Railway facility spatial feature big data analysis method

Abstract

The invention provides a railway facility spatial feature big data analysis method. The method comprises the steps of obtaining railway spatial data collected from multiple sources; the railway spatial data with the unified railway element data structure obtained through construction is stored in a geographic information database and / or a big data storage platform after data quality checking; spatial analysis, geographic information system analysis and track network analysis are carried out on the stored railway spatial data in the historical periods, and railway facility spatial features including spatial geometric features, spatial position features and facility state features are extracted; learning the railway facility spatial features through a clustering analysis model, and clustering facility safety accidents to identify potential safety risks; and through a supervised learning model, learning facility state features included in the railway facility spatial features in the historical time period, so as to analyze the fault mode of the railway facility. The method can analyze railway facility spatial characteristics, and lays a foundation for intelligent railways and digital railways.

Description

Technical Field

[0001] The present invention relates to the technical field of railway facility big data analysis, and in particular to a railway facility spatial feature big data analysis method. Background Art

[0002] With the emergence of various new spatial data collection methods, the number of data collection channels has increased, and more and more railway spatial data have been collected through new measurement technologies. At present, railway spatial data has the characteristics of large data volume, multiple data types, and complex spatial structure. Combined with big data analysis, it can realize massive spatial data mining, monitoring and early warning, decision support, etc. Typical applications of railway spatial big data analysis include facility failure prediction, EMU productivity layout, passenger and freight market price monitoring, logistics market demand monitoring, engineering construction progress organization, and rail quality evaluation. Figure 2 The following is a typical application of railway spatial big data analysis. With the help of big data analysis technology, spatial geometry analysis technology, and image feature extraction technology, we actively explore the spatial feature analysis method of railway big data infrastructure, and have achieved remarkable results in the fields of railway safety risk warning, peripheral foreign body intrusion prediction, and infrastructure identification. It can be applied to facility failure prediction, EMU productivity layout, passenger and freight market price monitoring, engineering construction progress organization, rail quality evaluation, and logistics time demand monitoring. These innovative practices not only provide new impetus for the progress of the railway industry, but also open up new prospects for the development of my country's overall transportation system. Among them, railway facility spatial data contains location information, spatial distribution, spatial association, spatial structure, spatial attributes, spatial changes, and spatial key points, and has rich characteristics. Railway facility spatial data is of great significance to the management, planning, and optimization of railway transportation systems.

[0003] However, the existing railway facility spatial feature analysis method lacks a standardized overall architecture and feature analysis process. Therefore, how to provide a standardized, unified and widely applicable railway facility spatial feature analysis method is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a method for analyzing big data of spatial characteristics of railway facilities to eliminate or improve one or more defects existing in the prior art.

[0005] One aspect of the present invention provides a method for analyzing big data of spatial features of railway facilities, the method comprising the following steps: obtaining railway spatial data collected from multiple sources, preprocessing the obtained railway spatial data, matching key fields of railway spatial data of different data types using an attribute fuzzy matching model, and constructing railway spatial data with a unified railway element data structure; storing the constructed railway spatial data with a unified railway element data structure in a geographic information database and / or a big data storage platform after data quality verification; performing spatial analysis, geographic information system analysis, and track network analysis on the stored railway spatial data of a historical period, and extracting railway facility spatial features including spatial geometric features, spatial location features, and facility status features; learning the railway facility spatial features through a clustering analysis model, clustering facility safety accidents, and identifying potential safety risks; and learning the facility status features within the historical period contained in the railway facility spatial features through a supervised learning model, and analyzing the failure modes of the railway facilities.

[0006] In some embodiments of the present invention, the collection sources of the railway spatial data include multiple types of drones, vehicle-mounted equipment, sensors, railway monitoring systems, geographic information systems and Beidou collection devices, and the railway spatial data include multiple types of equipment status data, operation record information, station information and line information; the types of the railway spatial data include shp, gdb and tif formats.

[0007] In some embodiments of the present invention, the step of preprocessing the acquired railway spatial data includes: performing data cleaning and formatting processing on the acquired railway spatial data, and performing unified processing on the railway spatial data by means of geographic calibration and coordinate system conversion.

[0008] In some embodiments of the present invention, the step of verifying the data quality of railway spatial data includes: statistically analyzing the distribution information of collection points of railway spatial data, analyzing the distance deviation and line continuity of the collection points, and screening out railway spatial data whose collection point quality is lower than a preset threshold according to preset rules; verifying the integrity and accuracy of the attribute information content structure of railway spatial data with a unified railway element data structure, and screening out incomplete or inaccurate railway spatial data.

