Intelligent mapping system for catenary equipment in tunnel
By combining multi-dimensional sensors and neural network models, a four-dimensional fusion dataset was constructed, which solved the problems of accurate detection and environmental adaptability of overhead contact line equipment in tunnels, realized the automated identification and intuitive presentation of equipment status, and improved operation and maintenance efficiency and safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAN BRANCH OF CHINA RAILWAY WUHAN ELECTRIFICATION BUREAU GRP CO LTD
- Filing Date
- 2025-09-18
- Publication Date
- 2026-05-15
AI Technical Summary
Traditional methods are insufficient for efficiently and accurately monitoring minute deformations and defects in overhead contact line equipment within tunnels, and they cannot respond in real time to the impact of environmental factors on equipment performance, resulting in the failure to detect potential safety hazards in a timely manner.
A multi-dimensional sensor system, including a laser scanner, a high-definition industrial camera, a BeiDou positioning component, and environmental sensors, is used in conjunction with a neural network model to construct a four-dimensional fusion dataset. This enables the automatic identification of the three-dimensional geometric framework and defects of the overhead contact line equipment, and the coordinate deviation is corrected by dynamically adapting the fusion coefficient formula.
It enables accurate detection and automated defect identification of overhead contact line equipment in tunnels, reduces missed and false detections during manual inspections, improves the intuitive presentation of equipment status and the scientific nature of operation and maintenance decisions, and reduces operation and maintenance costs.
Smart Images
Figure CN121163488B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of overhead contact line engineering inspection technology, specifically an intelligent surveying system for overhead contact line equipment in tunnels. Background Technology
[0002] In the railway transportation sector, overhead contact line equipment in tunnels, as a core component of power transmission, undertakes the critical task of providing continuous and stable power to trains. With the rapid expansion of high-speed railway networks and urban rail transit systems, the scale and complexity of tunnel construction have significantly increased, leading to higher distribution density and maintenance requirements for overhead contact line equipment. The tunnel environment has unique characteristics, such as insufficient lighting, large temperature and humidity fluctuations, and high dust concentrations. These factors pose challenges to the performance stability and service life of the overhead contact line equipment. At the same time, minor deformations or defects in overhead contact line equipment (such as chutes and droppers) may cause power supply failures, directly affecting the safety and efficiency of train operation.
[0003] However, traditional techniques for surveying and monitoring overhead contact line equipment in tunnels mainly rely on manual inspections and simple mechanical measuring tools. These methods have many limitations. First, manual inspections are inefficient and cannot achieve real-time, comprehensive monitoring of all equipment, especially in long tunnels and scenarios with high-density equipment distribution, making it difficult to meet maintenance needs. Second, traditional measuring tools have limited accuracy and cannot accurately capture minute deformations and defects in equipment, resulting in the failure to detect potential safety hazards in a timely manner. Third, manual data recording and processing are prone to errors, and data traceability and analysis capabilities are weak, which is not conducive to the scientific formulation of maintenance decisions. In addition, traditional methods lack comprehensive consideration of environmental factors, such as temperature, humidity, light, and dust, which have a significant impact on the performance and lifespan of overhead contact line equipment but are often overlooked in traditional inspections.
