A rail robot inspection system

Abstract

The present application relates to the technical field of intelligent operation and maintenance, in particular to a rail robot inspection system, the rail system is installed below a steel structure net rack or a space truss, providing an operation path covering all target detection areas for an expandable robot body; the expandable robot body moves along the operation path of the rail system, accurately positioning the operation position of the expandable robot and the target position according to a three-dimensional positioning system, and collecting inspection data through a data collection system; a distributed power supply and charging system provides continuous energy supply for the expandable robot; the data collection system cooperates with a camera module and a light supplement module to obtain inspection data; a data transmission system transmits the collected inspection data to a terminal control platform. The rail robot inspection system realizes automatic acquisition, accurate positioning and stable return of inspection data information of a concealed space of a steel structure, and provides reliable data for operation and maintenance decision-making.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology, specifically a track-mounted robot inspection system. Background Technology

[0002] Public buildings typically employ large-span steel structure grids or space truss systems, creating numerous hidden spaces that are difficult for people to access or stay in for extended periods. These hidden spaces contain a large number of critical load-bearing components, connection nodes, and ancillary facilities, and their operational status directly affects the structural and operational safety of large public buildings.

[0003] Current methods for inspecting concealed spaces in steel structures mainly rely on manual high-altitude operations or localized fixed monitoring equipment. Manual inspections not only suffer from high operational risks, high labor intensity, and limited inspection frequency, but are also constrained by space conditions, making it difficult to achieve systematic coverage of all components. Fixed monitoring equipment, on the other hand, suffers from high deployment costs, poor adaptability, and difficulty in covering complex spaces, making it difficult to meet the dynamic inspection needs of concealed spaces.

[0004] With the development of inspection robot technology, some track-based or mobile inspection robots have begun to be applied in the field of building inspection. However, existing technologies focus more on the design of the robot body and lack system-level adaptation for the hidden spaces of steel structures in large public buildings. There are still obvious deficiencies in terms of track layout logic, full-scene coverage capability, attitude adjustment range, long-distance data acquisition in low-light environments, accuracy of the correlation between the positioning system and the real scene space, long-term continuous operation and communication reliability, making it difficult to form a stable and reusable engineering inspection system.

[0005] Therefore, there is an urgent need for a dedicated inspection system for concealed spaces in the steel structures of large public buildings. This system should be designed systematically to address issues such as insufficient accessibility, incomplete coverage, inadequate data collection capabilities, inaccurate positioning, and weak continuous operation capabilities, thereby enabling safe, efficient, and intelligent inspection of concealed spaces. Summary of the Invention

[0006] (1) Technical problems to be solved

[0007] This invention provides a track-mounted robot inspection system to achieve stable, continuous, and full-coverage automatic inspection in concealed spaces within steel structures. It ensures the clarity and accuracy of data acquisition in low-light or no-light environments, while improving the inspection robot's environmental adaptability and data acquisition efficiency in complex steel structure spaces, ensuring long-term stable transmission of inspection data and the overall safety of the system operation.

[0008] (2) Technical solution

[0009] To achieve the above objectives, the present invention provides a track-type robot inspection system, the system comprising: a track system, an expandable inspection robot body, a three-dimensional positioning system, a data acquisition system, a distributed power supply and charging system, and a data transmission system;

[0010] The track system is installed under the steel structure grid or space truss, providing an operating path for the expandable inspection robot to cover the entire target detection area.

[0011] The expandable inspection robot moves along the track system's running path. The three-dimensional positioning system accurately locates the robot's running position and target position. The data acquisition system collects inspection data in full coverage, including image information within the scene and image information of the detected target.

[0012] The distributed power supply and charging system provides a continuous energy supply for the scalable inspection robot; the data acquisition system acquires inspection data through the collaborative operation of the camera module and the supplementary lighting module.

[0013] Preferably, the data transmission system forms a closed communication loop to stably transmit the collected inspection data back to the terminal control platform;

[0014] The terminal control platform also includes a structural state assessment module, which is used to collect relative displacement data between the track system and adjacent steel structure nodes at several preset calibration points of the track system, and form deformation time series data of each node.

[0015] Based on the inspection data collected by the scalable inspection robot, the target image information is used to detect and obtain the time series data of the rust area and crack length.

[0016] Correlation analysis and evolution increment statistics are performed on the time series data of deformation and the time series data of disease characteristics to calculate the disease-deformation coupling index, which characterizes the degree of correlation between disease evolution and deformation changes.

[0017] Based on the disease-deformation coupling index and the importance of the components, the risk weight of each component is dynamically corrected to generate updated risk weight parameters. The updated risk weight parameters are then provided to the inspection task scheduling module of the terminal control platform for priority and high-frequency inspection of high-risk components in subsequent inspection cycles.