[0009] In some embodiments of the present invention, the method also includes a step of fusing spatial features of railway facilities, including: learning the stored railway spatial data of historical periods through a neighborhood aggregation model and a recurrent neural network model to obtain node features and sequence features of the railway spatial data; using a location-aware multi-head self-attention network to discover hidden features of railway facilities; inputting the node features, the sequence features, the hidden features, the spatial geometry features, the spatial position features and the facility status features into a feature fusion model to obtain fused spatial features of the railway facilities.

[0010] In some embodiments of the present invention, the method also includes the step of railway environment safety situation awareness analysis, including: using railway spatial data collected from multiple sources to establish a railway environment infrastructure and safety range identification model, based on fixed-focus video monitoring, analyzing the pixel value change rate and grayscale projection of the monitored images to achieve railway facility identification; using railway spatial data collected from multiple sources, through a railway perimeter feature recognition model based on transfer learning, to detect and identify features with safety hazards on the railway perimeter.

[0011] In some embodiments of the present invention, the step of learning the facility status characteristics within the historical time period contained in the spatial characteristics of the railway facilities through a supervised learning model to analyze the failure modes of the railway facilities includes: using the supervised learning model to analyze and model the historical failure data of the railway facilities, predicting the failure modes of the railway facilities in the future time period, using a classification algorithm to train and identify different types of failure modes, and using a regression model to predict the future time period when the failure will occur.

[0012] In some embodiments of the present invention, the railway spatial data also includes passenger flow data of the railway system, and the method also includes: using a clustering model or an association rule mining model to analyze the passenger flow data of different areas of the railway system, and learning the passenger flow patterns and correlations between different areas; using a spatial regression model to learn the correlation between passenger flow and other railway spatial data, and training to predict future passenger flow.

[0013] In some embodiments of the present invention, the method also includes: when the collection point is fixed-point video surveillance along the railway, based on the spatiotemporal information of the perimeter environment obtained from the fixed-point video surveillance along the railway, a recognition model for identifying foreign object intrusion detection is constructed based on a deep learning model to achieve rapid positioning, identification and early warning of foreign objects; the method also includes the step of evaluating the model used for spatial feature analysis of railway facilities using a pre-trained maturity assessment model; wherein the maturity assessment model includes three dimensions: analysis application, analysis quality, and data architecture, each dimension is divided into 5 levels, the analysis application dimension includes five levels: descriptive application, diagnostic application, predictive application, exploratory application and decision support application, the analysis quality dimension is divided into five levels: descriptive analysis, diagnostic analysis, predictive analysis, control analysis and decision support analysis, and the data architecture dimension is divided into five levels: single structure, comprehensive, artificial organization, partially autonomous organization and autonomous organization.

[0014] In some embodiments of the present invention, the method further comprises: visually presenting the analysis results based on the spatial characteristics of railway facilities.

[0015] The railway facility spatial feature big data analysis method proposed in the present invention can extract multi-dimensional railway facility spatial features based on the proposed railway facility spatial feature analysis method, and train a big data model based on the extracted railway facility spatial features, laying the foundation for intelligent railways and digital railways.

[0016] Additional advantages, purposes, and features of the present invention will be described in part in the following description, and will become apparent to those skilled in the art after studying the following, or may be learned from the practice of the present invention. The purposes and other advantages of the present invention may be achieved and obtained by the structures specifically indicated in the specification and the accompanying drawings.

[0017] Those skilled in the art will appreciate that the objectives and advantages that can be achieved with the present invention are not limited to the above specific description, and the above and other objectives that can be achieved by the present invention will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of the present application, and do not constitute a limitation of the present invention. In the drawings:

[0019] Figure 1 This is a flow chart of a method for analyzing big data of railway facility spatial characteristics in one embodiment of the present invention.

[0020] Figure 2 It is a typical application of railway spatial big data analysis.

[0021] Figure 3The figure is a schematic diagram of a maturity assessment model for railway facility spatial feature analysis in one embodiment of the present invention.

[0022] Figure 4 This is a general architecture diagram of the basic platform of the railway facility spatial feature analysis system in one embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0024] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0025] It should be emphasized that the term “include / comprises” when used herein refers to the presence of features, elements, steps or components, but does not exclude the presence or addition of one or more other features, elements, steps or components.

[0026] It should also be noted that, unless otherwise specified, the term “connection” herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.

[0027] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0028] In order to overcome the problems existing in the prior art, this application proposes a big data analysis method for the spatiotemporal characteristics of railway facilities, which integrates technologies such as big data analysis and spatial geometry analysis, and conducts analysis and research on the spatiotemporal characteristics of railway spatiotemporal big data infrastructure identification, feature extraction, hierarchical classification and data mining.

