Automatic inspection and identification system and method for full-bridge cable force of suspension bridge based on sliding rail moving platform and machine vision

Abstract

The invention belongs to the technical field of suspension bridge full-bridge cable force automatic inspection, and particularly relates to a suspension bridge full-bridge cable force automatic inspection recognition system and method based on a sliding rail moving platform and machine vision. The system comprises a fixed rail subsystem which is fixedly installed on the side face of a suspension bridge stiffening beam; the mobile detection vehicle subsystem is movably mounted on the fixed track subsystem and is used for walking autonomously along the track, collecting sling vibration videos and environmental parameters and transmitting the data in real time; and the background data processing and analyzing subsystem is in remote communication connection with the mobile detection vehicle subsystem and is used for receiving, storing and processing data transmitted by the mobile detection vehicle subsystem. According to the invention, a three-level collaborative architecture of'fixed track-mobile detection vehicle-background processing 'is constructed, full-bridge detection coverage is realized through the fixed track subsystem, autonomous movement and accurate detection are completed through the mobile detection vehicle subsystem, data integration and health assessment are realized through the background data processing and analysis subsystem, and the detection efficiency is improved.

Description

Technical Field

[0001] This invention belongs to the field of automatic inspection technology of cable tension of the entire suspension bridge, and specifically relates to an automatic inspection and identification system and method for cable tension of the entire suspension bridge based on a sliding rail moving platform and machine vision. Background Technology

[0002] Suspension bridges, as one of the mainstream forms of long-span bridges, are widely used in transportation construction in complex geographical environments such as rivers, lakes, seas, and deep mountain valleys due to their advantages of strong spanning capacity and reasonable stress distribution. As the core load-bearing component of a suspension bridge, the suspenders play a crucial role in transferring the weight of the stiffening girder and the bridge deck load to the main cable. Their health directly determines the overall structural safety and service life of the bridge. Cable tension is a core indicator reflecting the health status of the suspenders. During long-term service, suspenders are susceptible to damage such as tension decay, loosening, and wire breakage due to alternating loads, environmental corrosion (such as wind and rain erosion and salt spray corrosion), material aging, and fatigue damage. Failure to detect and address abnormal cable tension in suspenders in a timely manner can lead to load imbalance, triggering a chain reaction of damage to the main cable, stiffening girder, and other components, and in severe cases, even causing major safety accidents such as bridge collapse. Therefore, regularly and accurately testing the cable tension of all suspenders in a suspension bridge is a critical step in ensuring the safe operation and maintenance of the bridge structure.

[0003] Currently, cable tension detection for suspension bridges has become a key research focus and engineering challenge in the field of bridge health monitoring. With the continuous increase in traffic volume and the aging of bridges, increasingly higher demands are being placed on the accuracy, efficiency, automation, and safety of cable tension detection. Traditional detection methods and existing technical solutions have gradually revealed many limitations in practical engineering applications, making it difficult to meet the needs of efficient, accurate, and automated inspection of the entire bridge's cable tension. Therefore, developing an automated inspection and identification system for the entire bridge's cable tension, featuring autonomous movement, accurate identification, real-time data uploading, and visual analysis, is urgently needed. Currently, suspension bridge cable tension detection technologies are mainly divided into two categories: contact detection and non-contact detection. Each type of technology has significant shortcomings in practical applications, as detailed below: 1. Contact-based testing technologies (such as vibration method, pressure sensor method, magnetic flux method): ① Cumbersome operation and extremely low testing efficiency. These methods require testing personnel to use aerial work platforms (such as bridge inspection vehicles or suspended platforms) to approach the suspension cables and manually install sensors or testing devices. For large-span suspension bridges with hundreds of suspension cables, the entire bridge inspection cycle can take several days or even weeks, seriously affecting normal bridge traffic; ② Poor safety. The high-altitude working environment is complex and easily affected by natural environmental factors such as wind and precipitation, posing safety hazards such as personnel falls and equipment drops; ③ Testing accuracy is greatly affected by human operation. Differences in installation positions and operating techniques among different testing personnel may lead to large dispersion in testing data, making it difficult to guarantee data consistency and reliability; ④ Real-time dynamic testing cannot be achieved. The testing process requires traffic interruption or vehicle flow restriction, causing significant disruption to traffic.

[0004] 2. Traditional non-contact detection technologies (such as machine vision inspection based on fixed cameras and laser Doppler vibration detection): ① Limited detection range: Fixed cameras can only cover a few cables in a specific area. To achieve full-bridge inspection, a large number of cameras need to be deployed on the bridge, resulting in high equipment costs, difficult installation and maintenance, and susceptibility to obstruction by the bridge structure, leading to blind spots. ② Poor environmental adaptability: Laser Doppler detection is easily affected by environmental factors such as atmospheric turbulence, dust, rain, and snow, significantly reducing detection accuracy. Traditional machine vision detection algorithms are sensitive to changes in lighting and cannot function properly at night or in low-light conditions. They also struggle to effectively extract minute vibration signals from the cables, resulting in low cable force identification accuracy. ③ Lack of autonomous movement and obstacle avoidance capabilities: They cannot autonomously adjust their detection position according to the cable distribution and cannot avoid temporary obstacles on the bridge (such as construction equipment and fallen objects), easily causing equipment damage or detection interruption. ④ Low data processing and visualization: Detection data requires manual processing and analysis and cannot be uploaded to the backend in real time for full-bridge cable force integration and health assessment, making it difficult to quickly locate cables with abnormal cable force.