[0004] Therefore, an intelligent surveying system for overhead contact line equipment in tunnels was developed. Summary of the Invention
[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an intelligent mapping system for overhead contact line equipment in tunnels. This invention, by deploying multi-dimensional sensors, addresses the limitations of traditional manual inspections or reliance on single sensors in environments such as insufficient lighting, high dust concentration, and large temperature and humidity fluctuations within tunnels. A laser scanner can penetrate dust to acquire the three-dimensional geometric framework of the equipment. A BeiDou positioning component is linked to the data acquisition action in real time, solving the positioning deviation problem caused by weak GPS signals in tunnels. An environmental sensor group monitors parameters such as temperature, humidity, and lighting in real time, providing environmental adaptation basis for data correction. Through a dynamic adaptation fusion coefficient formula, sensor data weights are calculated to correct original coordinate deviations and construct a four-dimensional fusion dataset. This solves the errors caused by environmental interference or data isolation in traditional methods, and addresses the problems of missed detections during manual visual inspections in low light conditions and the inability of mechanical measuring tools to correlate with environmental parameters.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an intelligent surveying system for overhead contact line equipment in tunnels, the system comprising:
[0007] Multi-source data acquisition module: Deploys multi-dimensional sensors to acquire multi-source data from the overhead contact line, and aggregates and processes the multi-source data through an integrated interface to form a raw dataset;
[0008] Multi-source data fusion module: preprocesses multi-source data in the original dataset, corrects the original coordinates by fusing data weights, and constructs a four-dimensional fused dataset of image, geometry, location and environment;
[0009] Identification and parameter calculation module: Through the neural network model trained on the training set, it determines the defects of key components of the overhead contact line equipment and outputs the type and three-dimensional position coordinates of the effective defects. At the same time, it calculates the sliding groove spacing and the evaluation value of the dropper installation status, and generates a parameter deviation list after comparing it with the design standard.
[0010] Modeling and Interaction Module: Based on the parameter deviation list, a 3D mesh model of the contact network equipment is constructed using laser point clouds. The appearance of the equipment is restored through texture mapping, defect identification results and parameter calculation data are embedded, the geometric errors between the 3D mesh model and the entity are corrected, and 3D interactive functions are developed, supporting model scaling, 360° rotation and translation. A cross-section cutting function is added, and the internal structure can be displayed in a custom planar position. Color markings are configured according to the parameter deviation list, with out-of-tolerance areas marked in red and normal areas marked in green, so that the equipment status is presented visually.
[0011] Data management and application output module: Integrates raw datasets, four-dimensional fusion datasets, parameter deviation lists and three-dimensional mesh models to build a distributed database, supports multi-dimensional retrieval, develops standardized interfaces to connect with the construction and maintenance platform, and generates equipment status assessment reports, three-dimensional model files and deviation adjustment suggestion files, which can be accessed and viewed through mobile applications.
[0012] Furthermore, the multi-source data acquisition module is equipped with multi-dimensional sensors, including a laser scanner, a high-definition industrial camera, a Beidou positioning component, and an environmental sensor group.
[0013] The laser scanner performs a panoramic scan of the target area to obtain the three-dimensional geometric framework of the overhead contact line equipment. The scanning rate is ≥1 million points / second, and the ranging accuracy is ≤±0.5mm.
[0014] The high-definition industrial camera is used to capture images of key components of the overhead contact line equipment, including components such as chutes and suspension cables, with a resolution of ≥12 million pixels.
[0015] The BeiDou positioning component is bound to the data acquisition action in real time, updates the location information periodically, has a positioning accuracy of ≤1m, and supports differential positioning.
[0016] The environmental sensor group includes a temperature and humidity sensor, a light sensor, and a dust concentration sensor, which collect parameters such as temperature, relative humidity, light intensity, and dust concentration inside the tunnel.
[0017] The collected multi-source data is aggregated and processed through an integration interface to form a raw dataset containing image data, laser point cloud data, positioning data, and environmental parameter data.
[0018] Furthermore, in the multi-source data fusion module, the image data in the original dataset is denoised, enhanced, and distortion corrected; the laser point cloud data is filtered and denoised; the positioning data and environmental parameter data are smoothed and corrected; and the data is spatiotemporally aligned based on timestamps and coordinate transformations. Then, the data fusion weights of the laser scanner and the high-definition industrial camera are calculated through a dynamic adaptation fusion coefficient formula, and the original coordinates are corrected using a three-dimensional coordinate collaborative correction formula to construct a four-dimensional fusion dataset of image, geometry, position, and environment.