[0029] The spatial characteristics of railway facilities are mainly aimed at business needs such as railway line operation, equipment and facility management, and train driving safety, covering multiple aspects and multiple levels. The analysis of the spatial characteristics of railway facilities includes: (1) Station and line distribution characteristics, which mainly studies the distribution of railway stations and lines, including information such as the distance between stations, station density, line length, and line safety coverage. (2) Railway network structure characteristics, which mainly analyzes the topological structure of the railway network, network connectivity, and the connection between different lines. (3) Facilities and equipment spatial layout characteristics, which analyzes the spatial layout of railway facilities and equipment, such as the form and spatial distribution of signal systems, tracks, and vehicle maintenance centers.

[0030] The method for analyzing the big data of the spatial features of railway facilities proposed in this application has the following route for analyzing the big data of the spatial features of railway facilities: (1) Based on the railway spatiotemporal big data such as remote sensing images, photographic images and surveillance videos collected throughout the railway, through data preprocessing, image segmentation and regional calibration, spatial features are extracted using structure, color, texture, motion state, etc., the extracted features are screened and processed, the most representative railway element features are selected, and dimensionality reduction is performed. (2) Machine learning models, deep learning models, etc. are applied to use the extracted features for tasks such as pattern recognition, classification, target detection or anomaly detection. The analysis results are evaluated and verified to check the accuracy and robustness of the model. The railway mileage linear reference system and Euclidean buffer analysis and other technologies are used to accurately locate infrastructure such as railway lines and stations, analyze environmental features along the line, and analyze the safety range of railway lines. Finally, the results are applied to actual scenarios, such as railway operation monitoring and safety management. (3) A maturity assessment model for the analysis of spatial features of railway facilities is constructed, such as Figure 3 As shown in the figure, the railway facility space big data analysis model is divided into 1-5 levels through the three dimensions of analysis application, analysis quality, and data architecture, and the maturity of the big data model is comprehensively evaluated. Among them, the big data model analysis application is divided into five levels: descriptive application, diagnostic application, predictive application, exploratory application, and decision support application; the analysis quality is divided into five levels: descriptive analysis, diagnostic analysis, predictive analysis, control analysis, and decision support analysis; the data architecture is divided into five levels: single structure, comprehensive, artificial organization, partially autonomous organization, and autonomous organization.

[0031] Figure 1 This is a flow chart of a method for analyzing big data of railway facility spatial characteristics in one embodiment of the present invention. Figure 1It includes five steps: data collection and preprocessing, spatial data storage and management, railway facility spatial feature extraction, data analysis and mining, and result visualization and application. A data quality management system is established to control the quality of the spatial data stored and managed. In data collection and processing, equipment status data from sensors, operation records from monitoring systems, site information and line information from geographic information systems, and remote sensing image information from Beidou acquisition devices are collected. In spatial data storage and management, data quality rules are set, and geographic information databases, big data storage platforms, and relational databases are set in the storage database to store different types of information. In the extraction of railway facility spatial features, features include spatial geometric features and spatial position, i.e., status features. Spatial geometric features include facility morphology and line safety range. Spatial range and status include equipment status and facility distribution. Data analysis and mining include buffer analysis, line mileage linear analysis, data mining, road network analysis, and trend analysis and prediction. Result visualization and application include multiple application directions such as infrastructure monitoring, safety area monitoring, safety warning, transportation organization statistics, and transportation organization statistics.

[0032] The railway facility spatial feature big data analysis method proposed in the present invention comprises the following steps:

[0033] Step S110: Acquire railway spatial data collected from multiple sources, pre-process the acquired railway spatial data, use the attribute fuzzy matching model to match key fields of railway spatial data of different data types, and construct railway spatial data with a unified railway element data structure.

[0034] In the specific implementation process, railway spatial data may include railway spatial data collected by sensors, monitoring systems, and geographic information systems. Generally speaking, railway facility spatial data may include environmental basic spatial data and railway facility-specific spatial data. Environmental basic spatial data include geological landforms such as topography, landforms, soil, and rocks, as well as natural resources such as water resources, mineral resources, and vegetation distribution. Railway facility spatial data may also refer to the spatial distribution data of various fixed equipment and mobile equipment within the railway system, including line data, bridge and tunnel data, station data, signal equipment data, and mobile equipment data. In some other embodiments of the present invention, railway spatial data may also include data content such as stations, lines, equipment status, and operation records.

[0035] Step S120: The constructed railway spatial data with a unified railway element data structure is stored in a geographic information database and / or a big data storage platform after data quality verification.