[0018] Preferably, the expandable inspection robot body is equipped with an attitude adjustment module, which drives the expandable inspection robot body to expand its body shape within the concealed space of the steel structure. The range of the body shape expansion is determined according to the height difference of the components in the concealed space of the steel structure, the complexity of the spatial orientation, and the degree of accessibility restriction.

[0019] Preferably, the attitude adjustment module works with the gimbal to perform multi-degree-of-freedom attitude adjustment, enabling the expandable inspection robot to adaptively adjust the detection distance, detection angle, and detection orientation according to different spatial location characteristics.

[0020] Preferably, the layout of the track system is comprehensively planned based on the effective detection distance range of the expandable inspection robot data acquisition module and the spatial morphological characteristics of the steel structure concealed space, and is deployed in a full-coverage manner along the main structural components within the steel structure concealed space.

[0021] Preferably, the three-dimensional positioning system, based on the correlation analysis between the three-dimensional model of the concealed space of the steel structure and the physical space, and combined with the positioning and perception module of the expandable inspection robot and the track system, realizes the position identification of the expandable inspection robot in the track system's operating path.

[0022] Preferably, the three-dimensional positioning system associates the inspection data collected during the inspection process of the scalable inspection robot with the corresponding spatial location to generate a structured dataset with spatial coordinate attributes.

[0023] Preferably, the camera module of the data acquisition system has a zoom function, which can acquire clear and identifiable inspection data under different detection distances and target scales by adjusting the acquisition field of view and imaging magnification.

[0024] The supplementary lighting module of the data acquisition system dynamically adjusts the lighting intensity and working status according to the real-time lighting conditions and detection requirements in the concealed space of the steel structure, and links with the camera module to ensure the clarity of inspection data acquisition in no light or low light environments.

[0025] The data acquisition system also includes an imaging control module, which divides the supplementary lighting module into multiple supplementary lighting zones and independently adjusts the on-state and illumination intensity of each supplementary lighting zone.

[0026] The saturation pixel ratio of the preview image is statistically analyzed during the preview imaging stage. and the proportion of the highlighted area An imaging evaluation function is constructed; the calculation formula for the imaging evaluation function is as follows:

[0027] ;

[0028] in, This serves as a reference sharpness value for the image. For the sharpness normalization term, Image sharpness index; These are weighting coefficients set according to actual needs, used to balance sharpness and glare reduction effects;

[0029] Within the adjustable parameter space of supplementary lighting zones, zoom ratio, and exposure time, an iterative search algorithm is used to find the parameter combination that enables the imaging evaluation function to achieve the preset optimization target, and the parameter combination is automatically adjusted when the imaging result does not meet the requirements for sharpness and reflection suppression.

[0030] Preferably, the distributed power supply and charging system has multiple power supply nodes deployed along the track system. Each power supply node adopts a contact charging method, providing autonomous navigation charging and long-term continuous operation charging for the expandable inspection robot through the power supply nodes.

[0031] Preferably, the terminal control platform centrally stores and classifies the returned inspection data; based on the spatial coordinate attributes provided by the three-dimensional positioning system, it visualizes and maps the inspection data with the three-dimensional model; and after performing trend analysis on the historical inspection data, it generates a steel structure condition assessment report.

[0032] (3) Beneficial effects

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] 1. By continuously deploying a track system along the concealed space of the steel structure and planning the path according to the effective distance of the inspection data collection, the expandable inspection robot can form a continuous running path covering the entire target detection area. It can stably reach concealed areas that are inaccessible to humans or have high risks, solving the problem of insufficient coverage of traditional manual inspection and point detection methods.

[0035] 2. By expanding the form factor of the inspection robot and rotating the gimbal at multiple angles, blind spots caused by component obstruction can be effectively avoided, thereby improving the integrity and reliability of inspection data collection.

[0036] 3. The data acquisition system, through the coordinated work of the supplementary lighting module and the camera module, enables the inspection robot to acquire clear and identifiable inspection data in low-light or no-light environments in the concealed space of the steel structure, thus meeting the data quality requirements for concealed space inspection.

[0037] 4. By establishing a correlation and positioning mechanism between the 3D model and the actual steel structure space, the image data collected during the inspection process is accurately mapped to its real spatial location, providing a reliable basis for subsequent condition assessment and maintenance decisions.

[0038] 5. Distributed power supply and charging provide continuous and stable energy support for the scalable inspection robot under long-distance track operation conditions, reducing the frequency of manual maintenance and improving the continuity and reliability of the overall system operation.

[0039] 6. By using track-mounted inspection robots to replace manual labor in high-altitude, concealed, and high-risk areas, the safety risks to personnel are significantly reduced, while the inspection of concealed spaces in steel structures is automated, continuous, and intelligent. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the composition of a track-type robot inspection system according to the present invention.