[0019] Furthermore, in the multi-source data fusion module, the data fusion weights of the laser scanner and the high-definition industrial camera are calculated using a dynamically adapted fusion coefficient formula, which is as follows: ,in, For the first Sensor-like data fusion weights, For indexes of different device types, For laser scanners, For high-definition industrial cameras, For the first The sensor rating ranges from 0 to 1. hour Score the integrity of the laser point cloud. hour Rate the image sharpness. For environmental adaptability, This represents the environmental sensitivity coefficient.
[0020] Furthermore, in the multi-source data fusion module, the original coordinates are corrected using a three-dimensional coordinate collaborative correction formula, which is as follows: ,in, The original coordinates are (x, y, z). The corrected coordinates are (x, y, z). This represents the BeiDou positioning deviation value. Distance from the tunnel entrance The attenuation coefficient is... This is the temperature correction factor. This represents the temperature difference between the measured temperature and the standard temperature.
[0021] Furthermore, in the identification and parameter calculation module, the training set for training the neural network model includes historical labeled images that cover common defects, including edge breakage and deformation in chute damage, blockage of the chute by foreign objects such as concrete blocks and metal fragments, and breakage or loosening of the suspension wire.
[0022] A subset of data on catenary equipment tracks and droppers is selected from the four-dimensional fusion dataset. Image frames are extracted from the image data. Through a trained neural network model, inference parameters are configured and the inference environment is initialized. The image frames are input into the neural network model to extract features and identify defects in the tracks and droppers. Based on the positioning information, the three-dimensional coordinate clusters of the defects in the laser point cloud data are extracted to determine the three-dimensional position coordinates of the defect area. The coordinates of the feature points at both ends of the track are extracted from the laser point cloud data. The actual effective track spacing is calculated using the track spacing calculation formula and the deviation is recorded by comparing it with the design standard. Based on the laser point cloud data and image data, the evaluation value of the dropper status is calculated using the dropper status evaluation formula to determine whether the dropper status is qualified. Finally, the defect information, parameter deviations, and status evaluation results are integrated to generate a parameter deviation list.
[0023] Furthermore, in the identification and parameter calculation module, the actual effective distance of the slide groove is calculated using the slide groove distance calculation formula, which is: ,in, D This represents the actual effective spacing of the chute. 、 The coordinates of the feature points at both ends of the chute are two-dimensional. Let be the defect area of the groove. The total area of the chute. This represents the damage impact coefficient.
[0024] Furthermore, in the identification and parameter calculation module, the evaluation value of the dropper state is calculated using the dropper state evaluation formula to determine whether the dropper state is qualified. The dropper state evaluation formula is as follows: ,in, This is the value for evaluating the condition of the dropper. This refers to the actual length of the suspension string. For the design length of the suspension string, This is due to the deviation in the installation angle of the dropper. Design speed for the line;
[0025] Pre-set according to design standards ,when At that time, the condition of the suspension wire is deemed acceptable.
[0026] Compared with existing technologies, this intelligent surveying system for overhead contact line equipment in tunnels has the following advantages:
[0027] I. This invention, by deploying multi-dimensional sensors, overcomes the limitations of traditional manual inspections or reliance on single sensors in environments such as insufficient lighting, high dust concentration, and large temperature and humidity fluctuations within tunnels. The laser scanner can penetrate the three-dimensional geometric framework of the dust acquisition equipment, and the BeiDou positioning component is bound to the acquisition action in real time, solving the positioning deviation problem caused by weak GPS signals in tunnels. The environmental sensor group monitors parameters such as temperature, humidity, and lighting in real time, providing environmental adaptation basis for data correction. Through a dynamic adaptation fusion coefficient formula, the sensor data weights are calculated to correct the original coordinate deviation and construct a four-dimensional fusion dataset. This solves the errors caused by environmental interference or data isolation in traditional methods, and addresses the problems of missed detections under low lighting conditions during manual visual inspections and the inability of mechanical measuring tools to correlate with environmental parameters.
