Train Type Identification Using Linear Camera Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing train type identification methods, such as wheelbase measurement, are inadequate for distinguishing between different train types, especially when standards vary by country and when trains are refitted, leading to inaccuracies in identifying the type and carrying objects.
Innovation Solution
A method and system using a linear camera to continuously photograph a train, generate sub-images, splice them into a train image, extract characteristic parameters like wheelbase, height, and color, and compare these with prestored templates to automatically determine the train type, with adjustments for relative speed and distortion correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If wheelbase measurement method is used to identify train types, then identification can be performed based on standardized dimensions, but it fails to accurately identify trains when standards vary by country or when carriages are refitted
Solution Approach 1:
The train is divided into multiple characteristic parameters including wheelbase, carriage height, window quantity, and color. Instead of relying solely on wheelbase measurement, the system segments the identification process into multiple independent features that are extracted and analyzed separately, then combined for comprehensive train type classification
Solution Approach 2:
The system transitions from single-dimensional wheelbase measurement to multi-dimensional identification by incorporating vertical dimension (carriage height), visual dimension (window quantity and color), and combining these with the horizontal wheelbase measurement. This multi-dimensional approach enables accurate identification across different national standards and refitted trains
2Measurement precision
If multiple train inspection points with magnetic steel are arranged along the railway to detect wheelbase, then wheelbase measurement can be achieved, but the system cannot identify trains with non-standard wheelbases or refitted carriages
Solution Approach 1:
The inspection system is enhanced with multi-functional capabilities by adding imaging devices that can capture multiple types of train characteristics (wheelbase, height, windows, color) simultaneously. This universal inspection approach allows the system to reliably identify various train types including those with non-standard wheelbases or refitted carriages
Solution Approach 2:
Image processing technology serves as an intermediary between the physical train and the identification system. The imaging devices capture visual information, and image processing algorithms extract characteristic parameters, acting as a mediator that enables accurate identification without being constrained by magnetic steel detection limitations
3Loss of information
If a linear camera continuously photographs the train to generate multiple sub-images, then comprehensive train characteristics can be captured, but image splicing and processing complexity increases
Solution Approach 1:
The system performs preliminary actions by using the linear camera to continuously photograph the train and generate multiple sub-images during train passage. These sub-images are captured in advance with proper timing and positioning, containing all necessary characteristic information before processing begins
Solution Approach 2:
The system creates multiple copies of train images at different positions and angles through continuous photography. These image copies are then spliced and processed to extract characteristic parameters. The copying approach ensures comprehensive data capture while using standard image processing techniques to manage complexity
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
The present disclosure relates to a train type identification method and system, and a security inspection system and system. The train type identification method includes: continuously photographing a to-be-inspected train by using a linearity camera in motion relative to the to-be-inspected train, and generating (S202) a plurality of train sub-images; splicing (S204) the plurality of train sub-images to acquire a train image of the to-be-inspected train; extracting (S206) at least one train characteristic parameter from the train image; comparing (S208) the at least one train characteristic parameter with a prestored train type template; and automatically determining (S210) a type of the to-be-inspected train based on a comparison result.