[0041] Figure 2 This is a schematic diagram of the track system, the hanger system, and the expandable inspection robot of a track-type robot inspection system according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Before describing the embodiments, it is necessary to explain the application scenarios of the present invention. The present invention is aimed at the concealed spaces of steel structures in large public buildings (such as railway stations, airport terminals, stadiums, convention centers, etc.). Through systematic design, it solves the core problems of accessibility, full coverage, data validity, positioning accuracy, and continuous operation capability of space inspection in the concealed spaces of steel structures in such public buildings, so as to achieve safe, efficient, intelligent and automatic inspection of concealed spaces.

[0044] Example 1: As Figure 1 and Figure 2 As shown, this embodiment provides a track-type robot inspection system, which includes: a track system, an expandable inspection robot body, a three-dimensional positioning system, a data acquisition system, a distributed power supply and charging system, and a data transmission system;

[0045] The track system is installed under the steel structure grid or space truss, providing an operating path for the expandable inspection robot to cover the entire target detection area.

[0046] The layout of the track system is comprehensively planned based on the effective detection distance range of the expandable inspection robot data acquisition module and the spatial morphological characteristics of the steel structure concealed space, and is deployed to fully cover the main structural components within the steel structure concealed space.

[0047] For example, this embodiment uses a large railway passenger station as the application object. The main structure of the station is a large-span steel space frame, with a roof span of approximately 120m × 80m, a rise of approximately 15m, and a grid size of approximately 3m × 3m. The chords and web members are mainly composed of circular steel pipes. A suspended ceiling system is installed below the space frame, forming a concealed space of approximately 1.5 to 3.0m in height between the suspended ceiling and the lower chord of the space frame. This space contains densely distributed electromechanical pipelines, fire protection system pipes, and power distribution facilities, making manual entry difficult and posing a high operational risk. This embodiment demonstrates the overall deployment and operational verification of the track-mounted robot inspection system described in this invention.

[0048] Based on the spatial morphological characteristics of the station building's steel structure grid, the distribution of key nodes, and the effective detection distance range (10-15m) of the expandable inspection robot data acquisition system, the track system path is comprehensively planned, and the system is deployed in a full-coverage manner along the main structural components within the concealed space of the steel structure.

[0049] The track path is laid out along the extension direction of the lower chord members of the space frame, adopting a loop topology. A main track is laid approximately every 12m longitudinally, with connecting tracks running laterally, forming a grid-like path system of "longitudinal main track + transverse connecting track". The spacing between adjacent tracks is controlled between 10 and 14m, covering all component areas and node areas requiring key inspection. Single-track systems are used in locally confined areas (such as areas with dense piping), while double-track systems are used in wider areas to ensure both flexibility of movement and efficiency of data collection.

[0050] The track is fixed to the lower chord node of the space frame via a dedicated hanger system. The hangers use adjustable rigid connections, and the installation height deviation is controlled within ±5mm to ensure track smoothness. The total track length of the entire station is approximately 2000m, forming a closed loop without breaks, and can cover the entire concealed space target detection area of ​​the station building's steel structure.

[0051] The expandable inspection robot body is equipped with an attitude adjustment module, which drives the expandable inspection robot body to expand its body shape within the concealed space of the steel structure. The range of the body shape expansion is determined according to the height difference of the components in the concealed space of the steel structure, the complexity of the spatial orientation, and the degree of accessibility restriction.

[0052] The attitude adjustment module works with the gimbal to perform multi-degree-of-freedom attitude adjustment, enabling the expandable inspection robot to adaptively adjust the detection distance, detection angle, and detection orientation according to different spatial location characteristics.

[0053] For example, the expandable inspection robot is equipped with a posture adjustment module, allowing for vertical expansion of 0-400mm. The range of this expansion is determined by a combination of factors, including the height difference between components in the concealed steel structure space (approximately 1.5-3.0m), spatial complexity (dense pipeline intersections), and limited accessibility (local clearance less than 400mm). The posture adjustment module, in conjunction with a dual-axis gimbal, enables multi-degree-of-freedom posture adjustment, with a horizontal rotation range of ±180° and a pitch range of +90° / −30°. This allows the expandable inspection robot to adaptively adjust its detection distance, angle, and orientation based on different spatial characteristics, adapting to variations in ceiling height and the inspection needs of components with different orientations. At preset detection points, after precise docking, the control unit automatically drives the robot to expand to the set height based on the spatial characteristics of the detection point and adjusts the gimbal angle so that the camera module's field of view is aligned with the target component surface.

[0054] The expandable inspection robot is also equipped with an environmental perception module consisting of lidar and ultrasonic sensors to perceive surrounding obstacles in real time. If an obstacle is detected in front of it during the journey, it will automatically shrink to a safe size, pass through the obstacle area, and then return to its working form to ensure the robot's safe passage in densely piped areas.

[0055] The data acquisition system acquires inspection data through the coordinated operation of a camera module and a supplementary lighting module.

[0056] The data acquisition system also includes an imaging control module, which divides the supplementary lighting module into multiple supplementary lighting zones and independently adjusts the on-state and illumination intensity of each supplementary lighting zone; during the preview imaging stage, it statistically analyzes the saturated pixel ratio and the area ratio of the bright area of ​​the preview image to construct an imaging evaluation function.