[0028] Second, this invention uses a neural network model trained on historical labeled images to extract images and point cloud data of chutes and droppers from a four-dimensional fusion dataset. It identifies defect types such as edge breaks and foreign object blockages, and locates three-dimensional coordinates to automate and standardize defect identification, solving the problems of missed and false detections in manual inspections. It also constructs a three-dimensional mesh model using laser point clouds, embedding defect data and parameter deviations, and supports model scaling, rotation, sectioning, and color annotation, making the equipment status intuitive. Moreover, maintenance personnel can access evaluation reports, three-dimensional models, and adjustment suggestions through mobile devices to accurately locate defects and quantify deviations, prioritize the handling of out-of-tolerance areas, reduce unnecessary inspections, thereby reducing maintenance costs and improving railway operation safety and decision-making efficiency.
[0029] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description
[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.
[0031] Figure 1 A flowchart of an intelligent surveying system for overhead contact line equipment in a tunnel;
[0032] Figure 2 This is a framework diagram of an intelligent surveying system for overhead contact line equipment inside a tunnel. Detailed Implementation
[0033] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.
[0034] Example 1:
[0035] Multi-source data acquisition module: In the quarterly maintenance scenario of the overhead contact line equipment in high-speed railway tunnels, during the high-speed rail shutdown period, the maintenance vehicle is driven into the tunnel to be inspected. Multi-dimensional sensor groups are installed on the top and sides of the vehicle. The laser scanner performs a panoramic scan of the overhead contact line equipment in the tunnel, completely capturing the three-dimensional geometric framework of the overhead contact line from the tunnel entrance to the exit. The high-definition industrial camera focuses on key components such as the sliding groove connection nodes and the connection parts of the droppers and catenary cables of the overhead contact line, capturing detailed images under the assistance of tunnel lighting. The Beidou positioning component moves with the vehicle, binding the acquisition action and updating the location information in real time to ensure that each data point corresponds to the accurate mileage in the tunnel. The environmental sensor group continuously monitors the temperature, humidity, lighting conditions and dust in the tunnel, especially when passing near the tunnel ventilation opening, focusing on recording changes in environmental parameters. All the collected images, laser point clouds, positioning and environmental data are aggregated through the integration interface to form a raw dataset covering the entire tunnel overhead contact line.
[0036] Multi-source data fusion module: For images in the original dataset, it first removes noise caused by flickering lights in the tunnel, enhances the clarity of the chute edges and suspension details, and corrects image distortion caused by camera shooting angle. It then filters the laser point cloud data to remove noise from the tunnel walls and interference points from the vehicle's own structure. The module smooths the positioning data and environmental parameters to eliminate errors caused by instantaneous fluctuations. Based on the timestamps and coordinate transformations during data acquisition, it aligns the images, laser point clouds, positioning data, and environmental data spatiotemporally, matching multi-source data from the same time and location. Finally, using a dynamic adaptation fusion coefficient formula, combined with the integrity of the laser point cloud, image clarity, current environmental adaptability, and environmental sensitivity coefficient, it calculates the data fusion weight between the laser scanner and the high-definition industrial camera. The dynamic adaptation fusion coefficient formula is as follows: ,in, For the first Sensor-like data fusion weights, For indexes of different device types, For laser scanners, For high-definition industrial cameras, For the first The sensor rating ranges from 0 to 1. hour Score the integrity of the laser point cloud. hour Rate the image sharpness. For environmental adaptability, Assuming an environmental sensitivity coefficient, the three-dimensional coordinate collaborative correction formula is used to incorporate the BeiDou positioning deviation, the distance from the current location to the tunnel entrance, the attenuation coefficient, the temperature correction coefficient, and the difference between the measured and standard temperatures to correct the original coordinates. The three-dimensional coordinate collaborative correction formula is as follows: ,in, The original coordinates are (x, y, z). The corrected coordinates are (x, y, z). This represents the BeiDou positioning deviation value. Distance from the tunnel entrance The attenuation coefficient is... This is the temperature correction factor. Based on the temperature difference between the measured temperature and the standard temperature, a four-dimensional fusion dataset containing image, geometric, location, and environmental information was finally constructed, such as... Figure 1 As shown.
