Track-Bound Rail Path Detection with Gap Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining a track-bound rail path in a track system face challenges in accurately identifying the rail path under varying environmental conditions, particularly with turnouts, leading to potential errors in image recognition.
Innovation Solution
The method involves capturing image data using an image sensor, identifying linear sequences of rails, checking for gaps, and classifying representative image sections to ensure accurate rail path determination, utilizing a software-based gap classifier with machine learning capabilities to enhance reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional image-based track detection methods are used, then the system can operate autonomously, but errors in image analysis and misjudgments occur particularly under varying environmental conditions and with turnouts
Solution Approach 1:
The patent segments the track detection task into multiple specialized components: a neural network for initial track section classification and a separate gap classifier for detecting interruptions in rail lines. This segmentation allows each component to specialize in specific aspects of track detection, improving overall reliability while maintaining autonomous operation capability
Solution Approach 2:
The patent introduces an intermediary gap classifier that acts as a mediator between the neural network's initial classification and the final track path determination. This intermediary component specifically addresses gaps in rail lines caused by turnouts and environmental factors, refining the raw neural network output into more reliable track path information
2Measurement precision
If neural networks are used for track classification, then classification capability is improved, but errors cannot be ruled out particularly with turnouts and environmental variations
Solution Approach 1:
The patent implements a feedback mechanism where the gap classifier analyzes the neural network's classification results and identifies gaps in rail lines. This feedback loop allows the system to detect and correct potential errors in neural network classification, particularly for turnouts and environmentally challenging sections, thereby reducing the overall error probability
Solution Approach 2:
The patent performs preliminary gap analysis using the gap classifier before final track path determination. By identifying potential gaps and issues in advance, the system can prepare appropriate corrections or alerts, preventing errors from propagating to the final autonomous driving decisions
3Ease of operation
If the rail path is determined from continuous image data, then the path can be established, but gaps in linear sequences (e.g., at turnouts) lead to misinterpretations
Solution Approach 1:
The patent extracts and separately analyzes gap information from the continuous image data using the dedicated gap classifier. By taking out the gap detection function from the general track classification process, the system can specifically address interruptions in rail lines without compromising the overall ease of track path determination
Solution Approach 2:
The patent applies local quality analysis by using the gap classifier to specifically examine regions where gaps are likely to occur (such as turnouts and switch areas) with higher scrutiny. This localized detailed analysis reduces misinterpretation risk in critical areas while maintaining efficient processing in straightforward track sections
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for determining a track-bound rail path (SPN) in a track system, in which image data of the track system are acquired using an image sensor and a computer, representative image sections are determined from the image data, and the rail path (SPN) is determined based on these representative image sections. The rail path (SPN) is determined by the computer by capturing the course of the rails (SH) of the relevant track (GL) as linear sequences (LZ1 ... LZ3), checking whether the linear sequences have gaps (LK1 ... LK2) in their course, and selecting specific representative image sections (BBA2, BBA3) based on which the gaps (LK1 ... LK2) are classified. The invention further relates to a monitoring device (UV1 ... UV2), a computer program, and a data provisioning device.