[0057] Within the adjustable parameter space of supplementary lighting zones, zoom ratio, and exposure time, an iterative search algorithm is used to find the parameter combination that enables the imaging evaluation function to achieve the preset optimization target, and the parameter combination is automatically adjusted when the imaging result does not meet the requirements for sharpness and reflection suppression.

[0058] Specifically, the supplementary lighting module of the data acquisition system consists of multiple independently controllable LED zones, including a left oblique light zone, a right oblique light zone, an upper ring light zone, and a near-coaxial light zone. The illumination intensity of each zone can be continuously adjusted within the range of 0 to 2000 lux. The camera module uses a 12x optical zoom industrial camera, which supports real-time adjustment of zoom ratio, exposure time, and gain parameters.

[0059] Once the expandable inspection robot reaches a preset inspection point and completes its posture adjustment, the imaging control module first activates some supplementary lighting zones with a preset low-brightness combination. For example, only the left oblique lighting zone is activated, with an illuminance of approximately 300 lux, while the other zones are deactivated. A preview image is then acquired, and the following processing is performed on the preview image:

[0060] Image dimensions: Width W, Height H, Total pixels. Grayscale / Brightness Values: Assuming the image has been converted to obtain a single-channel brightness map. x=1…W, y=1…H, with values ​​ranging from 0 to 255 (8 bits); "Saturated pixel": a pixel with a brightness of 255; "Highlight pixel": a pixel with a brightness greater than a certain highlight threshold. (e.g., 230); Count the number of saturated pixels. , This refers to the number of saturated pixels in the image with a brightness value equal to 255, i.e., the number of pixels that satisfy this value. pixels Count the number of pixels obtained; calculate the saturation pixel ratio. Reasonable range recommendations (Within 5%), otherwise there may be obvious overexposure or strong reflective areas.

[0061] Select a high-brightness threshold from the brightness histogram. (e.g., 230), count the number of highlighted pixels. , To meet pixels The number of pixels counted; the proportion of bright areas whose brightness exceeds the bright area threshold. The image is filtered using the Laplacian operator, and the variance of the Laplacian response is calculated. As an indicator of image sharpness. The larger the value, the richer the image edges and details, and the higher the sharpness. A reference image sharpness value can be selected by statistically analyzing a large number of samples. .

[0062] The imaging control module constructs the following imaging evaluation function:

[0063] .

[0064] in, This serves as a reference sharpness value for the image. For the sharpness normalization term, Image sharpness index; These are weighting coefficients set according to actual needs, used to balance sharpness and glare reduction.

[0065] A limited number of searches are performed within the set of selectable parameters for fill light zones, zoom ratio, and exposure time: On the one hand, by changing the fill light zone combination (such as turning off the near coaxial light zone and only retaining the left oblique light or left and right symmetrical oblique light), the positive strong reflection is reduced, and the area of ​​the bright spot is reduced; on the other hand, the zoom ratio is appropriately adjusted according to the current detection distance so that the diseased area occupies an appropriate proportion in the image, while by shortening the exposure time and appropriately increasing the side illumination, overexposure is avoided while ensuring clear details in the dark areas.

[0066] In actual testing, for a white, highly reflective coated steel pipe at a node of the lower chord of the space frame, medium-brightness illumination was applied using the upper annular light zone and near-coaxial light zone. In the preview images collected, a large area of ​​high brightness appeared in the center of the node. The saturation pixel ratio was statistically obtained. Highlighted area ratio Laplace variance Although the value is relatively high, the J value of the evaluation function is low, indicating a serious reflection problem.

[0067] Based on this, the imaging control module automatically shuts down the near-coaxial lighting zone, reduces the illuminance of the upper annular lighting zone to approximately 400 lux, while increasing the illuminance of the left and right oblique lighting zones to approximately 800 lux. The exposure time is shortened from 8 ms to 4 ms, and the zoom ratio is slightly increased from 6x to 8x to reduce frontal reflections and utilize side lighting to enhance the contrast of lesion edges. After adjusting these parameters and re-acquiring images, statistical results show that the saturated pixel ratio has decreased. The proportion of the bright area has decreased. Meanwhile, Laplace variance The value decreased only slightly, while the J value of the evaluation function increased significantly.

[0068] In the final image, the rust patches and fine crack boundaries that were previously covered by highlights are clearly visible. The defect feature extraction module can reliably identify the rust area and crack length, achieving accurate quantification of the defect's geometric scale. Engineering applications show that, through the aforementioned imaging control strategy of supplementary lighting-zoom-reflection suppression, the imaging stability of this invention is significantly improved in the high-reflectivity environment of concealed spaces in steel structures. The defect identification accuracy and repeatability are superior to traditional schemes that only use fixed supplementary lighting and fixed exposure parameters.