[0005] 3. Existing mobile inspection platform technologies (such as drone inspection: for example, patent CN 119223504 B): ① Limited endurance: Drones have short single-flight times, making it difficult to continuously inspect the cables of a long-span suspension bridge. Frequent battery replacements or charging are required, resulting in low inspection efficiency. ② Poor stability: Drones are greatly affected by wind and cannot accurately hover at the cable inspection position, leading to fuzzy vibration signal acquisition and large errors in cable force identification. ③ High operational requirements: Drones require remote control by professional operators and are prone to collisions in the complex structural environment of bridges, resulting in insufficient safety. ④ Limited data transmission: Wireless signals are easily interfered with in high-altitude environments, making it difficult to achieve real-time uploading of inspection data and synchronous analysis in the background.

[0006] In summary, the technical shortcomings are as follows: 1. Low level of automation: Existing visual inspection systems mostly require manual intervention and cannot achieve fully automated continuous inspection; 2. Insufficient spatial coverage: It is difficult to obtain vibration data of all suspension cables of the entire bridge in a short time and at low cost; 3. Weak anti-interference ability: Factors such as bridge environmental vibration, vehicle traffic, and changes in lighting can easily affect the recognition accuracy; 4. Low system integration: Data acquisition, processing, and visualization are often separated, lacking an integrated intelligent platform; 5. Poor scalability: Existing systems are difficult to adapt to suspension bridges with different spans and cable spacings. Summary of the Invention

[0007] To address the aforementioned issues, the purpose of this invention is to provide an automatic inspection and identification system and method for the cable tension of a suspension bridge based on a sliding rail mobile platform and machine vision. This invention constructs a three-level collaborative architecture of "fixed track - mobile inspection vehicle - back-end processing". The fixed track subsystem achieves full bridge inspection coverage, the mobile inspection vehicle subsystem enables autonomous movement and accurate inspection, and the back-end data processing and analysis subsystem achieves data integration and health assessment, thereby comprehensively improving inspection efficiency.

[0008] The technical solution of this invention is: an automatic inspection and identification system for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision, comprising: The fixed track subsystem is fixedly installed on the side of the stiffening girder of the suspension bridge, providing a running track and full bridge space coverage for mobile inspection; The mobile inspection vehicle subsystem is movably installed on the fixed track subsystem for autonomously moving along the track, collecting video of cable vibration and environmental parameters, and transmitting the data in real time. The background data processing and analysis subsystem is remotely connected to the mobile inspection vehicle subsystem. It is used to receive, store, and process the data transmitted by the mobile inspection vehicle subsystem, analyze and calculate the cable force through algorithms, and perform visualization and health status assessment. The fixed track subsystem, the mobile inspection vehicle system, and the background data processing and analysis subsystem work together to form a three-level automatic inspection and identification architecture.

[0009] The fixed track subsystem includes two parallel tracks laid on the sides of the stiffening girder of the suspension bridge. The length of the tracks covers the inspection area of ​​all the suspension cables of the entire bridge. The cross-section of the tracks is "I"-shaped, and the inner side is provided with guide grooves and racks. The guide grooves are used to guide the movement of the mobile inspection vehicle, and the racks are used to mesh with the drive gears of the mobile inspection vehicle. Safety docking platforms are provided at both ends of the tracks for parking, charging and maintenance of the mobile inspection vehicle.

[0010] The mobile inspection vehicle system includes a vehicle body and the following modules integrated on the vehicle body: The drive and walking module includes a drive motor, a gear set meshing with the track rack, four guide wheels embedded in the guide groove, and an electromagnetic braking unit. A transmission chain is provided between adjacent guide wheels. The gear set meshes with the transmission chain to drive the vehicle body to move along the track and achieve precise stopping. The high-precision positioning module integrates a GPS / BeiDou dual-mode satellite positioning unit and a laser ranging sensor, which is used to acquire and fuse the absolute position and relative displacement information of the vehicle body in real time; The environmental perception and obstacle avoidance module includes an environmental sensor group for collecting parameters such as wind speed, temperature, humidity, and light intensity, as well as a composite obstacle avoidance unit composed of lidar and ultrasonic sensors, used to perceive environmental conditions and detect obstacles on the travel path. The visual acquisition module includes a multi-angle adjustable bracket, a high-definition industrial camera mounted on the bracket, and a night lighting unit that works with the camera to acquire vibration video sequences of targets attached to slings. The data transmission module adopts a 5G communication module and has a built-in data cache memory to realize bidirectional data communication between the mobile detection vehicle subsystem and the background data processing and analysis subsystem; The energy module, including the main lithium battery pack and an auxiliary solar charging panel located on the top of the vehicle, is used to provide power to the entire mobile inspection vehicle subsystem and manage the charging and discharging process.

[0011] The background data processing and analysis subsystem includes: The data receiving and storage server adopts a dual backup architecture of cloud and local, which is used to receive and classify the vibration video, location information, and environmental parameter data from the mobile inspection vehicle system, as well as store bridge structural parameters and historical inspection data; The advanced cable force analysis algorithm library integrates a high-efficiency phase motion amplification algorithm module, a cable force calculation correction model module, and a trend analysis algorithm module. It is used to process received vibration videos to extract cable frequencies, calculate and correct cable force values, and analyze long-term cable force trends. The bridge cable force visualization and health assessment platform provides a human-computer interaction interface, supports the display of the bridge cable force distribution in the form of two-dimensional heat maps and three-dimensional models, and has functions such as automatic comparison and early warning, trend curve generation, automatic generation of detection reports, and remote control.

