Rail Position Detection via Edge Image Node Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting rail position information in vehicles, such as those using GNSS and map databases, face challenges in accuracy due to rail shapes, weather conditions, and GNSS errors, leading to unstable performance and the need for position correction in critical areas.
Innovation Solution
A position information detection device that includes a receiver for correlating node coordinates with rail data, a generator for creating edge images, a rail specifying unit for identifying overlapping node coordinates, and a vehicle position acquisition unit to determine accurate position information, integrating GNSS data with visual and map-based information for error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image features are extracted from camera images to detect rails, then rail position detection is performed, but detection becomes unstable due to rail shapes, weather, and time conditions
Solution Approach 1:
The patent combines multiple detection methods (image-based rail feature extraction, GNSS positioning, and map database matching) into a unified position information detection device. The control unit integrates results from these different methods to determine final vehicle position, thereby compensating for the weaknesses of individual methods and achieving both high accuracy and stability across various conditions.
Solution Approach 2:
The patent introduces map database information as an intermediary to bridge the gap between GNSS coordinates and actual rail positions. By matching detected rail positions with pre-stored map data, the system can correct GNSS errors and provide accurate position information even when direct image-based detection fails due to weather or rail shape variations.
2Reliability
If map database is interlocked with GNSS position information for rail detection, then stable performance is obtained, but position accuracy is insufficient in critical areas requiring correction
Solution Approach 1:
The system uses map database information as feedback to correct GNSS position errors. By comparing detected rail positions with expected positions from the map database, the control unit can identify and correct deviations, thereby improving position accuracy while maintaining the stability provided by the map-based approach.
Solution Approach 2:
The patent pre-stores detailed rail position information in the map database before operation. This preliminary preparation of reference data enables the system to quickly correct GNSS errors during operation without requiring real-time complex calculations, thus maintaining both stability and high accuracy.
3Measurement precision
If multiple detection methods are integrated for accurate positioning, then position accuracy improves, but device complexity increases
Solution Approach 1:
The control unit is designed as a multi-functional component that can process and integrate results from different detection methods (image processing, GNSS data processing, map database matching). This universal processor approach avoids the need for separate dedicated processing units for each method, thereby reducing overall system complexity while maintaining the ability to achieve high position accuracy through integration.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to one embodiment, a position information detection device includes a receiver configured to receive a plurality of data in which a plurality of node coordinates arranged at regular intervals in such a manner as to correspond to left and right rails are correlated with position information; a generator configured to generate an edge image by using an input image from an imaging device disposed in a vehicle; a rail specifying unit configured to specify the data including second rail node string information that is the node coordinates that are most overlapping with a position of an edge included in the edge image; and a vehicle position acquisition unit configured to acquire second position information based on the specified data.