection. Regular inspections can check the appearance cleanliness of the bridge and the surface integrity of the components, but due to limited maintenance funds, many maintenance units often reduce the frequency of such inspections, resulting in potential defects that are difficult to detect in time. Regular inspections can conduct a more comprehensive inspection of the bridge. Article 3.5.2 of the "Highway Bridge and Culvert Maintenance Code" (JTG 5120-2021) stipulates: "Regular inspections should approach each component to carefully inspect its defects." However, for areas that are difficult to reach manually, such as the bottom plate and outer web of the steel box girder of the cross-river bridge (generally without maintenance channels), it is necessary to rely on auxiliary equipment such as bridge inspection vehicles, which not only increases the difficulty and cost of inspection, such as occupying the lane and affecting traffic, but also brings additional safety risks, especially in the working environment under the water. In addition, the results of manual inspection are highly dependent on the professional knowledge and experience of the inspectors, which may lead to misjudgment due to differences in subjective judgment. At the same time, due to the frequency of inspection (for example, for bridges with maintenance inspection level I, the regular inspection cycle is generally 1 year), it is difficult to achieve continuous tracking and timely treatment of the apparent diseases of the anti-corrosion coating of the steel box girder, which may miss the key opportunity to prevent the development of diseases. Therefore, it is imperative to optimize the detection method of the apparent diseases of the anti-corrosion coating of the steel box girder, and it is necessary to explore the introduction of more efficient and accurate monitoring technologies to overcome the limitations of existing detection methods, improve monitoring efficiency and accuracy, and ensure the safety and durability of bridge structures.

[0004] In recent years, with the development of machine vision, artificial intelligence, drones and other technologies, drone inspection methods have been gradually applied to the detection of bridge surface diseases, and relevant engineering and academic fields have explored this. In the field of engineering, the actual bridge inspection operation currently relies on drone pilots to manually control drones to take bridge photos, and then the inspectors manually import the pictures into the computer, and then conduct manual visual inspection and comparison. In the academic field, Tang Yong et al. developed a bridge crack detection system based on drone image technology and SVM algorithm in machine learning. The crack pictures of the interchange section of Jiaoping Expressway in Henan Province were collected by drones. The results showed that the crack parameters can be automatically calculated, and the overall classification accuracy of 96% was obtained. Yang Guojun et al. constructed an integrated model for bridge crack recognition and segmentation based on improved YOLOv7 and SeaFormer, and proposed a pixel-level calculation method for crack length and width based on inscribed circles. The cracks were tested in public concrete bridge crack images. The results showed that the crack recognition and positioning accuracy reached 91.38%, the relative error of crack width detection was within 18%, and the relative error of crack length detection was within 10%. Ding Junfeng et al. combined the HABC algorithm with the greedy strategy to propose a greedy HABC algorithm to realize the path planning of multiple target points of bridge pier structure detected by UAV. The reliability of the algorithm was verified by simulation. The results showed that the algorithm running time can be effectively saved by sacrificing a small amount of total path length. Ning Chaoyang et al. described the fine route planning and rapid modeling method of high pier bridges in mountainous areas and the quantification method of pier geometric parameters based on point cloud. Through actual engineering applications, it was found that the refined route of UAV can effectively improve the accuracy of three-dimensional point cloud modeling of bridges. The quantification errors of the radius, height and volume of the piers were within 2%, 5% and 8% respectively.

[0005] However, existing literature mainly focuses on the detection of cracks in concrete bridges, and research on the detection of apparent defects in the coating of steel box girder bridges is relatively scarce. At the same time, although drone technology is increasingly used in bridge inspections, current drone inspection research focuses more on drone flight efficiency, coverage, and three-dimensional reconstruction of bridges. However, for specific structures such as cross-river steel box girder bridges, especially for hard-to-reach areas such as bottom plates and the outside of webs, the drone inspection path planning method is still unclear. In addition, the main purpose of drone inspections is to efficiently and accurately discover and evaluate the apparent defects of bridges, rather than just building a three-dimensional model of the bridge. Therefore, the existing drone inspection path planning methods and research results have great limitations in guiding the detection of apparent defects in the coating of cross-river steel box girder bridges.