[0012] The high-efficiency phase motion amplification algorithm module is specifically used for: converting RGB video to YIQ color space and extracting luminance components; performing dual-tree complex wavelet transform (DTCWT) decomposition on the luminance signal; amplifying the phase information of a specific frequency band in the complex domain; reconstructing the amplified coefficients using DTCWT to obtain the amplified luminance signal; recombining the processed luminance signal with the original chrominance signal and converting it back to RGB space to obtain the motion-amplified video; and extracting the inherent vibration frequency of the sling through spectral analysis of the amplified video.

[0013] The cable force calculation and correction model module establishes the relationship between cable force and temperature based on the coefficient of linear expansion and the temperature coefficient of elastic modulus. For a single cable, the cable force correction value caused by temperature change is... F corr Represented as: F corr =F meas - E⋅A⋅α⋅ Δ T In the formula: F meas The cable force value is the reference temperature of 20℃. E The elastic modulus of the material; A Let be the cross-sectional area of ​​the sling; α High-strength steel wire with a high coefficient of linear expansion for sling material α ≈ 1.2×10 −5 / ℃;Δ T This represents the change between the measured temperature and the reference temperature.

[0014] The trend analysis algorithm module employs a dual early warning mechanism: trend slope warning and cable force absolute value warning. Trend Slope Warning: Based on historical cable force attenuation slope statistics of healthy slings, a primary warning threshold and a secondary warning threshold are set. The primary warning threshold... When the overall attenuation slope of the target sling When the attenuation rate of a healthy sling exceeds 95%, a Level 1 trend warning is triggered; the Level 2 warning threshold... ,when When the decay rate is deemed abnormal, a secondary trend warning is triggered. The instantaneous attenuation slope is used to statistically analyze the linear attenuation slope of healthy suspenders installed in the same batch on the same bridge during the steady attenuation phase. Its average value is Standard deviation is An early warning is triggered when the instantaneous or fitted decay slope of the target sling exceeds the corresponding threshold. Absolute cable force warning: The first-level warning threshold is set at 95% of the design cable force value of the sling, and the second-level warning threshold is 90% of the design cable force value; the warning is triggered when the measured cable force value of the target sling is lower than the corresponding threshold.

[0015] The target affixed to the sling is a regular geometric shape with high contrast and is coated with an anti-reflective coating.

[0016] An automatic inspection and identification method for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision is disclosed. The method employs the aforementioned automatic inspection and identification system for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision, and includes the following steps: S1: System initialization, the inspection task is started through the background data processing and analysis subsystem, the sling database is loaded, and the mobile inspection vehicle subsystem is instructed to perform a self-check and then depart from the safety dock; S2: The mobile inspection vehicle system moves autonomously along a fixed track to the target sling position according to the preset inspection sequence, and monitors the environment and path safety in real time through the environmental perception and obstacle avoidance module; S3: After reaching the designated location, the mobile detection vehicle subsystem accurately positions itself and brakes, adjusts the vision acquisition module to align with the sling target, and collects the current environmental parameters; S4: The vision acquisition module acquires vibration videos of the suspension target and uploads them to the background data processing and analysis subsystem via the data transmission module; S5: The background data processing and analysis subsystem calls the advanced cable force analysis algorithm library to perform motion amplification and spectrum analysis on the vibration video to extract the frequency, and calculates the corrected cable force value by combining the cable force calculation correction model. S6: Display the calculated cable force value on the visualization platform and call the trend analysis algorithm to assess the health status and make early warning judgments; S7: After completing the inspection of the current sling, the mobile inspection vehicle subsystem unlocks and moves to the next sling, repeating steps S3 to S6 until the inspection of all slings of the entire bridge is completed; S8: After the full bridge inspection is completed, the mobile inspection vehicle subsystem automatically returns to the safe docking station, and the background data processing and analysis subsystem generates and outputs a complete inspection report.

[0017] During the movement in steps S2 and S7, if the environmental perception and obstacle avoidance module detects that the wind speed continuously exceeds 8m / s or identifies an obstacle in front, the mobile detection vehicle will automatically perform operations such as pausing, waiting, or returning to the safe docking platform.

[0018] The technical advantages of this invention are as follows: 1. This invention utilizes three core subsystems: a fixed track subsystem, a mobile inspection vehicle subsystem, and a backend data processing and analysis subsystem. The modules within each subsystem work collaboratively, and the subsystems interact through data transmission links to achieve information exchange, jointly completing the automatic, accurate, and efficient inspection and identification of the cable force of the entire bridge. 2. This invention integrates efficient phase motion amplification algorithms and machine vision technology, combined with the high-precision positioning capability of the mobile inspection vehicle subsystem, to accurately extract minute vibration signals of the cables, achieving high-precision identification of cable force. 3. This invention enhances the environmental adaptability of the mobile inspection vehicle subsystem, integrating modules for environmental perception, obstacle avoidance, nighttime supplementary lighting, and stable energy supply to ensure stable operation in complex environments. 4. This invention, through the advanced algorithms and visualization platform of the backend data processing and analysis subsystem, achieves real-time processing of inspection data, anomaly warning, historical comparison, and trend prediction, providing accurate data support for bridge operation and maintenance decisions. 5. This invention requires no manual intervention throughout the entire process, achieving automation, intelligence, and unmanned operation of the entire bridge cable force inspection, reducing inspection costs, ensuring the safety of inspection personnel, and minimizing interference with traffic.