[0037] The identification and parameter calculation module calls a neural network model trained on historical defect images (covering common issues such as chute edge fractures, deformation, concrete block blockages, and hanger cable breakage and loosening). It selects subsets of chute and hanger cable data from the four-dimensional fusion dataset, extracts image frames, and inputs them into the neural network model. By analyzing image features, it identifies defects such as edge fractures caused by long-term vibrations in the chute, blockages caused by falling objects from the tunnel ceiling, and slack in the hangers due to tension changes. It also extracts the three-dimensional coordinate clusters of the defect areas from the laser point cloud data using location information to determine the three-dimensional position coordinates of each defect. Furthermore, it captures the coordinates of feature points at both ends of the chute from the laser point cloud data and calculates the actual effective chute spacing using the chute spacing calculation formula (combining the two-dimensional coordinates of feature points, the chute defect area, the total area, and the damage influence coefficient). The chute spacing calculation formula is as follows: ,in, D This represents the actual effective spacing of the chute. 、 The coordinates of the feature points at both ends of the chute are two-dimensional. Let be the defect area of the groove. The total area of the chute. The damage impact coefficient is calculated, and the deviation is recorded after comparison with the design standard. Simultaneously, combining laser point cloud and image data, the dropper condition assessment value is calculated using the dropper condition assessment formula (incorporating the ratio of the actual dropper length to the design length, installation angle deviation, and line design speed) to determine whether the dropper condition is qualified. The dropper condition assessment formula is as follows: ,in, This is the value for assessing the condition of the dropper. This refers to the actual length of the suspension string. For the design length of the suspension string, This is due to the deviation in the installation angle of the dropper. Design speed for the line;
[0038] And pre-set according to design standards ,when At that time, the condition of the dropper is determined to be qualified, and all defect information, parameter deviations and condition assessment results are integrated to generate a detailed list of parameter deviations.
[0039] Modeling and Interaction Module: Based on laser point cloud data, a 3D mesh model of the entire tunnel contact network equipment is constructed. Texture mapping technology is used to map the appearance details captured by high-definition industrial cameras onto the model, restoring the actual appearance of the equipment. Identified defect results and parameter calculation data are embedded into the model, and geometric errors between the model and the actual object are corrected to ensure model accuracy. A 3D interactive function is developed, allowing maintenance personnel to zoom, rotate 360°, and translate the model by operating the equipment, observing equipment details from different angles. A cross-sectional cutting function is added, allowing for customization of the cutting plane position, clearly displaying the internal structure of the chute, the connection nodes between the hangers and the contact wire, and other internal structures. Color coding rules are configured according to the parameter deviation list, marking out-of-tolerance chute areas and unqualified hangers in red, and normal areas in green, making the equipment status visually intuitive.
[0040] The data management and application output module integrates the original dataset, the four-dimensional fusion dataset, the parameter deviation list, and the three-dimensional mesh model to build a distributed database. This allows maintenance personnel to retrieve the information they need by multiple dimensions, such as defect type, equipment location, and inspection time. It also develops standardized interfaces to connect with the overhead contact line operation and maintenance management platform to achieve data sharing. Furthermore, it generates equipment status assessment reports (including defect distribution and parameter deviation analysis), three-dimensional model files, and deviation adjustment suggestions (such as spacing adjustment schemes for out-of-tolerance slides, cleaning suggestions for blocked slides, and replacement plans for unqualified hangers). Maintenance personnel can access and view these documents anytime through a mobile application, providing precise guidance for on-site maintenance work.