[0006] At the same time, in actual bridge inspection operations, most of them rely on drone pilots to manually control drones to take photos of bridges. This method not only requires pilots to have superb beyond-visual-range flight skills, but also significantly increases the operational risks, especially in complex environments across rivers, such as strong winds and water environments. Once the safety boundary is misjudged, it is very easy to cause drone collision accidents. In addition, the inspection process requires inspectors to manually import the pictures into the computer, and then conduct manual visual inspection and comparison. This process is not only time-consuming and laborious, but also difficult to achieve automatic identification and evaluation of apparent defects in the coating of steel box girder bridges, which greatly limits the potential of drone inspection technology in improving inspection efficiency and accuracy. Summary of the invention

[0007] In response to the problems existing in the prior art, this application proposes a method and system for evaluating the apparent defects of steel box girder bridges based on drone inspection, which can realize the automatic identification and evaluation of steel box girder coating defects.

[0008] A method for evaluating the apparent defects of a steel box girder bridge based on drone inspection includes the following steps:

[0009] S1, deployment of drone airport and drones at the steel box girder bridge site;

[0010] S2, plan the photography route in the UAV flight control system, and then use the UAV with the aerial survey camera to collect photos of the steel box girder. After the shooting is completed, the photos are reconstructed in three dimensions to form a three-dimensional "rough model" of the steel box girder;

[0011] S3, calculate relative altitude:

[0012]

[0013] In formula (1), H represents the relative altitude, that is, the vertical distance between the camera lens and the object being photographed; f represents the focal length of the camera, that is, the distance between the image principal point and the rear node of the camera lens; GSD represents the ground resolution; p represents the pixel size;

[0014]

[0015] In formula (2) and formula (3), FOV_W represents the width of the camera field of view; FOV_V represents the height of the camera field of view; WD represents the working distance, that is, the distance from the lens to the target object being photographed, and WD is approximately equal to DC; W represents the width of the image sensor, in millimeters; V represents the height of the image sensor, in millimeters;

[0016]

[0017] In formula (4) and formula (5), RE_W represents the scale factor of the camera in the width direction, and the unit is millimeter per pixel; RE_V represents the scale factor of the camera in the height direction, and the unit is millimeter per pixel; PIX_W represents the number of pixels of the camera in the width direction; PIX_V represents the number of pixels of the camera in the height direction;

[0018] S4, input the relative flight height H, the pitch angle α of the web of the steel box girder, and the route speed v on the three-dimensional “rough mold” of the steel box girder to plan the inspection route;

[0019] S5, import the inspection route into the UAV flight control system, set the inspection frequency and time of the UAV, and set the camera scale factor parameters; use the UAV with the aerial survey camera to collect the surface coating photos of the web and bottom plate of the steel box girder;

[0020] S6, collect various types of steel box girder coating disease photos, build a steel box girder coating disease data set, input the data set into the target detection network model, train and optimize the model parameters;

[0021] S7, identifying the steel box girder coating defect photos collected by the drone with the aerial survey camera, and obtaining the apparent defect category and defect area of ​​the steel box girder anti-corrosion coating.

[0022] Preferably, in step S2, five tilt photography routes, namely, front, back, left, right, and top, are planned in the UAV flight control system.

[0023] Furthermore, in step S4, the inspection route includes an inspection route of the left web and left flange plate of the steel box girder, an inspection route of the bottom plate of the steel box girder, and a refined inspection route of the right web and right flange plate of the steel box girder.

[0024] Preferably, in step S6, the target detection network model is YOLOv8.

[0025] Furthermore, after obtaining the apparent defect category and defect area of ​​the anti-corrosion coating of the steel box girder, a step of evaluating the degradation scale of the anti-corrosion coating of the steel box girder is also included.

[0026] Furthermore, after obtaining the apparent defect category and defect area of ​​the anti-corrosion coating of the steel box girder, the method also includes the step of automatically generating an apparent defect detection report of the anti-corrosion coating of the steel box girder.

[0027] Optionally, according to the provisions of the Highway Bridge Technical Condition Assessment Standard, the degradation scale of the anti-corrosion coating of the web and bottom plates of the steel box girder is assessed by the area of ​​the damage, and a steel box girder anti-corrosion coating apparent damage inspection report is generated.

[0028] A system for executing the method for evaluating apparent defects of a steel box girder bridge, comprising:

[0029] A drone terminal, including a drone and an aerial survey camera, is used to remotely perform a steel box girder flight mission and collect photos of the steel box girder web and / or bottom plate and / or flange plate;

[0030] A wireless communication module is used to remotely transmit the photos taken by the drone to the storage module;

[0031] A storage module storing photos of the web plate and / or the bottom plate and / or the flange plate of the steel box girder;

[0032] The computing module is deployed on the cloud platform; the computing module includes a target detection network model, which is used to calculate the relative altitude of the drone, the proportional factors of the camera in the width and height directions, and identify the apparent disease category and disease area of ​​the anti-corrosion coating of the steel box girder.