[0069] The camera module of the data acquisition system has a zoom function, which can acquire clear and identifiable inspection data under different detection distances and target scales by adjusting the acquisition field of view and imaging magnification.

[0070] The supplementary lighting module of the data acquisition system dynamically adjusts the lighting intensity and working status according to the real-time lighting conditions and inspection requirements in the concealed space of the steel structure, and links with the camera module to ensure the clarity of inspection data acquisition in no light or low light environments.

[0071] For example, the camera module of the data acquisition system uses an industrial-grade zoom camera with a pixel resolution of 2048×1080. It has zoom and magnification functions, with an optical zoom of 12x. By adjusting the acquisition field of view and imaging magnification, it can acquire clear and identifiable inspection data under different detection distances and target scales. The maximum imaging clarity at a detection distance of 10m is no less than 2mm / pixel, which can effectively identify typical defects such as surface cracks, corrosion, and weld abnormalities of components.

[0072] The supplementary lighting module of the data acquisition system uses a high-power LED array light, with the illumination intensity continuously adjustable within the range of 200–2000 lux. The supplementary lighting module dynamically adjusts the illumination intensity and operating status based on real-time lighting conditions and inspection requirements within the concealed space of the steel structure, and is linked with the camera module: it automatically activates supplementary lighting when the ambient illuminance is below 300 lux, and dynamically adjusts the light intensity according to the acquisition distance to ensure uniform illumination of the target area and guarantee the clarity of inspection data acquisition in no-light or low-light environments. For a 10m acquisition distance, the supplementary lighting module operates at approximately 80W, ensuring that the image contrast meets recognition requirements. At each preset inspection point, the robot executes the acquisition task according to the following process: docking and positioning → shape expansion → gimbal adjustment → supplementary lighting activation → image acquisition → supplementary lighting deactivation → gimbal reset → shape retraction → moving to the next point.

[0073] The expandable inspection robot moves along the running path of the track system. The three-dimensional positioning system accurately locates the running position and target position of the expandable inspection robot. The data acquisition system collects inspection data in full coverage, including image information in the scene and image information of the detected target.

[0074] The three-dimensional positioning system, based on the correlation analysis between the three-dimensional model of the concealed space of the steel structure and the physical space, and combined with the positioning and perception module of the expandable inspection robot and the track system, realizes the position identification of the expandable inspection robot in the track system's operating path.

[0075] The three-dimensional positioning system associates the inspection data collected during the inspection process of the scalable inspection robot with the corresponding spatial location to generate a structured dataset with spatial coordinate attributes.

[0076] For example, the 3D positioning system uses the BIM 3D model of the station's steel structure as its spatial basis. Through correlation analysis between the 3D model and the physical space, and combined with the positioning perception module of the scalable inspection robot and the track system, it achieves position recognition of the scalable inspection robot along the track system's operating path. Specifically, a magnetic coded marker is placed approximately every 2.5m along the track system, each magnetic coded marker corresponding to a unique spatial coordinate in the BIM model; the robot's bottom is equipped with a magnetic strip reader as a positioning perception module, which can identify the coded markers it passes in real time, achieving a stopping positioning accuracy of ≤±5mm.

[0077] The 3D positioning system associates image data collected during the inspection process of the scalable inspection robot with corresponding spatial locations. It synchronously packages the image data, timestamps, and spatial location codes to generate a structured dataset with spatial coordinate attributes. After being uploaded to the terminal control platform, the platform automatically maps the data to the corresponding location in the BIM 3D model, achieving a "one point, one file" digital record. The 3D positioning system has a positioning error of ≤±5mm and a data association success rate of 100%. The terminal control platform can accurately retrieve historical inspection image data for any node using spatial coordinates, verifying the accuracy of the positioning system and the effectiveness of the data association.

[0078] The distributed power supply and charging system provides a continuous energy supply for the scalable inspection robot.

[0079] The distributed power supply and charging system has multiple power supply nodes deployed along the track system. Each power supply node adopts a contact charging method, which provides autonomous navigation charging and long-term continuous operation charging for the expandable inspection robot.

[0080] For example, the distributed power supply and charging system has multiple power supply nodes deployed along the track system. The total length of the track in the test scenario is about 2000m, and a contact charging node is deployed every 250m, evenly distributed throughout the entire track loop.

[0081] When fully charged (100% battery), the robot can operate continuously for approximately 8 hours in normal inspection mode, covering an inspection path of 2000 meters. When the battery level drops to 20%, the robot automatically plans a path to the nearest power supply node, recharges for 2 hours to restore full charge, and can continue performing inspection tasks, meeting the requirements for long-term continuous operation. The entire process requires no human intervention, and the task continuity is complete, verifying the feasibility of the distributed power supply and charging system supporting long-term continuous operation.