[0041] In summary, in the quarterly maintenance scenario of high-speed railway tunnel catenary, comprehensive raw data is acquired by deploying sensors through a multi-source data acquisition module. This data is then processed by a fusion module to construct a four-dimensional dataset. A trained neural network model is used to identify defects in the slipways and droppers, calculate key parameters, and generate a deviation list. Based on the deviation list, an annotated three-dimensional model is constructed. Finally, the data is integrated to form a report and recommendations. The entire process achieves accurate detection, intuitive presentation, and efficient management of the catenary equipment status, providing comprehensive data support for maintenance and improving the accuracy and efficiency of high-speed railway catenary operation and maintenance.
[0042] Example 2:
[0043] Multi-source data acquisition module: After the catenary equipment is installed in the subway tunnel, staff operate a special inspection vehicle to enter the tunnel for acceptance testing. The multi-dimensional sensor group on the inspection vehicle begins to work: the laser scanner performs a panoramic scan of the newly installed catenary, completely recording the three-dimensional geometry of the catenary from the starting point to the end point, including the installation position and connection relationship of each component; the high-definition industrial camera focuses on key installation parts such as the connection point between the chute and the fixed support at the top of the tunnel, and the connection point between the dropper and the busbar, capturing high-definition images to check the installation quality; the Beidou positioning component moves with the inspection vehicle, updating the location information in real time to ensure that the data can correspond to the specific mileage section in the tunnel; the environmental sensor group monitors the temperature, humidity, light, and dust conditions in the tunnel, especially in areas with more residual dust from tunnel construction, focusing on recording environmental parameters to assist in subsequent data processing; all collected data is aggregated through an integration interface to form a raw dataset containing images, laser point clouds, positioning, and environmental parameters.
[0044] Multi-source data fusion module: Processes images in the original dataset, removing noise such as image blurring caused by construction dust residue, enhancing the clarity of details such as installation gaps and connecting bolts, correcting distortion caused by camera angles, filtering and denoising laser point cloud data, removing interfering point clouds such as temporary construction supports and tools, smoothing positioning data and environmental parameters to eliminate the impact of instantaneous fluctuations, and achieving spatiotemporal alignment of multi-source data based on timestamps and coordinate transformations to ensure accurate matching of images, point clouds, positioning, and environmental data at the same location. Finally, it calculates the fusion weight between the laser scanner and the high-definition industrial camera using a dynamic adaptation fusion coefficient formula, which is: Then, the original coordinates are corrected using the three-dimensional coordinate co-correction formula, which is as follows: Ultimately, a four-dimensional fusion dataset was constructed, providing an accurate data foundation for subsequent detection.
[0045] The identification and parameter calculation module uses a trained neural network model to extract relevant data on the chute and hanger from the four-dimensional fusion dataset. By analyzing image features, it identifies problems such as concrete blocks accidentally falling into the chute during installation and excessive angle deviations in the hanger installation. It also extracts feature points at both ends of the chute from the laser point cloud data and calculates the actual chute spacing using the chute spacing calculation formula: A comparison with the design values revealed that some spacing did not meet the standards. The evaluation value was calculated using the dropper condition assessment formula (combining actual length, design length, installation angle deviation, and line design speed). The dropper condition assessment formula is as follows: Pre-set according to design standards ,when At that time, the condition of the dropper is determined to be qualified, and all test results are integrated to generate a parameter deviation list, recording in detail the non-conformities, such as... Figure 2 As shown.
[0046] Modeling and Interaction Module: Based on laser point clouds, a 3D mesh model of the contact network equipment is constructed. The appearance of the newly installed equipment is restored through texture mapping. Defect identification results and parameter calculation data are embedded into the model to correct geometric errors. A 3D interactive function is developed to allow acceptance personnel to view the equipment installation details by zooming, rotating, and translating the model. A cross-sectional cutting function is added to customize the plane, allowing users to view the connection structure between the slide and the fixed support, the installation nodes of the suspension cable, and other internal conditions. According to the color marking rules, out-of-tolerance areas are marked in red, and normal areas are marked in green, which intuitively shows the equipment installation quality.
