Railway Asset Extraction from LIDAR Point Clouds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for extracting railway assets, locations, and attributes from LIDAR point cloud data are time-consuming and require significant manual effort, as existing software tools are not fully automated and struggle with accurately identifying rail locations, especially in switch areas and new track sections.
Innovation Solution
A system that automatically recognizes railway assets by processing LIDAR survey point cloud files, using a management platform to generate an engineering-grade survey with accurate location data, including a map of the guideway and associated objects, by applying a Kalman filter and object templates to identify track points and objects, and processing zones to handle switch areas and new sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual extraction methods are used to identify and map railway assets from point cloud data, then flexibility and adaptability are maintained, but time consumption and labor effort increase significantly
Solution Approach 1:
The system performs automated extraction of railway assets, locations, and attributes from point cloud data without requiring manual intervention. The software independently identifies guideway features, extracts geometric properties, and generates mapping information, making the extraction process self-sufficient and eliminating time-consuming manual operations.
Solution Approach 2:
The patent replaces manual mechanical extraction processes with an automated software-based system that processes point cloud data computationally. Instead of human operators manually identifying and mapping railway assets, the system uses algorithms to automatically detect features, calculate positions, and extract attributes from the point cloud data.
2Extent of automation
If existing software tools are used for railway asset identification, then some automation is achieved, but accuracy deteriorates in switch areas and new track sections
Solution Approach 1:
The system applies different processing strategies and algorithms tailored to specific railway features and conditions. For switch areas and new track sections, the software uses specialized detection methods that account for the unique geometric characteristics and challenges of these regions, thereby maintaining high accuracy across diverse railway configurations.
Solution Approach 2:
The patent employs adjustable parameters and thresholds that can be optimized for different extraction scenarios. By dynamically modifying detection parameters based on the specific characteristics of the point cloud data and railway features being analyzed, the system maintains high precision whether processing standard track sections, switch areas, or newly constructed rails.
3Productivity
If fully automated recognition is implemented, then productivity increases, but system complexity increases
Solution Approach 1:
The automated extraction system is divided into distinct functional modules that handle different aspects of the processing task. The software segments the extraction process into steps such as point cloud preprocessing, feature detection, parameter calculation, and result generation, allowing each module to be independently optimized and managed while achieving high overall productivity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method includes detecting an initialization position of a processing zone within a graphical user interface, the processing zone having boundaries and a predefined direction extending away from the initialization position, the graphical user interface comprising displayed point cloud data, the displayed point cloud data being based on a scanning of a three dimensional space. The method also includes applying a Kalman filter to the track points to identify a trajectory of a guideway and generating a model of the guideway based on the track points and the trajectory. The method further includes detecting one or more of a turnout region or an object associated with the guideway. The method additionally includes generating a map comprising the model of the guideway and one or more of the turnout region or the object, and at least one label identifying the turnout region or the object included in the map.