[0082] The data transmission system deploys eight wireless communication nodes along the track system, forming a closed communication loop to stably transmit the collected inspection data back to the terminal control platform. During the test, the power supply to each communication node was manually disconnected sequentially to simulate node failure scenarios, and the system communication switching time and data transmission continuity were recorded.

[0083] The test results showed that when a single communication node failed, the system automatically switched to the backup communication path within 3 seconds using the redundant path of the closed communication loop. Data transmission was not interrupted during the switching process, and the data packet loss rate was 0%. This verified the high reliability of the closed communication loop architecture, which can effectively cope with communication interference and node failure scenarios in the concealed space of complex steel structures, and ensure that inspection data is stably transmitted back to the terminal control platform.

[0084] The terminal control platform also includes a structural state assessment module, which is used to collect relative displacement data between the track system and adjacent steel structure nodes at several preset calibration points of the track system, and form deformation time series data of each node.

[0085] Based on the inspection data collected by the scalable inspection robot, the target image information is used to detect and obtain the time series data of the rust area and crack length.

[0086] Correlation analysis and evolution increment statistics are performed on the time series data of deformation and the time series data of disease characteristics to calculate the disease-deformation coupling index, which characterizes the degree of correlation between disease evolution and deformation changes.

[0087] Based on the disease-deformation coupling index and the importance of the components, the risk weight of each component is dynamically corrected to generate updated risk weight parameters. The updated risk weight parameters are then provided to the inspection task scheduling module of the terminal control platform for priority and high-frequency inspection of high-risk components in subsequent inspection cycles.

[0088] For example, a track calibration point is set approximately every 12m along the lower chord direction of the track structure. Each calibration point is connected to an adjacent lower chord node via a rigid connector, establishing a fixed relative position. A laser displacement sensor is installed at each calibration point to measure the relative displacement between the track bottom surface and the lower chord node. The sensor has a range of 0–20mm and a resolution of no less than 0.01mm.

[0089] During a complete inspection cycle, the scalable inspection robot runs along the track system. When it passes near a calibration point, it precisely stops above the calibration point according to the spatial coordinate information provided by the 3D positioning system. It then reads the instantaneous measurement value of the laser displacement sensor corresponding to that calibration point via the communication bus, and adds the current timestamp and spatial coordinate code to form the node deformation time-series data. For the i-th node, the relative displacement collected during the k-th inspection is denoted as... Data was collected N times within a year to form time series data of deformable variables. .

[0090] Meanwhile, the scalable inspection robot captures high-definition images at a resolution of 2048×1080 each time it stops near the node. After receiving the images, the terminal control platform uses a disease feature extraction algorithm to automatically segment and identify rust and crack areas in the images: extracting the pixel area of ​​the rust area through a semantic segmentation network. The effective length of the crack is obtained through crack detection and skeleton extraction algorithms. .

[0091] Since the pixel and spatial dimensions of the camera module have been calibrated during the calibration phase, the size conversion factor at the current detection distance is known. (mm / pixel), then the corrosion area and crack length of node i during the k-th inspection can be expressed as:

[0092] The actual area corresponding to the pixel area of ​​the rust region is: ,unit: .

[0093] Converting the effective crack length to the actual length: , Unit: mm.

[0094] The terminal control platform further constructs a comprehensive disease index for node i. :

[0095] .

[0096] Among them, the reference value of corrosion area For example, selecting 1000 mm² according to specifications or experience indicates the level of "significant corrosion," and the crack length is a reference value. For example, selecting 10mm indicates the level of "cracks that require special attention".

[0097] and The weighting coefficients, set based on engineering experience, are used to balance the impact of corrosion and cracking on structural performance and meet the requirements. Common settings For example, if corrosion is as important as cracks, then... .

[0098] To measure the evolution rate of deformation and disease at node i over a recent period, the terminal control platform calculates the average increment of deformation and disease index based on the most recent M inspection data (e.g., M=6):

[0099] ; .

[0100] ; .

[0101] in, mm / time represents the average deformation increment per inspection cycle; , dimensionless / times, represents the average increase in the disease index per cycle; N represents the number of inspections completed on node i within a specified statistical period (e.g., one year). If or This indicates that the quantity is decreasing, which can be expressed in subsequent formulas. Negative values ​​are treated as 0.

[0102] The time series data of deformation at node i are calculated using the Pearson correlation coefficient. Time series data of disease characteristics correlation coefficient , , Value range: This is used to characterize the degree of synchronous change between the two. A reference value for deformation increment is introduced. Reference values ​​for disease increase These are used as threshold parameters to determine whether the deformation development rate and the disease deterioration rate exceed the engineering warning level, respectively. Deformation increment reference value. The reference value for the increase in defects is determined based on the structural type, the importance of the components, and engineering operation and maintenance experience. The values ​​are determined based on engineering experience regarding the impact of disease characteristics (corrosion area, crack length) on structural performance, and are configured by professionals during system deployment; a typical example is... A value of 1.0 mm / time indicates that an average increase of more than 1 mm in deformation increment per inspection cycle is considered an excessive deformation rate. A value of 0.2 per inspection indicates that an average increase of more than 0.2 in the disease index per inspection cycle is considered a relatively rapid disease development. The deformation increment and disease increment are divided by their respective reference values, then normalized using dimensionless methods, and negative values ​​are truncated to 0. ,in A value greater than 1 indicates that the increment exceeds the reference value. Combining correlation and increment, a disease-deformation coupling index is defined: When deformation and disease are both increasing rapidly and are highly correlated, It will be significantly greater than 1; when the deformation or disease changes are not significant, or Approaching 0 It will also be very small.