[0047] Data Management and Application Output Module: Integrates raw datasets, four-dimensional fusion datasets, parameter deviation lists, and three-dimensional mesh models to construct a distributed database. It supports multi-dimensional retrieval by installation mileage, component type, defect type, etc., and connects with the engineering acceptance platform through standardized interfaces to generate equipment acceptance reports, three-dimensional model files, and adjustment suggestions (such as cleaning plans for blocked chutes, reinstallation guidelines for chutes with out-of-tolerance spacing, and correction methods for angle deviation droppers). Staff can access and view these documents through a mobile application, serving as an important basis for determining whether the overhead contact line equipment has passed acceptance.
[0048] In summary, in the acceptance scenario of newly built subway tunnel catenary systems, multi-source data from the installed equipment is collected by sensors mounted on the inspection vehicle. This data is then fused to obtain a four-dimensional dataset. A neural network model is used to identify installation defects, calculate spacing and dropper state parameters, and generate a deviation list. Subsequently, a color-coded three-dimensional model is constructed. After data integration, an acceptance report and adjustment suggestions are generated, supporting multi-dimensional retrieval and platform integration. This process comprehensively verifies the equipment installation quality and provides a reliable basis for acceptance work through intuitive three-dimensional models and detailed data, ensuring the initial installation accuracy of the subway catenary.
[0049] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.
Claims
1. An intelligent surveying system for overhead contact line equipment in tunnels, characterized in that, The system includes: Multi-source data acquisition module: Deploys multi-dimensional sensors to acquire multi-source data from the overhead contact line, and aggregates and processes the multi-source data through an integrated interface to form a raw dataset; Multi-source data fusion module: This module preprocesses the multi-source data in the original dataset, corrects the original coordinates by fusing data weights, and constructs a four-dimensional fused dataset encompassing image, geometry, location, and environment. It performs noise reduction, enhancement, and distortion correction on the image data in the original dataset, filters and denoises the laser point cloud data, and smooths and corrects the positioning and environmental parameter data. It also performs spatiotemporal alignment of the data based on timestamps and coordinate transformations. Finally, it calculates the data fusion weights of the laser scanner and high-definition industrial camera using a dynamic adaptation fusion coefficient formula, corrects the original coordinates using a three-dimensional coordinate collaborative correction formula, and constructs a four-dimensional fused dataset encompassing image, geometry, location, and environment. The dynamic adaptation fusion coefficient formula is as follows: ,in, For the first Sensor-like data fusion weights, For indexes of different device types, For laser scanners, For high-definition industrial cameras, For the first The sensor rating ranges from 0 to 1. hour Score the integrity of the laser point cloud. hour Rate the image sharpness. For environmental adaptability, The environmental sensitivity coefficient is used; its three-dimensional coordinate collaborative correction formula is: ,in, The original coordinates are (x, y, z). The corrected coordinates are (x, y, z). This represents the BeiDou positioning deviation value. Distance from the tunnel entrance The attenuation coefficient is... This is the temperature correction factor. This represents the temperature difference between the measured temperature and the standard temperature. Identification and parameter calculation module: Through the neural network model trained on the training set, it determines the defects of key components of the overhead contact line equipment and outputs the type and three-dimensional position coordinates of the effective defects. At the same time, it calculates the sliding groove spacing and the evaluation value of the dropper installation status, and generates a parameter deviation list after comparing it with the design standard. Modeling and Interaction Module: Based on the parameter deviation list, a 3D mesh model of the contact network equipment is constructed using laser point cloud as the basis. The appearance of the equipment is restored through texture mapping, defect identification results and parameter calculation data are embedded, the geometric error between the 3D mesh model and the entity is corrected, and 3D interactive functions and cross-section cutting functions are developed. Color marking is configured according to the parameter deviation list. Data management and application output module: Integrates raw datasets, four-dimensional fusion datasets, parameter deviation lists and three-dimensional mesh models to build a distributed database, supports multi-dimensional retrieval, develops standardized interfaces to connect with the construction and maintenance platform, and generates equipment status assessment reports, three-dimensional model files and deviation adjustment suggestion files, which can be accessed and viewed through mobile applications.
