Crack Boundary Extraction Using Skeleton Graph Pruning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer vision methods for detecting structural defects in images, such as cracks in civil engineering structures, face challenges in speed and accuracy, particularly due to the time-intensive and costly process of obtaining high-quality annotations, and existing weak supervision techniques are insufficient to optimize the labeling of training data.
Innovation Solution
A method involving image preprocessing, skeletonization, and graph pruning is employed to generate a revised skeletal structure of cracks, allowing for rapid and accurate annotation of defect boundaries, which can be refined interactively through user input and integrated into a machine learning pipeline.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional computer vision methods are used for defect detection, then detection capability is achieved, but annotation time and cost increase significantly
Solution Approach 1:
The patent creates a simplified copy of the defect structure through skeletonization, transforming complex defect boundaries into skeletal graphs that retain essential topological information. This skeletal copy serves as a streamlined representation that is much faster to process while preserving the critical structural characteristics needed for accurate defect detection and annotation.
Solution Approach 2:
The patent extracts the essential topological structure from the full defect boundary by generating a skeletal graph that contains only the critical structural elements. By separating the core topological information from the complete boundary details, the system achieves faster processing while maintaining detection accuracy.
2Measurement precision
If detailed boundary annotation is performed manually, then annotation accuracy is improved, but productivity decreases
Solution Approach 1:
The patent performs preliminary skeletonization and graph pruning to create a simplified structural representation before the actual boundary annotation process. This pre-processing step establishes the essential topological framework in advance, enabling faster and more accurate annotation by providing a structured guide that reduces the complexity of the original defect boundary.
Solution Approach 2:
The patent replaces the manual mechanical process of tracing complex defect boundaries with an automated computational process that generates skeletal graphs. This substitution transforms the annotation task from labor-intensive manual tracing to efficient algorithmic processing, dramatically improving productivity while maintaining precision through the structured skeletal representation.
3Loss of information
If complete defect boundaries are processed, then measurement completeness is achieved, but processing complexity increases
Solution Approach 1:
The patent segments the defect boundary processing into distinct stages: initial boundary detection, skeletonization to extract topological structure, graph generation, and selective pruning. This segmentation divides the complex processing task into manageable components, each handling specific aspects of the boundary information, thereby reducing overall processing complexity while maintaining completeness.
Solution Approach 2:
The patent introduces skeletal graphs as an intermediary representation between the original defect boundary and the final annotation output. This intermediary structure serves as a simplified bridge that preserves essential topological information while reducing processing complexity, enabling efficient manipulation and analysis without dealing directly with the full complexity of the original boundary.
Data Source
AI summary
According to embodiments, a method, computer system, and computer program product for obtaining boundaries of structural defects of materials in images of structures is provided. The present invention may include loading an image of a structure made of a material having one or more structural defects and running a pipeline according to certain processing parameters. Running the pipeline pre-processes the loaded image to obtain an initial segmentation mask, where the mask defines an initial boundary of the crack. Based on the initial segmentation mask obtained, a graph of a skeletal structure of the crack is generated, where the skeletal structure comprises a backbone and outer substructures. The graph is pruned by cutting away one or more outer subgraphs corresponding to respective outer substructures to obtain a revised skeletal structure. A revised boundary of the crack is obtained based on both the loaded image and the revised skeletal structure.