[0103] During the system initialization phase, the terminal control platform assigns an initial risk weight to each node based on the component type (main chord, secondary members, web members), the component's importance in the structural system, and the initial condition of defects. Based on structural importance and initial disease setting, the typical range can be set to 1–5. When the disease-deformation coupling index… After the calculation is completed, the structural state assessment module dynamically adjusts the risk weights according to the following formula: ;in, This is the magnification factor, typically ranging from 0.5 to 2; when A larger value significantly increases the risk weight of that node.

[0104] For example, in the past 6 inspections, the deformation of a certain lower chord node i slowly increased from 2.0 mm to 6.0 mm, with an average increase of approximately 0.67 mm per inspection; the corresponding disease index increased from 0.1 to 0.75, with an average increase of approximately 0.13 per inspection. The correlation coefficient between the two is... The calculated disease-deformation coupling index is approximately 0.85. The risk level is greater than the preset threshold of 1.0. Based on this, the terminal control platform upgrades the node from "medium risk" to "high risk", increases its risk weight by approximately 2.3 times relative to the initial value, and shortens the inspection cycle of the node from 6 months to 1 month in the inspection task scheduling module. At the same time, the node is highlighted in red in the 3D model to prompt maintenance personnel to prioritize on-site verification and handling of this node in the next maintenance.

[0105] The data transmission system forms a closed communication loop, stably transmitting the collected inspection data back to the terminal control platform.

[0106] The terminal control platform centrally stores and classifies the returned inspection data; based on the spatial coordinate attributes provided by the three-dimensional positioning system, it visualizes and maps the inspection data with the three-dimensional model; and it generates a steel structure condition assessment report after performing trend analysis on historical inspection data.

[0107] For example, the terminal control platform centrally stores and categorizes the inspection data transmitted back from the data transmission system. The platform automatically categorizes and archives the data according to component type (chord members, web members, nodes, ancillary facilities, etc.), inspection batch, and spatial area, forming a structured digital inspection archive. Based on the station's BIM 3D model, the terminal control platform automatically maps the image data collected at each inspection point to the corresponding component location in the BIM model, using the spatial coordinate attributes provided by the 3D positioning system, and displays the visualization. The platform supports directly clicking on any component in the 3D model view to retrieve all historical inspection images and corresponding inspection data for that component, with an operation response time of no more than 2 seconds.

[0108] The accuracy of data mapping was verified for 50 selected typical component locations. The platform mapping results were compared with the actual measured component locations on site. The average position deviation was 3.2 mm, and the maximum deviation was 7.8 mm. The positioning accuracy met the requirements of engineering applications (≤±10 mm), which verified the accuracy of the three-dimensional visualization mapping display function.

[0109] The terminal control platform performs trend analysis on historical inspection data and generates a steel structure condition assessment report accordingly. For a specific lower chord node of a space frame, data from the last six inspections were retrieved for a longitudinal comparative analysis of the node's surface condition: After the fourth inspection, pitting rust spots with a diameter of approximately 5 mm appeared on the node; by the fifth inspection, the rust spot area had expanded to approximately 12 mm; and by the sixth inspection, it had expanded to approximately 18 mm, showing a clear trend of rust expansion. This node was marked with a prominent color in the 3D model, and a steel structure condition assessment report was generated, recommending that maintenance personnel address this node during the next routine inspection. The verification results demonstrate that the terminal control platform can effectively utilize historical inspection data to conduct longitudinal trend analysis of the structural condition, providing reliable data support for preventative maintenance decisions.

[0110] Using the remote scheduling function of the terminal control platform, a temporary priority instruction was issued to the robot performing a routine inspection task, requiring the robot to immediately proceed to five designated nodes to perform key re-inspection tasks after completing the current inspection point. Upon receiving the instruction, the robot completed path replanning within 4.3 seconds, abandoned the original path, and headed to the first designated re-inspection node. After completing the re-inspection of the five nodes, it automatically returned to the original interrupted task point to continue performing routine inspections. The entire operation was characterized by accurate response, reasonable path planning, and smooth task switching, verifying the effectiveness and real-time performance of the remote scheduling function of the terminal control platform.