[0019] The following will provide further explanation in conjunction with the accompanying drawings. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the composition structure of an automatic inspection and identification system for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision, according to the present invention.

[0021] Figure 2 This is a schematic diagram of the fixed track subsystem of the present invention.

[0022] Figure 3 This is a schematic diagram of the guide groove and rack of the present invention.

[0023] Figure 4 This is a schematic diagram of the structure of the mobile detection vehicle subsystem of the present invention.

[0024] Figure 5 This is a flowchart illustrating an automatic inspection and identification method for the cable force of a suspension bridge based on a sliding rail moving platform and machine vision, according to the present invention.

[0025] Figure 6 This is a flowchart of the efficient phase motion amplification algorithm of the present invention.

[0026] Reference numerals: 1-Fixed track subsystem; 2-Mobile inspection vehicle subsystem; 3-Back-end data processing and analysis subsystem; 11-Track; 12-Safety docking platform; 111-Guide groove; 112-Rack; 21-Vehicle body; 22-Drive and walking module; 23-High-precision positioning module; 24-Environmental perception and obstacle avoidance module; 25-Vision acquisition module; 26-Data transmission module; 27-Energy module; 221-Drive motor; 222-Gear set; 223-Guide wheel; 224-Electromagnetic braking unit; 225-Transmission chain; 241-Environmental sensor group; 242-LiDAR; 243-Ultrasonic sensor; 251-Bracket; 252-High-definition industrial camera; 253-Nighttime supplementary lighting unit; 271-Main lithium battery pack; 272-Auxiliary solar charging panel. Detailed Implementation Example 1

[0027] like Figures 1-4 As shown, an automatic inspection and identification system for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision includes: Fixed track subsystem 1 is fixedly installed on the side of the stiffening girder of the suspension bridge, providing a running track and full bridge space coverage for mobile detection; The mobile detection vehicle subsystem 2 is movably installed on the fixed track subsystem 1 and is used to autonomously move along the track, collect video of the sling vibration and environmental parameters, and transmit the data in real time. The background data processing and analysis subsystem 3 is remotely connected to the mobile inspection vehicle subsystem 2. It is used to receive, store, and process the data transmitted by the mobile inspection vehicle subsystem 2, analyze and calculate the cable force through algorithms, and perform visualization and health status assessment. The fixed track subsystem 1, the mobile detection vehicle system 2, and the background data processing and analysis subsystem 3 work together to form a three-level automatic inspection and identification architecture.

[0028] In operation, after system startup, the fixed track subsystem 1 is installed along both sides of the stiffening girder of the suspension bridge, providing a track for the mobile inspection vehicle to cover the entire bridge. The mobile inspection vehicle subsystem 2 autonomously moves along the track to the location of each cable, collecting vibration videos and environmental data, which are then transmitted in real time to the backend data processing and analysis subsystem 3 via a 5G network. The backend system processes, analyzes, calculates cable force, and performs health assessments on the data, ultimately displaying the cable force distribution and early warning information for the entire bridge on a visualization platform. This invention achieves automated and continuous inspection of the cable force of the entire bridge without manual intervention; the three-level collaborative architecture enhances the system's integration and intelligence; it is applicable to suspension bridges with different spans and cable spacings and has good scalability. Example 2

[0029] Based on Embodiment 1, in this embodiment, preferably, the fixed track subsystem 1 includes two parallel tracks 11 arranged on the side of the stiffening girder of the suspension bridge. The length of the tracks 11 covers the inspection area of ​​all the suspension cables of the entire bridge. The cross-section of the tracks 11 is "I"-shaped, and guide grooves 111 and racks 112 are provided on their inner sides. The guide grooves 111 are used to guide the mobile inspection vehicle to move, and the racks 112 are used to mesh with the drive gear of the mobile inspection vehicle. Safety docking platforms 12 are provided at both ends of the tracks 11 for parking, charging and maintenance of the mobile inspection vehicle.

[0030] In use, the track 11 is laid parallel to the side of the stiffening beam, covering the entire bridge cable inspection area. The mobile inspection vehicle is precisely guided and driven by the guide groove 111 and the rack 112. After the inspection task is completed, the mobile inspection vehicle returns to the safe docking platform 12 for charging and maintenance. The track structure of this invention is stable, has strong wind vibration resistance, and is adaptable to the complex environment of bridges; the guide and rack design ensures smooth movement and precise stopping of the mobile inspection vehicle; the safe docking platform provides integrated parking, charging, and maintenance functions, improving the system's endurance and availability. Example 3