[0036] Step S130: Perform spatial analysis, geographic information system analysis and track network analysis on the stored railway spatial data of the historical period, and extract the railway facility spatial characteristics including spatial geometric characteristics, spatial location characteristics and facility status characteristics.

[0037] In the specific implementation process, the railway space characteristics include station and line distribution characteristics, railway network structure characteristics and facility equipment space layout characteristics. In another embodiment, the railway space characteristics may also include facility morphology characteristics, equipment operation status characteristics and safety range characteristics.

[0038] Step S140: learning the spatial characteristics of the railway facilities through a cluster analysis model, clustering facility safety accidents to identify potential safety risks; learning the facility status characteristics within the historical period contained in the spatial characteristics of the railway facilities through a supervised learning model to analyze the failure modes of the railway facilities.

[0039] The railway facility spatial feature big data analysis method proposed in the present invention can extract multi-dimensional railway facility spatial features based on the proposed railway facility spatial feature analysis method, and train a big data model based on the extracted railway facility spatial features, laying the foundation for intelligent railways and digital railways.

[0040] In some embodiments of the present invention, the sources of railway spatial data collection include multiple types of drones, vehicle-mounted equipment, sensors, railway monitoring systems, geographic information systems and Beidou collection devices, and the railway spatial data include multiple types of equipment status data, operation record information, station information and line information.

[0041] In some embodiments of the present invention, the types of railway spatial data include shp, gdb and tif formats.

[0042] By adopting the embodiment of the invention, it is possible to perform big data analysis on the spatial characteristics of railway facilities by combining railway spatial data collected from different sources, thereby being applicable to railway facility management tasks in different scenarios.

[0043] In some embodiments of the present invention, the step of preprocessing the acquired railway spatial data includes: performing data cleaning and formatting processing on the acquired railway spatial data, and performing unified processing on the railway spatial data by means of geographic calibration and coordinate system conversion.

[0044] By adopting this embodiment of the invention, the railway spatial data used for big data analysis can be made purer, which is beneficial to improving the model training effect.

[0045] In some embodiments of the present invention, the step of verifying the data quality of railway spatial data includes: statistically analyzing the distribution information of collection points of railway spatial data, analyzing the distance deviation and line continuity of the collection points, and screening out railway spatial data whose collection point quality is lower than a preset threshold according to preset rules; verifying the integrity and accuracy of the attribute information content structure of railway spatial data with a unified railway element data structure, and screening out incomplete or inaccurate railway spatial data.

[0046] By adopting the embodiment of the invention, it is possible to further check the data quality of railway spatial data, thereby further improving the effect of subsequent model training.

[0047] In some embodiments of the present invention, the method also includes a step of fusing spatial features of railway facilities, including: (1) learning the railway spatial data of the stored historical period through a neighborhood aggregation model and a recurrent neural network model to obtain the node features and sequence features of the railway spatial data; (2) using a location-aware multi-head self-attention network to discover the hidden features of railway facilities; (3) inputting the node features, the sequence features, the hidden features, the spatial geometry features, the spatial location features and the facility status features into a feature fusion model to obtain the fused spatial features of the railway facilities.

[0048] By adopting the embodiment of the invention, the quality of the spatial characteristics of railway facilities can be improved, and the effect of the final big data analysis of the spatial characteristics of railway facilities can be improved.

[0049] In some embodiments of the present invention, the method also includes the step of railway environment safety situation awareness analysis, including: (1) using railway spatial data collected from multiple sources to establish a railway environment infrastructure and safety range identification model, based on fixed-focus video monitoring, analyzing the pixel value change rate and grayscale projection of the monitored image to achieve railway facility identification; (2) using railway spatial data collected from multiple sources, through a railway perimeter feature recognition model based on transfer learning, to detect and identify features with safety hazards on the railway perimeter.

[0050] By adopting the embodiment of the invention, it is possible to use railway spatial data collected from multiple sources to conduct situational awareness and analysis of railway environmental safety and identify factors in the environment that affect safety.

[0051] In some embodiments of the present invention, the step of learning the facility status characteristics within the historical time period contained in the spatial characteristics of the railway facilities through a supervised learning model to analyze the failure mode of the railway facilities includes: using the supervised learning model to analyze and model the historical failure data of the railway facilities, predicting the failure mode of the railway facilities in the future time period, using the classification algorithm to train and identify different types of failure modes, and using the regression model to predict the future time period when the failure will occur.

[0052] By adopting the embodiment of the invention, the status of railway facilities can be monitored and predicted from the dimension of big data analysis, and possible railway facility failures can be responded to and handled in a more timely manner.

