Train Door Trapping Detection via Image Difference Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle door trapping detection systems using physical sensors are limited in detecting thin objects, and human error can occur when crew members incorrectly assess situations, leading to potential train accidents.
Innovation Solution
A monitoring system utilizing a camera to image train doors and a control device that determines trapping by calculating the difference between reference and observation images, focusing on the movement of central coordinates to accurately detect trapped objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If physical sensors are used for trapping detection, then the detection system is simple and reliable, but thin objects such as stroller wheels cannot be detected
Solution Approach 1:
The patent replaces physical sensors (mechanical/pressure-based detection) with a camera-based image processing system. The monitoring camera captures images of the door area, and image processing determines whether an object is trapped by analyzing pixel differences between images taken when the door is open and closed. This substitution enables detection of thin objects that physical sensors cannot detect.
Solution Approach 2:
The system creates visual copies (images) of the door area using a monitoring camera. By capturing and comparing image data rather than using direct physical sensing, the system can identify trapped objects through visual analysis of pixel differences, enabling detection of objects that would be invisible to traditional sensors.
2Ease of operation
If crew members manually assess trapping situations, then the system is simple to operate, but human error can occur leading to incorrect judgments
Solution Approach 1:
The system performs automatic trapping detection through image processing without requiring crew member intervention for the detection itself. The monitoring camera and control device automatically capture images, calculate pixel differences, and determine whether trapping has occurred, eliminating human error in the detection and judgment process while maintaining ease of operation.
Solution Approach 2:
The system provides automatic feedback to the crew member through the monitoring device, displaying whether trapping is detected based on image analysis. This feedback mechanism replaces subjective human judgment with objective system-generated information, improving reliability while keeping the interface simple for crew members.
3Measurement precision
If image processing determination is implemented for trapping detection, then detection accuracy for all object types improves, but the system complexity increases
Solution Approach 1:
The monitoring camera serves multiple functions: it monitors the door area for trapping detection, provides visual records for review, and can potentially detect other anomalies. The same image processing system that detects trapping can also analyze other aspects of door operation, reducing the need for separate specialized devices and minimizing overall system complexity.
Solution Approach 2:
The patent combines the monitoring camera and image processing determination into an integrated system. Rather than adding separate complex detection devices, the system merges trapping detection functionality with the existing monitoring infrastructure, using the control device's processing capabilities to analyze images and determine trapping status, thereby limiting the increase in overall system complexity.
Data Source
Figure 1(a)~1(b)
Figure 2
Figure 3(a)~3(b)
AI summary
Provided is a technology for monitoring train doors which improves the accuracy of detection of trapping in vehicle doors. A server compares the difference between a static image (reference image 91), from each monitoring camera, of a normal state in which there is no trapping in vehicle doors, said static images being held in the server in advance, and a static image (an observation image 92) first acquired in a prescribed acquisition time. If a difference is present, the difference (difference image 94) between the reference image 91 and the observation image 92 acquired in the acquisition time is acquired. A quadrilateral F which covers the four sides of an object constituting the difference is subsequently rendered, the centre point of the quadrilateral is obtained, it is determined that trapping has occurred if movement of a centre coordinate C1, C2, ... Cn, i.e. the difference between the centre coordinates C, is lower than a prescribed threshold value, and red frame rendering of video is indicated on a monitor.