2. The intelligent mapping system for overhead contact line equipment in a tunnel according to claim 1, characterized in that, The multi-source data acquisition module is equipped with multi-dimensional sensors, including a laser scanner, a high-definition industrial camera, a Beidou positioning component, and an environmental sensor group. The laser scanner performs a panoramic scan of the target area to obtain the three-dimensional geometric framework of the overhead contact line equipment. The scanning rate is ≥1 million points / second, and the ranging accuracy is ≤±0.5mm. The high-definition industrial camera is used to capture images of key components of the overhead contact line equipment, including slide rails and suspension cables, with a resolution of ≥12 million pixels. The BeiDou positioning component is bound to the data acquisition action in real time, updates the location information periodically, has a positioning accuracy of ≤1m, and supports differential positioning. The environmental sensor group includes a temperature and humidity sensor, a light sensor, and a dust concentration sensor, which collect parameters such as temperature, relative humidity, light intensity, and dust concentration inside the tunnel. The collected multi-source data is aggregated and processed through an integration interface to form a raw dataset containing image data, laser point cloud data, positioning data, and environmental parameter data.
3. The intelligent surveying system for overhead contact line equipment in a tunnel according to claim 1, characterized in that, In the identification and parameter calculation module, the training set for training the neural network model includes historical labeled images that cover common defects, including edge breakage and deformation in chute damage, foreign object blockage in the chute caused by concrete blocks and metal debris, and defects such as breakage or loosening of the suspension wire. A subset of data on catenary equipment tracks and droppers is selected from the four-dimensional fusion dataset. Image frames are extracted from the image data. Through a trained neural network model, inference parameters are configured and the inference environment is initialized. The image frames are input into the neural network model to extract features and identify defects in the tracks and droppers. Based on the positioning information, the three-dimensional coordinate clusters of the defects in the laser point cloud data are extracted to determine the three-dimensional position coordinates of the defect area. The coordinates of the feature points at both ends of the track are extracted from the laser point cloud data. The actual effective track spacing is calculated using the track spacing calculation formula and the deviation is recorded by comparing it with the design standard. Based on the laser point cloud data and image data, the evaluation value of the dropper status is calculated using the dropper status evaluation formula to determine whether the dropper status is qualified. Finally, the defect information, parameter deviations, and status evaluation results are integrated to generate a parameter deviation list.
4. The intelligent mapping system for overhead contact line equipment in a tunnel according to claim 3, characterized in that, In the identification and parameter calculation module, the actual effective distance of the slide groove is calculated using the slide groove distance calculation formula, which is: ,in, D This represents the actual effective spacing of the chute. 、 The coordinates of the feature points at both ends of the chute are two-dimensional. Let be the defect area of the groove. The total area of the chute. This represents the damage impact coefficient.
5. The intelligent mapping system for overhead contact line equipment in a tunnel according to claim 3, characterized in that, In the identification and parameter calculation module, the evaluation value of the dropper state is calculated using the dropper state evaluation formula to determine whether the dropper state is qualified. The dropper state evaluation formula is as follows: ,in, This is the value for assessing the condition of the dropper. This refers to the actual length of the suspension string. For the design length of the suspension string, This is due to the deviation in the installation angle of the dropper. Design speed for the line; Pre-set according to design standards ,when At that time, the condition of the suspension wire is deemed acceptable.