[0053] In some embodiments of the present invention, the railway spatial data also includes passenger flow data of the railway system, and the method further includes: (1) using a clustering model or an association rule mining model to analyze the passenger flow data of different areas of the railway system to learn the passenger flow patterns and correlations between different areas; (2) using a spatial regression model to learn the correlation between passenger flow and other railway spatial data, and training to predict future passenger flow.

[0054] By adopting the embodiment of the invention, it is possible to monitor and predict the passenger flow data of the railway system on a large scale, thereby responding to changes in passenger flow data that have occurred or may occur in a more timely manner.

[0055] In some embodiments of the present invention, the method also includes: when the collection point is fixed-point video surveillance along the railway, based on the spatiotemporal information of the perimeter environment obtained from the fixed-point video surveillance along the railway, a recognition model for identifying foreign object intrusion detection is constructed based on a deep learning model to achieve rapid positioning, identification and early warning of foreign objects.

[0056] In some embodiments of the present invention, the method also includes the step of evaluating the model for spatial feature analysis of railway facilities using a pre-trained maturity assessment model; wherein the maturity assessment model includes three dimensions: analysis application, analysis quality, and data architecture, each dimension is divided into 5 levels, the analysis application dimension includes five levels: descriptive application, diagnostic application, predictive application, exploratory application, and decision support application, the analysis quality dimension includes five levels: descriptive analysis, diagnostic analysis, predictive analysis, control analysis, and decision support analysis, and the data architecture dimension includes five levels: single structure, comprehensive, artificial organization, partially autonomous organization, and autonomous organization. Figure 3 FIG. 1 is a schematic diagram of a maturity assessment model for analyzing the spatial characteristics of railway facilities in one embodiment of the present invention. The structure of the maturity assessment model is as follows: Figure 3 shown.

[0057] By adopting this embodiment, the maturity of the effect of the railway facility spatial characteristic analysis can be evaluated, and timely adjustments can be made to models or analysis methods that are not mature enough.

[0058] Furthermore, for each of the above invention embodiments, the method also includes: visually presenting the analysis results based on the spatial characteristics of railway facilities.

[0059] In some embodiments of the present invention, Figure 4The figure is an overall architecture diagram of the basic platform of the railway facility spatial feature analysis system in one embodiment of the present invention. The dynamic data architecture of the basic platform includes a raw data layer, a data processing layer, a mechanism model layer and an application layer. The data foundation of the architecture is called metadata, and the types of metadata include structured data, unstructured data, remote sensing images and training data sets. The processing of the data processing layer includes data cleaning, geographic calibration, formatting, projection transformation, data fusion and data classification. The tasks of the mechanism model layer include feature extraction, as well as security range analysis, detection identification and prediction tasks based on the extracted features. The tasks of the application layer include security situation awareness of the railway environment, intelligent analysis of foreign body intrusion, and spatial analysis of remote sensing image feature level. The data processing layer provides organizational processing of raw data, the mechanism model layer generates a mechanism model for data formatting, and the mechanism model layer provides business applications provided by interface combinations to the application layer.

[0060] In the specific implementation process, the following steps (1)-(6) constitute the overall process of big data analysis of railway facility spatial characteristics. The entire process is an iterative process, and in actual application, it will be customized, adjusted and optimized according to different needs and scenarios.

[0061] (1) Step 1: Data collection and preprocessing

[0062] Obtain railway spatial data from various sources (sensors, monitoring systems, geographic information systems, etc.), including stations, lines, equipment status, operation records, etc. Data types include shp, gdb, tif and other formats. Data needs to be cleaned, formatted and preprocessed to ensure data quality and availability. Since railway spatial data has the characteristics of large data volume, multiple versions, and close association with lines, the attribute fuzzy matching method is used to match key fields of multiple versions of different types of data to build a unified railway feature data structure; geographic calibration, coordinate system conversion and other methods are used to preprocess railway spatial data such as lines, stations, equipment and facilities to build a unified railway spatial location.

[0063] (2) Step 2: Spatial Data Storage and Management

[0064] Effectively store and manage railway spatial data. Due to the large amount of railway spatial data, the spatial data of the line, station, and equipment and facilities along a railway line alone can reach hundreds of GB of storage requirements. Usually, geographic information databases or big data storage platforms are used, and data quality verification rules are constructed, including: analyzing the distance deviation and line continuity of the collected coordinate points along the railway line, screening out data problems such as collection point offset, discontinuity, and loss; analyzing whether the content structure of railway element attribute information is complete and accurate, and screening out problems such as incomplete and incorrect element information. Ensure the scalability, efficiency, and security of data.