[0031] Based on Embodiment 1, in this embodiment, preferably, the mobile detection vehicle subsystem 2 includes a vehicle body 21 and the following modules integrated on the vehicle body 21: The drive and walking module 22 includes a drive motor 221, a gear set 222 meshing with the track rack 112, four guide wheels 223 embedded in the guide groove 111, and an electromagnetic braking unit 224. A transmission chain 225 is provided between adjacent guide wheels 223. The gear set 222 meshes with the transmission chain 225 to drive the vehicle body 21 to move along the track 11 and achieve precise stopping. The high-precision positioning module 23 integrates a GPS / BeiDou dual-mode satellite positioning unit and a laser ranging sensor, which is used to acquire and fuse the absolute position and relative displacement information of the vehicle body 21 in real time. The environmental perception and obstacle avoidance module 24 includes an environmental sensor group 241 for collecting parameters such as wind speed, temperature and humidity and light intensity, and a composite obstacle avoidance unit composed of a lidar 242 and an ultrasonic sensor 243, for sensing environmental conditions and detecting obstacles on the travel path. The visual acquisition module 25 includes a multi-angle adjustable bracket 251, a high-definition industrial camera 252 mounted on the bracket, and a night lighting unit 253 that works with the camera to acquire vibration video sequences of a target attached to a sling. The data transmission module 26 adopts a 5G communication module and has a built-in data cache memory to realize bidirectional data communication between the mobile detection vehicle subsystem 2 and the background data processing and analysis subsystem 3. The energy module 27, including the main lithium battery pack 271 and the auxiliary solar charging panel 272 located on the top of the vehicle body, is used to provide power to the entire mobile inspection vehicle subsystem 2 and manage the charging and discharging process.

[0032] In use, the mobile inspection vehicle travels along the track to the target sling position, and the drive and walking module 22 achieves precise stopping; the high-precision positioning module 23 integrates GPS and laser ranging data to provide real-time location information; the environmental perception and obstacle avoidance module 24 monitors wind speed and obstacles to ensure driving safety; the visual acquisition module 25 adjusts the camera angle to acquire vibration video of the sling target; the data transmission module 26 uploads data in real time; and the energy module 27 uses solar-assisted charging to ensure long-term inspection. This invention utilizes multiple modules working collaboratively to achieve autonomous movement, precise data acquisition, and real-time transmission; the environmental perception and obstacle avoidance functions enhance the system's adaptability and safety in complex bridge environments; and solar-assisted power supply extends the operating time, supporting continuous full-bridge inspection. Example 4

[0033] Based on Embodiment 1, in this embodiment, preferably, the background data processing and analysis subsystem 3 includes: The data receiving and storage server 31 adopts a dual backup architecture of cloud and local, and is used to receive and classify the vibration video, location information, environmental parameter data from the mobile inspection vehicle subsystem 2, as well as store bridge structural parameters and historical inspection data. The Advanced Cable Force Analysis Algorithm Library 32 integrates a high-efficiency phase motion amplification algorithm module, a cable force calculation correction model module, and a trend analysis algorithm module. It is used to process received vibration videos to extract cable frequencies, calculate and correct cable force values, and analyze long-term cable force trends. The whole bridge cable force visualization and health assessment platform 33 provides a human-computer interaction interface, supports the display of the whole bridge cable force distribution in the form of two-dimensional heat map and three-dimensional model, and has the functions of automatic comparison and early warning, trend curve generation, automatic generation of detection report and remote control.

[0034] In this invention, the data receiving and storage server 31 receives video, location, and environmental data uploaded by the mobile inspection vehicle and performs dual backup storage. The advanced cable force analysis algorithm library 32 calls the phase motion amplification algorithm to extract frequencies and calculates cable forces in combination with the correction model. The full-bridge cable force visualization and health assessment platform 33 displays the cable force distribution in the form of heat maps, 3D models, etc., and automatically generates early warning reports. This invention achieves integrated real-time data processing, storage, and analysis; the algorithm library supports high-precision cable force extraction and temperature correction, improving data reliability; and the visualization platform facilitates maintenance personnel to quickly grasp the full-bridge cable force status and abnormal locations. Example 5

[0035] Based on Example 1, in this example, as Figure 6 As shown, preferably, the high-efficiency phase motion amplification algorithm module is specifically used for: converting RGB video to YIQ color space and extracting luminance components; performing dual-tree complex wavelet transform (DTCWT) decomposition on the luminance signal; amplifying the phase information of a specific frequency band in the complex domain; reconstructing the amplified coefficients using DTCWT to obtain the amplified luminance signal; recombining the processed luminance signal with the original chrominance signal and converting it back to RGB space to obtain the motion-amplified video; and extracting the inherent vibration frequency of the sling through spectral analysis of the amplified video.

[0036] In this invention, the system converts the acquired RGB video to the YIQ color space, extracts the luminance component, and performs dual-tree complex wavelet transform (DTCWT) decomposition. Phase information in specific frequency bands is amplified, and the reconstructed video yields a motion-amplified video. The inherent frequency of the sling is then extracted through spectral analysis. This invention significantly enhances the accuracy of minute vibration signals from the sling; the algorithm adapts to different lighting and weather conditions, exhibiting strong robustness; and it supports real-time processing, meeting inspection efficiency requirements.

