Train Obstacle Detection via Ground-Based Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional obstacle detection systems for trains face challenges in managing storage capacity due to the need to store captured images, leading to increased storage requirements.
Innovation Solution
A train control system that offloads image processing to a ground-based obstacle detection device, which compares image information with stored data, reducing the need for on-train storage and enabling efficient obstacle detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If captured images are stored in a storage device installed in the train for obstacle detection, then obstacle detection capability is maintained, but the capacity of the storage device increases
Solution Approach 1:
The patent extracts the image storage function from the train-mounted system and relocates it to a ground-based server. The train only retains the imaging device and basic processing capabilities, while captured images are transmitted and stored externally, thereby reducing on-train storage requirements while maintaining obstacle detection functionality.
Solution Approach 2:
The patent introduces a ground-based server as an intermediary between the train's imaging device and the obstacle detection process. This server acts as a mediator that receives, stores, and processes images, allowing the train system to maintain detection capabilities without requiring large onboard storage capacity.
2Reliability
If image processing is performed on the train, then obstacle detection is enabled, but the complexity of the train system increases
Solution Approach 1:
The patent extracts the complex image processing and analysis functions from the train system and relocates them to a ground-based server. The train retains only the imaging device and basic transmission capabilities, significantly reducing system complexity while maintaining obstacle detection functionality through centralized processing.
Solution Approach 2:
Instead of performing all image processing on the train (traditional approach), the patent inverts the architecture by having the train only capture and transmit images, while all processing occurs on the ground. This inversion simplifies the train system while achieving the same detection goals.
3Measurement precision
If all captured images are stored for comparison, then detection accuracy is improved, but storage requirements increase
Solution Approach 1:
The patent applies partial action by selectively storing and comparing only relevant images rather than all captured images. The system compares current images with previously stored images to detect obstacles, maintaining detection accuracy while avoiding the need to store every single captured image indefinitely.
Solution Approach 2:
The system discards images after they have served their comparison purpose, rather than permanently storing all images. Once an image has been used for obstacle detection comparison, it can be discarded, freeing up storage space while maintaining the ability to detect obstacles through comparison with retained historical images.
Data Source
AI summary
A train control system includes: an image information conversion device that is mounted on a train and outputs first image information obtained by adding train information to image information imaged by an imaging device capable of capturing an area in front of the train; and an obstacle detection device that is installed on the ground and includes: an image information storage unit that records the first image information output from the image information conversion device; an image information comparison unit that compares the first image information with second image information recorded in the image information storage unit; and an image information result output unit that outputs a result from the image information comparison unit.