[0065] (3) Step 3: Extraction of spatial features of railway facilities

[0066] Analyze and process the stored railway spatial data to extract the spatial characteristics of railway facilities. This includes spatial analysis, geographic information system (GIS) analysis, and track network analysis of the data to extract characteristic information such as facility form, equipment operating status, and safety range.

[0067] (4) Step 4: Feature Fusion and Integration

[0068] Integrate and fuse multiple spatial features extracted from different sources to establish a comprehensive, multi-dimensional railway spatial feature data set for subsequent analysis and application. Neighborhood aggregation and recurrent neural networks are used to learn node features and sequence features of railway spatial data respectively, and then a custom similarity calculation formula is used to retrieve historical achievement data and standard sample data as context to solve the problem of scarce information. Then, the location-aware multi-head self-attention network is used to fully explore the hidden features of railway elements. Finally, the model realizes the fusion processing of railway spatial features.

[0069] (5) Step 5: Spatial data analysis and mining

[0070] This stage involves in-depth analysis of the integrated spatial feature data, using techniques such as spatial data mining, machine learning, and buffer analysis to discover patterns, regularities, and trends in the data, such as facility safety risk analysis, equipment failure modes, and passenger flow trends.

[0071] For facility safety risk analysis, cluster analysis technology is used to cluster facility safety accidents in spatial feature data to identify potential safety risk areas. In this way, areas with frequent accidents can be discovered to help decision makers take targeted preventive measures. At the same time, based on the location of the facility, a buffer zone centered on the facility is created, and the number and type of safety accidents within the buffer zone are analyzed. This helps to identify potential safety risks around the facility and develop targeted management strategies.

[0072] For equipment failure mode analysis, supervised learning algorithms are used to analyze and model the historical failure data of the equipment to predict the future failure mode of the equipment. Classification algorithms can be used to identify different types of failure modes, and regression algorithms can be used to predict the possible time of failure. The equipment failure data is converted into time series data, and then time series analysis techniques such as ARIMA models are used to analyze and predict the patterns and trends of equipment failures. This helps to detect possible equipment failures in advance and take preventive maintenance measures.

[0073] For passenger flow trend analysis, cluster analysis or association rule mining techniques are used to analyze passenger flow data and discover passenger flow patterns and correlations between different regions. This can help railway managers understand the distribution of passenger flow and optimize train operation plans and vehicle configurations. Spatial regression models are used to analyze the relationship between passenger flow and various spatial features (such as station location, surrounding environment, etc.). This helps predict future passenger flow and provide decision support for passenger flow management.

[0074] (6) Step 6: Results visualization and application

[0075] The results of the analysis are presented in a visual way, including vector maps, chart statistics, etc., to provide railway managers, decision makers and other stakeholders with intuitive and easy-to-understand information to support decision-making and optimize management. For train operation monitoring, vector maps are used to display the current location of all trains and their paths in the railway network. Different types of trains (freight, passenger, etc.) can be marked on the map, and color coding is used to indicate the operating status (normal, delayed, canceled, etc.). For railway facility maintenance, combined with geographic information system technology, an interactive map is created to show the location and status of facilities such as railway lines, stations, bridges, and tunnels. The maintenance needs and plans of different facilities are displayed through coding or symbol marking. Heat maps are used in combination with remote sensing images to show the maintenance urgency of various areas or facilities. For safety accident analysis, the location of safety accidents is drawn on the map, and different symbols are used to indicate the types of accidents (derailment, fire, collision, etc.). This can help railway managers identify accident-prone areas and take preventive measures.

[0076] In one embodiment of the present invention, the railway environment security situation perception step is included: based on comprehensive analysis of railway multi-source spatiotemporal big data obtained by various collection methods such as drones and vehicles, a railway environment infrastructure and safety range identification model is established, and a railway facility identification method based on fixed-focus video monitoring, analyzing pixel value change rate and grayscale projection is proposed to realize the monitoring and identification of infrastructure targets within the safety range of high-speed railways; a railway perimeter feature identification algorithm based on transfer learning is proposed to realize the detection and identification of features with safety hazards in the railway perimeter; a line mileage linear analysis and buffer zone analysis algorithm are proposed to realize the calculation of the safety of facilities along the railway line; a multi-dimensional and multi-perspective railway environment and infrastructure big data fusion mechanism is studied, and the information characteristics of various spatiotemporal data are fully combined to construct a railway facility and safety range identification model supported by uncertainty reasoning through complementary, collaborative and fusion processing. It is used to solve the problems of foreign body intrusion monitoring in the safety range under special terrain, safety hazard detection of perimeter environment railway facilities, etc., and accurately identify the safety situation of the railway environment. It is also possible to establish a railway environment infrastructure and safety range identification model based on railway multi-source spatiotemporal big data by building a railway facility and safety range identification model, and then identify the safety risks in the railway environment.