[0037] To overcome the drawbacks of traditional phase-Euler-based video amplification algorithms, which suffer from strong parameter dependence and computational complexity, the DTCWT technique is introduced. DTCWT is a highly efficient signal processing method based on a binary tree structure. Its core idea is to improve the directional resolution and reconstruction quality of the signal through two parallel filter banks. Compared with traditional discrete wavelet transform, DTCWT has two significant advantages: approximate translation invariance and excellent multi-directional selectivity. These characteristics enable it to effectively suppress frequency aliasing, a common phenomenon in traditional wavelet transform, and significantly reduce information loss during signal processing. Therefore, it has been widely used in signal denoising, edge detection, and image enhancement. The DTCWT system structure consists of two parallel filter trees, which generate the real and imaginary parts of the signal, respectively. During decomposition, each filter tree achieves fine analysis of the signal at multiple scales and directions by alternately applying high-pass and low-pass filters. It is particularly noteworthy that the two filter trees maintain a strict phase relationship during signal processing: the imaginary tree maintains a delay interval of one sample value relative to the real tree. This design ensures that the sampling points of the imaginary part tree are precisely located at the midpoint of adjacent sampling points in the real part tree, thus forming a complementary information mechanism between the real and imaginary parts. This unique structural design not only endows DTCWT with the important property of approximate translation invariance but also significantly improves the computational efficiency of the entire system. Simultaneously, to optimize the computational efficiency and reduce the complexity of the PBVM algorithm, a luminance Y-channel processing mechanism is employed for efficient motion magnification. Specifically, after converting the image from the RGB color space to the YIQ space, the algorithm extracts only the luminance component Y-channel for processing, while retaining the chroma information, leaving the I and Q channels unchanged. Next, only the luminance channel undergoes relevant processing steps, including spatial domain decomposition, temporal filtering, smoothing, motion magnification, and video reconstruction. After these processes, the enhanced luminance information is recombinated with the original chroma information to restore the complete YIQ representation. Finally, by converting the result back from YIQ to RGB space, the final motion-magnified color video is generated. This mechanism significantly reduces computational complexity while ensuring the motion magnification effect and effectively preserving color realism and consistency. Example 6

[0038] Based on Example 1, in this embodiment, preferably, the cable force calculation correction model module establishes the relationship between cable force and temperature based on the coefficient of linear expansion and the temperature coefficient of elastic modulus. For a single cable, the cable force correction value caused by temperature change is... F corr Represented as: F corr =F meas - E⋅A⋅α⋅ Δ T In the formula:F meas The cable force value is the reference temperature of 20℃. E The elastic modulus of the material; A Let be the cross-sectional area of ​​the sling; α High-strength steel wire with a high coefficient of linear expansion for sling material α ≈ 1.2×10 −5 / ℃;Δ T This represents the change between the measured temperature and the reference temperature.

[0039] When used in this invention, the system corrects the cable force measurement value at the reference temperature based on the cable material parameters (elastic modulus E, cross-sectional area A, coefficient of linear expansion α) and the measured temperature ΔT to obtain the true cable force value. This invention eliminates the influence of temperature changes on cable force measurement, improving data accuracy; the model is simple and practical, suitable for actual engineering applications; and it supports long-term data comparison and trend analysis. Example 7

[0040] Based on Example 1, in this embodiment, preferably, the trend analysis algorithm module adopts a dual early warning mechanism of trend slope warning and cable force absolute value warning: Trend Slope Warning: Based on historical cable force attenuation slope statistics of healthy slings, a primary warning threshold and a secondary warning threshold are set. The primary warning threshold... When the overall attenuation slope of the target sling When the attenuation rate of a healthy sling exceeds 95%, a Level 1 trend warning is triggered; the Level 2 warning threshold... ,when When the decay rate is deemed abnormal, a secondary trend warning is triggered. The instantaneous attenuation slope is used to statistically analyze the linear attenuation slope of healthy suspenders installed in the same batch on the same bridge during the steady attenuation phase. Its average value is Standard deviation is An early warning is triggered when the instantaneous or fitted decay slope of the target sling exceeds the corresponding threshold. Absolute cable force warning: The first-level warning threshold is set at 95% of the design cable force value of the sling, and the second-level warning threshold is 90% of the design cable force value; the warning is triggered when the measured cable force value of the target sling is lower than the corresponding threshold.

[0041] In use, this invention calculates the mean and standard deviation of the attenuation slope based on historical data of healthy slings, and sets primary and secondary warning thresholds. When the attenuation slope of the target sling or the measured sling force exceeds the threshold, the system automatically triggers an warning. This invention achieves early identification and warning of abnormal sling force; the dual warning mechanism improves the accuracy and reliability of warnings; and it supports long-term health status assessment and maintenance decisions. Example 8

[0042] Based on Example 1, in this example, preferably, the target attached to the sling is a regular geometric shape with high contrast and the surface is coated with an anti-reflective coating.

[0043] In use, this invention involves attaching high-contrast, anti-reflective geometric targets to each suspension cable, facilitating clear identification and tracking by the visual acquisition module under complex lighting conditions. This invention improves image recognition stability and accuracy; the anti-reflective coating reduces light interference, making it suitable for all-weather detection; and the target structure is simple, making installation and maintenance convenient. Example 9