[0033] Furthermore, the system also includes:

[0034] The evaluation module is electrically connected or communication-connected with the calculation module and is used to evaluate the degradation scale of the anti-corrosion coating of the steel box girder.

[0035] Furthermore, the system further comprises:

[0036] The inspection report generation module is electrically connected or communication-connected to the calculation module, and is used to automatically generate an inspection report on the surface defects of the anti-corrosion coating of the steel box girder.

[0037] Compared with the existing methods, this application has the following beneficial effects:

[0038] (1) Through unmanned airports and drone terminals, inspection tasks are remotely issued, and the flexibility and efficiency of drones are fully utilized to achieve rapid acquisition of surface images of steel box girders, laying the foundation for the detection of surface defects of anti-corrosion coatings on steel box girders. This acquisition method replaces the traditional manual detection method with aerial platforms and bridge inspection vehicles, solving the problems of low efficiency, high cost, and difficulty in reaching hidden areas faced by manual detection, and eliminating the collision risks faced by manually controlled drones in strong winds and complex water environments, greatly improving field efficiency and frequency, and reducing costs. It is especially suitable for steel box girder bridges across rivers.

[0039] (2) The inspection route for steel box girder bridges was determined to meet the requirements of scheduled and quantitative automated inspections of routine and key areas, greatly improving the efficiency of drone inspections;

[0040] (3) The integrated target detection algorithm realizes the automatic identification of the surface coating defects of steel box girders, and can automatically evaluate the degree of coating degradation and automatically generate inspection reports, solving the problem of strong subjectivity, low efficiency, and easy misjudgment and omission in manual photo interpretation. It realizes the automatic identification and evaluation of coating defects on steel box girders, ensures the inspection quality, and improves the efficiency of internal inspection.

[0041] (4) The detection result dataset can provide data support for the damage maintenance module of the bridge information management system, and play a positive role in promoting the realization of intelligent bridge maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a system module according to an embodiment of the present application;

[0043] Figure 2 The 3D model of the main bridge of Chongqi Bridge;

[0044] Figure 3 A schematic diagram of a camera pitch angle according to an embodiment of the present application;

[0045] Figure 4 A schematic diagram of route planning according to an embodiment of the present application;

[0046] Figure 5 This is a schematic diagram of the segmentation of a paint defect instance using the YOLOv8 algorithm in an embodiment of the present application. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings.

[0048] It should be clear that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0049] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0050] It should be understood that the term "and / or" used in this article is merely a way to describe the same field of associated objects, indicating that three relationships may exist.

[0051] In a first aspect, the present application provides a method for evaluating the apparent defects of a steel box girder bridge based on drone inspection, comprising the following steps:

[0052] S1, deployment of drone airport and drones at the steel box girder bridge site;

[0053] S2, plan the photography route in the UAV flight control system, and then use the UAV with the aerial survey camera to collect photos of the steel box girder. After the shooting is completed, the photos are reconstructed in three dimensions to form a three-dimensional "rough model" of the steel box girder;

[0054] S3, calculate relative altitude:

[0055]

[0056] In formula (1), H represents the relative altitude, that is, the vertical distance between the camera lens and the object being photographed; f represents the focal length of the camera, that is, the distance between the image principal point and the rear node of the camera lens; GSD represents the ground resolution; p represents the pixel size;

[0057] like Figure 3 The figure shows a schematic diagram of the camera pitch angle, where α represents the camera pitch angle, i.e. the pitch angle of the web of the steel box girder, and this parameter is determined from the construction drawings of the bridge;

[0058]

[0059] In formula (2) and formula (3), FOV_W represents the width of the camera field of view; FOV_V represents the height of the camera field of view; WD represents the working distance, that is, the distance from the lens to the target object, and WD is approximately equal to DC; the image sensor is the core component of the camera mounted on the drone, W represents the width of the image sensor, in millimeters (mm); V represents the height of the image sensor, in millimeters (mm).