[0077] In one embodiment of the present invention, it includes a step of intelligent analysis of foreign body intrusion: based on the massive spatiotemporal information of the perimeter environment obtained from fixed-point video surveillance along the railway, a foreign body intrusion image intelligent analysis model based on deep learning is constructed, and an identification algorithm for foreign body target intrusion detection based on image intelligent recognition and adapted to day and night is proposed to achieve rapid positioning, identification, and early warning of foreign bodies; a recognition algorithm for foreign body target intrusion detection based on syntactic pattern recognition is proposed to achieve the disassembly of complex semantics and analyze the type of intrusion target through basic sub-patterns; a study is conducted based on the sparsification method to obtain parameters with less impact on the algorithm, and a spatiotemporal big data model compression technology is proposed to achieve a significant reduction in the complexity of the algorithm. A railway perimeter target intrusion detection model can also be built to achieve the extraction of key objects of concern such as fallen rocks and flying birds along the railway. The railway perimeter target intrusion detection based on image intelligent recognition and syntactic pattern recognition that is adapted to day and night is described.

[0078] In one embodiment of the present invention, the step of spatial analysis of remote sensing image feature level is included: using theories such as deep learning, image processing, object recognition and classification of multi-source and multi-period remote sensing images along the railway are performed, spatial feature information of objects such as tracks, vehicles and equipment is extracted, and a feature level recognition model of remote sensing image of railway infrastructure along the railway is established to realize automatic detection and recognition and position calibration; a remote sensing image feature enhancement technology of high-resolution panchromatic image and multi-spectral image fusion is proposed to realize super-resolution reconstruction of low-resolution images by multi-source resolution remote sensing images, provide effective information for railway facility identification, and improve identification and classification effects. It is also possible to build a remote sensing image feature recognition model, based on the remote sensing image feature enhancement of high-resolution panchromatic image and multi-spectral image fusion, and establish a feature level recognition model of remote sensing image of railway infrastructure along the railway, so as to realize the feature recognition of safety hazards such as colored steel tiles and plastic greenhouses around the railway.

[0079] The advantages of the railway facility spatiotemporal feature big data analysis method proposed in the present invention include: (1) The categories of railway facility spatial features are designed, including station and line distribution features, railway network structure features and facility equipment spatial layout features, and the railway facility spatial feature big data analysis route is designed. (2) The overall process of railway facility spatial feature big data analysis is designed, including data collection and preprocessing, spatial data storage and management, railway facility spatial feature extraction, feature fusion and integration, spatial data analysis and mining, and result visualization and application. At the same time, the overall architecture of railway facility spatial feature analysis is proposed, including the system data architecture and structural associations such as the original data layer, data processing layer, mechanism model layer, and application layer. (3) A maturity assessment model for railway facility spatial feature analysis based on big data is built to evaluate the maturity of railway facility analysis. (4) A variety of railway facility recognition algorithms are proposed, including recognition methods based on video surveillance, pixel value change rate, transfer learning, and line mileage linear analysis, buffer zone analysis, etc. It can be applied to railway facility recognition and analysis tasks in various application scenarios. (5) Based on the multi-source spatiotemporal characteristics of railways, a railway environmental infrastructure and safety range identification model was established using big data analysis methods. (6) A visualization method for railway spatial characteristics under different application scenarios was established, and customized scenario models were constructed for scenarios such as train operation monitoring, railway facility maintenance, and safety accident analysis.

[0080] The key points of the solution proposed in the present invention are: (1) Based on the proposed railway facility spatial feature analysis method, it is possible to extract multi-dimensional railway facility spatial features, and train a big data model based on the extracted railway facility spatial features, laying the foundation for intelligent railways and digital railways. (2) The overall architecture of railway facility spatial feature analysis proposed by this method can comprehensively design the core elements and correlation relationships of data collection and preprocessing, spatial data storage and management, spatial feature extraction, feature fusion and integration, spatial data analysis and mining, result visualization and application. (3) It is possible to combine technologies such as Geographic Information System (GIS), spatial data mining and spatial analysis, map matching and trajectory analysis, image processing and object recognition, spatiotemporal data modeling and prediction, etc. to construct a railway facility spatial feature analysis model to support practical applications such as railway operation monitoring and safety management.

[0081] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0082] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.