[0111] Finally, it should be noted that although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A track-mounted robot inspection system, characterized in that, The system includes: a track system, an expandable inspection robot body, a three-dimensional positioning system, a data acquisition system, a distributed power supply and charging system, and a data transmission system; The track system is installed under the steel structure grid or space truss, providing the expandable inspection robot body with an operating path covering the entire target detection area; The expandable inspection robot moves along the running path of the track system. The three-dimensional positioning system accurately locates the running position and target position of the expandable inspection robot. The data acquisition system collects inspection data in full coverage, including image information in the scene and image information of the detected target. The distributed power supply and charging system provides a continuous energy supply for the scalable inspection robot; the data acquisition system acquires inspection data through the collaborative operation of the camera module and the supplementary lighting module. The data transmission system forms a closed communication loop, stably transmitting the collected inspection data back to the terminal control platform; The terminal control platform also includes a structural state assessment module, which is used to collect relative displacement data between the track system and adjacent steel structure nodes at several preset calibration points of the track system, and form deformation time series data of each node. Based on the inspection data collected by the scalable inspection robot, the target image information is used to detect and obtain the time series data of the rust area and crack length. Correlation analysis and evolution increment statistics are performed on the time series data of deformation and the time series data of disease characteristics to calculate the disease-deformation coupling index, which characterizes the degree of correlation between disease evolution and deformation changes. Based on the disease-deformation coupling index and the importance of the components, the risk weight of each component is dynamically corrected to generate updated risk weight parameters. The updated risk weight parameters are then provided to the inspection task scheduling module of the terminal control platform for priority and high-frequency inspection of high-risk components in subsequent inspection cycles. The camera module of the data acquisition system has a zoom function, which can acquire clear and identifiable inspection data under different detection distances and target scales by adjusting the acquisition field of view and imaging magnification. The supplementary lighting module of the data acquisition system dynamically adjusts the lighting intensity and working status according to the real-time lighting conditions and detection requirements in the concealed space of the steel structure, and links with the camera module to ensure the clarity of inspection data acquisition in no light or low light environments. The data acquisition system also includes an imaging control module, which divides the supplementary lighting module into multiple supplementary lighting zones and independently adjusts the on-state and illumination intensity of each supplementary lighting zone. The saturation pixel ratio of the preview image is statistically analyzed during the preview imaging stage. and the proportion of the highlighted area An imaging evaluation function is constructed; the calculation formula for the imaging evaluation function is as follows: ； in, This serves as a reference sharpness value for the image. For the sharpness normalization term, Image sharpness index; These are weighting coefficients set according to actual needs, used to balance sharpness and glare reduction effects; Within the adjustable parameter space of supplementary lighting zones, zoom ratio, and exposure time, an iterative search algorithm is used to find the parameter combination that enables the imaging evaluation function to achieve the preset optimization target, and the parameter combination is automatically adjusted when the imaging result does not meet the requirements for sharpness and reflection suppression.

2. The track-mounted robot inspection system according to claim 1, characterized in that, The expandable inspection robot body is equipped with an attitude adjustment module, which drives the expandable inspection robot body to expand its body shape within the concealed space of the steel structure. The range of the body shape expansion is determined according to the height difference of the components in the concealed space of the steel structure, the complexity of the spatial orientation, and the degree of accessibility restriction.

3. The track-mounted robot inspection system according to claim 2, characterized in that, The attitude adjustment module works with the gimbal to perform multi-degree-of-freedom attitude adjustment, enabling the expandable inspection robot to adaptively adjust the detection distance, detection angle, and detection orientation according to different spatial location characteristics.

4. The track-mounted robot inspection system according to claim 1, characterized in that, The layout of the track system is comprehensively planned based on the effective detection distance range of the expandable inspection robot data acquisition module and the spatial morphological characteristics of the steel structure concealed space, and is deployed to fully cover the structural components within the steel structure concealed space.

5. The track-mounted robot inspection system according to claim 1, characterized in that, The three-dimensional positioning system, based on the correlation analysis between the three-dimensional model of the concealed space of the steel structure and the physical space, and combined with the positioning and perception module of the expandable inspection robot and the track system, realizes the position identification of the expandable inspection robot in the track system's operating path.

6. The track-mounted robot inspection system according to claim 5, characterized in that, The three-dimensional positioning system associates the inspection data collected during the inspection process of the scalable inspection robot with the corresponding spatial location to generate a structured dataset with spatial coordinate attributes.

7. The track-mounted robot inspection system according to claim 1, characterized in that, The distributed power supply and charging system has multiple power supply nodes deployed along the track system. Each power supply node adopts a contact charging method, which provides autonomous navigation charging and long-term continuous operation charging for the expandable inspection robot.

8. The track-mounted robot inspection system according to claim 1, characterized in that, The terminal control platform centrally stores and classifies the returned inspection data; based on the spatial coordinate attributes provided by the three-dimensional positioning system, it visualizes and maps the inspection data with the three-dimensional model. A steel structure condition assessment report is generated after trend analysis of historical inspection data.

Images