[0060]

[0061] In equations (4) and (5), RE_W represents the scale factor (resolution) of the camera in the width direction, with the unit of mm / pixel; RE_V represents the scale factor (resolution) of the camera in the height direction, with the unit of mm / pixel; PIX_W represents the number of pixels of the camera in the width direction; PIX_V represents the number of pixels of the camera in the height direction;

[0062] S4, inputting parameters such as relative flight height H, pitch angle α of the web of the steel box girder, and route speed v on the three-dimensional “rough model” of the steel box girder to plan the inspection route;

[0063] S5, import the inspection route into the UAV flight control system, set the inspection frequency and time of the UAV, and set the camera scale factor parameters; use the UAV with the aerial survey camera to collect the surface coating photos of the web and bottom plate of the steel box girder;

[0064] S6, collect various types of steel box girder coating disease photos, build a steel box girder coating disease data set, input the data set into the target detection network model, train and optimize the model parameters;

[0065] S7, identifying the steel box girder coating defect photos collected by the drone with the aerial survey camera, and obtaining the apparent defect category and defect area of ​​the steel box girder anti-corrosion coating.

[0066] Exemplarily, in step S2, five-way tilted photography routes, namely, front, back, left, right, and top, are planned in the UAV flight control system to ensure that the routes cover the entire steel box girder.

[0067] Exemplarily, in step S4, the inspection routes include the inspection routes of the left web and left flange plates of the steel box girder, the inspection routes of the bottom plates of the steel box girder, and the refined inspection routes of the right web and right flange plates of the steel box girder.

[0068] Exemplarily, in step S6, the target detection network model is YOLOv8, which improves the detection accuracy through a deeper and more complex network structure and improved training techniques; in addition, YOLOv8 can better detect objects of different sizes by introducing improved multi-scale prediction technology, and can play a better real-time detection feature in the application scenario of this application, and can achieve a high frame rate even on a lower hardware configuration. However, this application does not limit the specific type of the target detection network model, and those skilled in the art can select a suitable detection model according to actual inspection needs.

[0069] Furthermore, after obtaining the apparent disease category and disease area of ​​the anti-corrosion coating of the steel box girder, step S6 also includes the step of evaluating the deterioration scale of the anti-corrosion coating of the steel box girder. Still further, it also includes the step of automatically generating a detection report on the apparent disease of the anti-corrosion coating of the steel box girder. Preferably, according to the provisions of Table 5.2.1-1 and Table 5.2.1-2 of the "Technical Condition Assessment Standard for Highway Bridges" (JTG / T H21-2011), the deterioration scale of the anti-corrosion coating of the web and bottom plate of the steel box girder is assessed by the disease area, and the apparent disease detection report on the anti-corrosion coating of the steel box girder is generated.

[0070] Table 5.2.1-1 Coating Deterioration

[0071]

[0072]

[0073] Table 5.2.1-2 Rust

[0074]

[0075] In a second aspect, the present application provides a system for executing the method for evaluating apparent defects of a steel box girder bridge, comprising:

[0076] A drone terminal, including a drone and an aerial survey camera, is used to remotely perform a steel box girder flight mission and collect photos of the steel box girder web and / or bottom plate and / or flange plate;

[0077] A wireless communication module is used to remotely transmit the photos taken by the drone to the storage module;

[0078] A storage module storing photos of the web plate and / or the bottom plate and / or the flange plate of the steel box girder;

[0079] The calculation module includes a target detection network model, which is used to calculate the relative altitude of the UAV, the proportional factors of the camera in the width and height directions, and identify the surface defect category and defect area of ​​the anti-corrosion coating of the steel box girder.

[0080] Furthermore, the system also includes an evaluation module, which is electrically connected or communicationally connected to the calculation module and is used to evaluate the degradation scale of the anti-corrosion coating of the steel box girder.

[0081] Furthermore, the system also includes a test report generation module, which is electrically connected or communicatively connected to the calculation module for automatically generating a steel box girder anti-corrosion coating surface defect detection report.

[0082] The implementation of the present application is described below in conjunction with a preferred embodiment.

[0083] like Figure 1 As shown, the inspection system of this embodiment includes a drone terminal, a wireless communication module, a storage module, a calculation module, an evaluation module and a detection report generation module.

[0084] The drone terminal includes drones and drone airports. The human-machine interface uses the DJI M300RTK aircraft, which is used to remotely perform steel box girder route flight missions and collect photos of the steel box girder web, bottom plate, and flange plate.

[0085] The wireless communication module is used to remotely transmit the photos taken by the drone to the storage module.