[0083] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with features of other embodiments or replace features of other embodiments.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A railway facility spatial feature big data analysis method, characterized in that: The method comprises the following steps: Acquire railway spatial data collected from multiple sources, pre-process the acquired railway spatial data, use the attribute fuzzy matching model to match the key fields of railway spatial data of different data types, and construct railway spatial data with a unified railway element data structure; The constructed railway spatial data with a unified railway element data structure is stored in a geographic information database and / or a big data storage platform after data quality verification; Perform spatial analysis, geographic information system analysis and track network analysis on the stored railway spatial data of historical periods, and extract the spatial characteristics of railway facilities including spatial geometric characteristics, spatial location characteristics and facility status characteristics; The spatial characteristics of the railway facilities are learned through a cluster analysis model, and facility safety accidents are clustered to identify potential safety risks. The facility status characteristics within a historical period contained in the spatial characteristics of the railway facilities are learned through a supervised learning model to analyze the failure modes of the railway facilities.

2. The method according to claim 1, characterized in that The railway spatial data collection sources include multiple types of drones, on-board equipment, sensors, railway monitoring systems, geographic information systems and Beidou collection devices, and the railway spatial data include multiple types of equipment status data, operation record information, station information and line information; The types of the railway spatial data include shp, gdb and tif formats.

3. The method according to claim 1, characterized in that The step of preprocessing the acquired railway spatial data includes: performing data cleaning and formatting processing on the acquired railway spatial data, and performing unified processing on the railway spatial data by using geographic calibration and coordinate system conversion.

4. The method according to claim 1, characterized in that: The steps for data quality verification of railway spatial data include: Collect statistics on the distribution information of the collection points of railway spatial data, analyze the distance deviation and line continuity of the collection points, and screen out railway spatial data with collection point quality lower than the preset threshold according to the preset rules; Verify the integrity and accuracy of the attribute information content structure of railway spatial data with a unified railway element data structure, and screen out incomplete or inaccurate railway spatial data.

5. The method according to claim 1, characterized in that The method also includes a railway facility spatial feature fusion step, including: The node features and sequence features of railway spatial data are obtained by learning the railway spatial data stored in historical periods through the neighborhood aggregation model and the recurrent neural network model. Using location-aware multi-head self-attention networks to discover hidden features of railway facilities; The node features, the sequence features, the hidden features, the spatial geometry features, the spatial position features and the facility status features are input into a feature fusion model to obtain fused railway facility spatial features.

6. The method according to claim 1, characterized in that The method also includes the step of railway environment safety situation awareness analysis, including: Using railway spatial data collected from multiple sources, a railway environmental infrastructure and safety range identification model is established. Based on fixed-focus video surveillance, the pixel value change rate and grayscale projection of the monitored images are analyzed to achieve railway facility identification. By utilizing railway spatial data collected from multiple sources and adopting a railway perimeter feature recognition model based on transfer learning, features posing safety hazards on the railway perimeter are detected and identified.

7. The method according to claim 1, characterized in that The step of learning the facility status characteristics in the historical period included in the railway facility spatial characteristics through the supervised learning model to analyze the failure mode of the railway facility includes: The supervised learning model is used to analyze and model the historical failure data of railway facilities, predict the failure modes of railway facilities in the future period, use the classification algorithm to train and identify different types of failure modes, and use the regression model to predict the future time period when the failure will occur.

8. The method according to claim 1, characterized in that The railway spatial data also includes passenger flow data of the railway system, and the method further includes: Use clustering models or association rule mining models to analyze passenger flow data in different areas of the railway system and learn passenger flow patterns and associations between different areas; The spatial regression model is used to learn the relationship between passenger flow and other railway spatial data, and the training is used to predict future passenger flow.

9. The method according to claim 1, characterized in that: The method further comprises: When the collection point is fixed-point video surveillance along the railway, based on the spatiotemporal information of the perimeter environment obtained from the fixed-point video surveillance along the railway, a recognition model for foreign object intrusion detection is constructed based on the deep learning model to achieve rapid positioning, recognition and early warning of foreign objects; The method also includes the step of evaluating the model for analyzing the spatial characteristics of railway facilities using a pre-trained maturity assessment model; wherein the maturity assessment model includes three dimensions: analysis application, analysis quality, and data architecture, each dimension is divided into five levels, the analysis application dimension includes five levels: descriptive application, diagnostic application, predictive application, exploratory application, and decision support application, the analysis quality dimension includes five levels: descriptive analysis, diagnostic analysis, predictive analysis, control analysis, and decision support analysis, and the data architecture dimension includes five levels: single structure, comprehensive, artificial organization, partially autonomous organization, and autonomous organization.

10. The method according to any one of claims 1 to 9, characterized in that The method further comprises: Visualize the analysis results based on the spatial characteristics of railway facilities.

Images