Fatigue Crack Detection via Feature Tracking Video Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting fatigue cracks in steel bridges are time-consuming, labor-intensive, and prone to errors, and existing non-destructive testing techniques require complex and costly equipment, making them inefficient for early detection and monitoring.
Innovation Solution
A computer-vision-based method that captures video of a structure under dynamic loading, detects feature points, tracks relative movement, and qualifies crack characteristics using a processing device to identify and quantify fatigue cracks without relying on edge features or expensive equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human inspection is used to visually examine fatigue cracks, then detection capability is provided, but time consumption and labor intensity increase significantly
Solution Approach 1:
The patent replaces manual human inspection with an automated computer vision system that captures images, detects feature points, and analyzes crack patterns automatically. This substitution eliminates the time-consuming and labor-intensive nature of human inspection while maintaining or improving detection reliability through consistent algorithmic analysis.
Solution Approach 2:
The system enables self-service inspection by automatically performing crack detection without human intervention. The computer vision system independently captures images, processes them through feature point detection, and generates crack analysis results, making the inspection process autonomous and significantly reducing both time and labor requirements.
2Measurement precision
If non-destructive testing techniques using acoustic emissions and piezoelectric sensors are used, then inspection accuracy improves, but system complexity and cost increase
Solution Approach 1:
The patent extracts the essential function of crack detection from complex NDT systems and implements it using a simplified computer vision approach. By removing unnecessary sensors, signal generators, and complex processing equipment, the system achieves comparable or superior accuracy with significantly reduced complexity and cost.
Solution Approach 2:
The patent replaces expensive, complex NDT equipment with affordable, readily available digital cameras and standard image processing software. This substitution maintains inspection accuracy while dramatically reducing system cost and complexity, making the solution accessible for widespread use.
3Reliability
If strain-based monitoring technologies with sensors and cabling are installed, then fatigue crack detection capability is provided, but installation complexity and cost increase
Solution Approach 1:
The patent replaces physical strain sensors and cabling with a non-contact optical measurement system. Digital images capture surface deformations and crack patterns directly, eliminating the need for sensor installation, wiring, and associated complexity while maintaining reliable crack detection capability.
Solution Approach 2:
The computer vision system serves multiple functions simultaneously: it captures images for documentation, detects feature points for analysis, identifies crack locations, and measures crack characteristics. This multi-functionality replaces multiple specialized devices with a single unified system, reducing overall complexity.
4Difficulty of detecting and measuring
If traditional edge feature-based crack detection is used, then crack location can be identified, but detection reliability decreases under low light or complex surface conditions
Solution Approach 1:
The patent changes the detection parameter from relying on visual edge features to tracking the motion and displacement of feature points. By analyzing how feature points move relative to each other under applied loads, the system can detect cracks reliably regardless of lighting conditions or surface textures, as the mechanical response to loading remains consistent.
Solution Approach 2:
The patent performs preliminary detection of feature points and establishes their initial positions and relationships before applying loads. This preliminary action creates a reference framework that enables reliable crack detection through subsequent motion analysis, independent of environmental conditions during the actual inspection.
Data Source
AI summary
A computer-vision-based fatigue crack detection approach using a short video is described. Feature tracking is applied to the video for tracking the surface motion of the monitored structure under repetitive load. Then, a crack detection and localization algorithm is established to search for differential features at different frames in the video. The effectiveness of the proposed approach is validated through testing two experimental specimens with in-plane and out-of-plane fatigue cracks. Results indicate that the proposed approach can robustly identify fatigue cracks, even when the cracks are under ambient lighting conditions, surrounded by other crack-like edges, covered by complex surface textures, or invisible to human eyes due to crack closure. The approach enables accurate quantification of crack openings under fatigue loading with good accuracy.