[0086] The storage module can use cloud platforms such as Alibaba Cloud to store a large number of steel box girder photos.

[0087] The calculation module is mainly used to calculate the relative flight height of the UAV, the proportional factor and the apparent defect area of ​​the steel box girder.

[0088] The assessment module is used to evaluate the degradation scale of anti-corrosion coating on steel box girders.

[0089] The inspection report generation module is used to automatically generate the steel box girder anti-corrosion coating surface defect inspection report.

[0090] Taking the main bridge of Chongqi Bridge as an example, the inspection method of this embodiment includes the following contents:

[0091] Step 1: Deploy the drone airport and drones at the steel box girder bridge site;

[0092] Step 2: Plan the five-direction tilt photography routes in DJI Pilot 2: front, back, left, right, and top. Use the DJI M300RTK aircraft with the DJI Zenmuse P1 aerial camera to collect photos of the steel box girder. After shooting, import the photos into DJI Terra for 3D reconstruction to form a 3D "rough model" of the steel box girder, such as Figure 2 shown.

[0093] Step 3: Calculate the relative altitude H of the web of the steel box girder, the camera field of view width FOV_W, the camera field of view height FOV_V, and the camera scale factors RE_W and RE_V in the width and height directions;

[0094] Step 4: Import the 3D rough template of the steel box girder into DJI Terra, input the relative altitude H, the pitch angle α of the steel box girder web, the route speed v and other parameters on the rough template, and plan three types of inspection routes, namely the inspection route of the left web and left flange plate of the steel box girder; the inspection route of the bottom plate of the steel box girder; and the refined inspection route of the right web and right flange plate of the steel box girder, such as Figure 4 shown.

[0095] Step 5: Specify the frequency and time of drone inspections; import the refined inspection routes into DJI Pilot 2, and use the DJI M300RTK aircraft with the DJI Zenmuse P1 aerial camera to collect photos of the surface coating of the web and bottom plates of the steel box girder.

[0096] Step 6: Collect photos of various types of steel box girder paint defects, build a steel box girder paint defect dataset, input the dataset into the YOLOv8 convolutional neural network, train and optimize the model parameters, and deploy it to the cloud platform as a computing module.

[0097] Step 7: The painted photos of the steel box girder are transmitted to the cloud platform’s steel box girder anti-corrosion coating apparent disease storage module and calculation module via 5G. The calculation module transmits the disease category and area to the steel box girder anti-corrosion coating apparent disease assessment module and inspection report module, thereby realizing automatic assessment of the steel box girder apparent disease and automatic generation of inspection reports.

[0098] The module for evaluating the apparent damage of the anti-corrosion coating on steel box girders is based on the provisions of Table 5.2.1-1 and Table 5.2.1-2 of the Technical Condition Assessment Standard for Highway Bridges (JTG / T H21-2011). It evaluates the degradation scale of the anti-corrosion coating on the web and bottom plates of steel box girders through the coating cracking area.

[0099] In summary, this embodiment uses a drone in combination with an unmanned airport to automatically collect bridge images; based on the principle of oblique photography, a three-dimensional model of the bridge is constructed; based on the principle of photogrammetry, the drone shooting distance (that is, the distance from the drone to the shooting target, which is expressed in relative altitude in this application) and the scale factor are calculated to determine the key inspection route for the steel box girder; the type and area of ​​coating defects are identified based on target detection and image segmentation models; the degree of coating deterioration is evaluated based on the "Technical Condition Assessment Standard for Highway Bridges" (JTG / T H21-2011), and a test report is automatically generated, ultimately achieving automated identification and evaluation of coating defects on steel box girders. The implementation process is feasible and effective.

[0100] The above embodiments are only for illustrating the technical concept and features of the present application, and their purpose is to enable technicians familiar with the technical field to understand the content of the present application and implement it accordingly, and they cannot be used to limit the protection scope of the present application. For ordinary technicians in the technical field, they can make several improvements and modifications without departing from the principles of the present application, and these improvements and modifications should also be regarded as the protection scope of the present application.