[0044] like Figure 5 As shown, an automatic inspection and identification method for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision is described above. The method includes the following steps: S1: System initialization, the inspection task is started through the background data processing and analysis subsystem 3, the sling database is loaded, and the mobile inspection vehicle subsystem 2 is instructed to perform a self-inspection and then depart from the safety dock 12; S2: The mobile inspection vehicle subsystem 2 moves autonomously along the fixed track 11 to the target sling position according to the preset inspection sequence, and monitors the environment and path safety in real time through the environmental perception and obstacle avoidance module 24; S3: After reaching the designated location, the mobile detection vehicle subsystem 2 accurately positions and brakes, adjusts the vision acquisition module 25 to align with the sling target, and collects the current environmental parameters; S4: The visual acquisition module 25 acquires the vibration video of the suspension target and uploads it to the background data processing and analysis subsystem 3 through the data transmission module 26; S5: The background data processing and analysis subsystem 3 calls the advanced cable force analysis algorithm library 32 to perform motion amplification processing and spectrum analysis on the vibration video to extract the frequency, and calculates the corrected cable force value by combining the cable force calculation correction model. S6: Display the calculated cable force value on the visualization platform 33, and call the trend analysis algorithm to assess the health status and make early warning judgments; S7: After completing the inspection of the current sling, the mobile inspection vehicle subsystem 2 is unlocked and moved to the next sling. Repeat steps S3 to S6 until the inspection of all slings of the entire bridge is completed. S8: After the full bridge inspection is completed, the mobile inspection vehicle subsystem 2 automatically returns to the safe docking station 12, and the background data processing and analysis subsystem 3 generates and outputs a complete inspection report.

[0045] During the movement in steps S2 and S7, if the environmental perception and obstacle avoidance module 24 detects that the wind speed continuously exceeds 8m / s or identifies an obstacle in front, the mobile detection vehicle 2 will automatically perform operations such as pausing, waiting, or returning to the safe docking platform 12.

[0046] When in use, the system executes inspection tasks according to a preset process: initialization → movement to the sling position → precise positioning → acquisition of video and environmental data → uploading and processing → sling force calculation and evaluation → movement to the next sling → completion of full bridge inspection → generation of report. This invention automates the process, reducing human error; supports continuous full bridge inspection, improving inspection efficiency; and possesses the ability to autonomously handle abnormal situations, such as exceeding wind speed limits or obstacle recognition.

[0047] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. An automatic inspection and identification system for the cable tension of a suspension bridge based on a sliding rail moving platform and machine vision, characterized in that: include: The fixed track subsystem (1) is fixedly installed on the side of the stiffening girder of the suspension bridge to provide a running track and full bridge space coverage for mobile detection; The mobile detection vehicle subsystem (2) is movably installed on the fixed track subsystem (1) for autonomously moving along the track, collecting video of cable vibration and environmental parameters, and transmitting the data in real time; The background data processing and analysis subsystem (3) is remotely connected to the mobile inspection vehicle subsystem (2) for receiving, storing and processing the data transmitted by the mobile inspection vehicle subsystem (2), calculating the cable force through algorithm analysis, and performing visualization and health status assessment. The fixed track subsystem (1), the mobile inspection vehicle subsystem (2), and the background data processing and analysis subsystem (3) work together to form a three-level automatic inspection and identification architecture.

2. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 1, characterized in that, The fixed track subsystem (1) includes two parallel tracks (11) laid on the side of the stiffening girder of the suspension bridge. The length of the track (11) covers the inspection area of ​​all the suspension cables of the entire bridge. The cross section of the track (11) is "I" shaped. A guide groove (111) and a rack (112) are provided on its inner side. The guide groove (111) is used to guide the mobile inspection vehicle to move. The rack (112) is used to mesh with the drive gear of the mobile inspection vehicle. Safety docking platforms (12) are provided at both ends of the track (11) for parking, charging and maintenance of the mobile inspection vehicle.

3. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 2, characterized in that, The mobile detection vehicle subsystem (2) includes a vehicle body (21) and the following modules integrated on the vehicle body (21): The drive and walking module (22) includes a drive motor (221), a gear set (222) meshing with the track rack (112), four guide wheels (223) embedded in the guide groove (111), and an electromagnetic braking unit (224). A transmission chain (225) is provided between the front and rear adjacent guide wheels (223). The gear set (222) meshes with the transmission chain (225) to drive the vehicle body (21) to move along the track (11) and achieve precise stopping. The high-precision positioning module (23) integrates a GPS / BeiDou dual-mode satellite positioning unit and a laser ranging sensor, which is used to acquire and fuse the absolute position and relative displacement information of the vehicle body (21) in real time; The environmental perception and obstacle avoidance module (24) includes an environmental sensor group (241) for collecting parameters such as wind speed, temperature and humidity and light intensity, and a composite obstacle avoidance unit composed of lidar (242) and ultrasonic sensor (243) for sensing environmental conditions and detecting obstacles on the travel path. The visual acquisition module (25) includes a multi-angle adjustable bracket (251), a high-definition industrial camera (252) mounted on the bracket, and a night lighting unit (253) that works with the camera to acquire a vibration video sequence of a target attached to a sling; The data transmission module (26) adopts a 5G communication module and has a built-in data cache memory to realize bidirectional data communication between the mobile detection vehicle subsystem (2) and the background data processing and analysis subsystem (3); The energy module (27), including the main lithium battery pack (271) and the auxiliary solar charging panel (272) located on the top of the vehicle body, is used to provide power to the entire mobile inspection vehicle subsystem (2) and manage the charging and discharging process.

4. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 3, characterized in that, The background data processing and analysis subsystem (3) includes: The data receiving and storage server (31) adopts a dual backup architecture of cloud and local, and is used to receive and classify the vibration video, location information, environmental parameter data from the mobile inspection vehicle subsystem (2), as well as store bridge structure parameters and historical inspection data; Advanced cable force analysis algorithm library (32) integrates a high-efficiency phase motion amplification algorithm module, a cable force calculation correction model module and a trend analysis algorithm module, which are used to process the received vibration video to extract the cable frequency, calculate and correct the cable force value, and analyze the long-term trend of cable force change; The whole bridge cable force visualization and health assessment platform (33) provides a human-computer interaction interface, supports the display of the whole bridge cable force distribution in the form of two-dimensional heat map and three-dimensional model, and has the functions of automatic comparison and early warning, trend curve generation, automatic generation of detection report and remote control.

5. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 4, characterized in that, The high-efficiency phase motion amplification algorithm module is specifically used for: converting RGB video to YIQ color space and extracting luminance components; performing dual-tree complex wavelet transform (DTCWT) decomposition on the luminance signal; and amplifying the phase information of a specific frequency band in the complex domain. The amplified coefficients are reconstructed using dual-tree complex wavelet transform (DTCWT) to obtain the amplified luminance signal; the processed luminance signal is then recombined with the original chrominance signal and converted back to RGB space to obtain the motion magnified video; the natural vibration frequency of the sling is extracted through spectral analysis of the magnified video.

6. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 4, characterized in that, The cable force calculation and correction model module establishes the relationship between cable force and temperature based on the coefficient of linear expansion and the temperature coefficient of elastic modulus. For a single cable, the cable force correction value caused by temperature change is... F corr Represented as: F corr =F meas - E⋅A⋅α⋅ D T In the formula: F meas The cable force value is the reference temperature (20℃). E The elastic modulus of the material; A Let be the cross-sectional area of ​​the sling; α The coefficient of linear expansion of the sling material (high-strength steel wire) α ≈ 1.2×10 −5 / ℃); Δ T This represents the change between the measured temperature and the reference temperature.

7. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 4, characterized in that, The trend analysis algorithm module employs a dual early warning mechanism: trend slope warning and cable force absolute value warning. Trend Slope Warning: Based on historical cable force attenuation slope statistics of healthy slings, a primary warning threshold and a secondary warning threshold are set. The primary warning threshold... When the overall attenuation slope of the target sling When the attenuation rate of a healthy sling is determined to be faster than 95%, a Level 1 trend warning is triggered; the Level 2 warning threshold... ,when When the decay rate is deemed abnormal, a secondary trend warning is triggered. The instantaneous attenuation slope is used to statistically analyze the linear attenuation slope of healthy suspenders installed in the same batch on the same bridge during the steady attenuation phase. Its average value is Standard deviation is An early warning is triggered when the instantaneous or fitted decay slope of the target sling exceeds the corresponding threshold. Absolute cable force warning: The first-level warning threshold is set at 95% of the design cable force value of the sling, and the second-level warning threshold is 90% of the design cable force value; the warning is triggered when the measured cable force value of the target sling is lower than the corresponding threshold.

8. The automatic inspection and identification system for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 3, characterized in that, The target affixed to the sling is a regular geometric shape with high contrast and is coated with an anti-reflective coating.

9. A method for automatic inspection and identification of cable tension in a suspension bridge based on a sliding rail moving platform and machine vision, employing the automatic inspection and identification system for cable tension in a suspension bridge based on a sliding rail moving platform and machine vision as described in claim 1, characterized in that... Includes the following steps: S1: System initialization, start the inspection task through the background data processing and analysis subsystem (3), load the sling database, and instruct the mobile inspection vehicle subsystem (2) to perform a self-inspection and then depart from the safety dock (12); S2: The mobile inspection vehicle subsystem (2) moves autonomously to the target sling position along the fixed track (11) according to the preset inspection sequence, and monitors the environment and path safety in real time through the environmental perception and obstacle avoidance module (24); S3: After reaching the designated location, the mobile detection vehicle subsystem (2) accurately positions and brakes, adjusts the vision acquisition module (25) to align with the sling target, and collects the current environmental parameters; S4: The visual acquisition module (25) acquires the vibration video of the suspension target and uploads it to the background data processing and analysis subsystem (3) through the data transmission module (26); S5: The background data processing and analysis subsystem (3) calls the advanced cable force analysis algorithm library (32) to perform motion amplification processing and spectrum analysis on the vibration video to extract the frequency, and calculates the corrected cable force value by combining the cable force calculation correction model; S6: Display the calculated cable force value on the visualization platform (33), and call the trend analysis algorithm to assess the health status and make early warning judgments; S7: After completing the current sling inspection, the mobile inspection vehicle subsystem (2) unlocks and moves to the next sling, repeating steps S3 to S6 until the inspection of all slings of the entire bridge is completed; S8: After the full bridge inspection is completed, the mobile inspection vehicle subsystem (2) automatically returns to the safe docking station (12), and the background data processing and analysis subsystem (3) generates and outputs a complete inspection report.

10. The automatic inspection and identification method for cable tension of a suspension bridge based on a sliding rail moving platform and machine vision according to claim 9, characterized in that, During the movement in steps S2 and S7, if the environmental perception and obstacle avoidance module (24) detects that the wind speed continuously exceeds 8m / s or identifies an obstacle in front, the mobile detection vehicle (2) will automatically perform the operation of pausing, waiting or returning to the safe dock (12).

Images