Claims

1. A method for evaluating the apparent defects of steel box girder bridges based on drone inspection, characterized in that: The steps include: S1, deployment of drone airport and drones at the steel box girder bridge site; S2, plan the photography route in the UAV flight control system, and then use the UAV with the aerial survey camera to collect photos of the steel box girder. After the shooting is completed, the photos are reconstructed in three dimensions to form a three-dimensional "rough mold" of the steel box girder; S3, calculate relative altitude: In formula (1), H represents the relative altitude, that is, the vertical distance between the camera lens and the object being photographed; f represents the focal length of the camera, that is, the distance between the image principal point and the rear node of the camera lens; GSD represents the ground resolution; p represents the pixel size; In formula (2) and formula (3), FOV_W represents the width of the camera field of view; FOV_V represents the height of the camera field of view; WD represents the working distance, that is, the distance from the lens to the target object being photographed, and WD is approximately equal to DC; W represents the width of the image sensor, in millimeters; V represents the height of the image sensor, in millimeters; In formula (4) and formula (5), RE_W represents the scale factor of the camera in the width direction, and the unit is millimeter per pixel; RE_V represents the scale factor of the camera in the height direction, and the unit is millimeter per pixel; PIX_W represents the number of pixels of the camera in the width direction; PIX_V represents the number of pixels of the camera in the height direction; S4, input the relative flight height H, the pitch angle α of the web of the steel box girder, and the route speed v on the three-dimensional "rough mold" of the steel box girder to plan the inspection route; S5, import the inspection route into the UAV flight control system, set the inspection frequency and time of the UAV, and set the camera scale factor parameters; use the UAV with the aerial survey camera to collect the surface coating photos of the web and bottom plate of the steel box girder; S6, collect various types of steel box girder coating disease photos, build a steel box girder coating disease data set, input the data set into the target detection network model, train and optimize the model parameters; S7, identifying the steel box girder coating defect photos collected by the drone with the aerial survey camera, and obtaining the apparent defect category and defect area of ​​the steel box girder anti-corrosion coating.

2. The method for evaluating apparent defects of steel box girder bridges according to claim 1, characterized in that: In step S2, five tilt photography routes, namely, front, back, left, right, and top, are planned in the UAV flight control system.

3. The method for evaluating apparent defects of steel box girder bridges according to claim 1, characterized in that: In step S4, the inspection routes include the inspection routes of the left web and left flange plates of the steel box girder, the inspection routes of the bottom plates of the steel box girder, and the refined inspection routes of the right web and right flange plates of the steel box girder.

4. The method for evaluating apparent defects of steel box girder bridges according to claim 1, characterized in that: In step S6, the target detection network model is YOLOv8.

5. The method for evaluating apparent defects of steel box girder bridges according to claim 1, characterized in that: After obtaining the apparent defect category and defect area of ​​the anti-corrosion coating of the steel box girder, the method also includes a step of evaluating the degradation scale of the anti-corrosion coating of the steel box girder.

6. The method for evaluating apparent defects of steel box girder bridges according to claim 5, characterized in that: After obtaining the apparent disease category and disease area of ​​the anti-corrosion coating of the steel box girder, the method also includes the step of automatically generating an apparent disease detection report of the anti-corrosion coating of the steel box girder.

7. The method for evaluating apparent defects of steel box girder bridges according to claim 6, characterized in that: According to the provisions of the "Technical Condition Assessment Standard for Highway Bridges", the degradation scale of the anti-corrosion coating on the web and bottom plates of steel box girders is assessed by the defect area, and a surface defect inspection report for the anti-corrosion coating on steel box girders is generated.

8. A system for executing the method for evaluating apparent defects of a steel box girder bridge according to any one of claims 1 to 7, characterized in that: include: A drone terminal, including a drone and an aerial survey camera, is used to remotely perform a steel box girder flight mission and collect photos of the steel box girder web and / or bottom plate and / or flange plate; A wireless communication module is used to remotely transmit the photos taken by the drone to the storage module; A storage module storing photos of the web plate and / or the bottom plate and / or the flange plate of the steel box girder; Computing module, deployed on the cloud platform; The calculation module includes a target detection network model, which is used to calculate the relative altitude of the drone, the proportional factors of the camera in the width and height directions, and identify the apparent disease category and disease area of ​​the anti-corrosion coating of the steel box girder.

9. The system according to claim 8, characterized in that Also includes: The evaluation module is electrically connected or communication-connected with the calculation module and is used to evaluate the degradation scale of the anti-corrosion coating of the steel box girder.

10. The system according to claim 8 or 9, characterized in that Also includes: The inspection report generation module is electrically connected or communication-connected to the calculation module, and is used to automatically generate an inspection report on the surface defects of the anti-corrosion coating of the steel box girder.

Images