Steel Carbide Particle Detection Using Watershed Image Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating carbide morphology in steel materials using scanning electron microscopy (SEM) images struggle to accurately identify and count individual carbide particles due to binarization techniques that merge multiple particles into a single structure, making it difficult to assess their size and number accurately.
Innovation Solution
An apparatus and method utilizing image processing techniques, including binarization and watershed separation, to extract and separate carbide particles based on length thresholds and characteristic value distributions, allowing for precise identification of individual carbides.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If binarization is performed on SEM images to extract carbide particles, then the extraction process is simplified, but multiple particles are merged into a single structure making accurate counting impossible
Solution Approach 1:
The patent applies segmentation by dividing the binarized image into multiple regions based on carbide particle characteristics. The separation unit segments connected carbide particles into individual entities by detecting branching points and calculating path lengths, thereby resolving the merging issue while maintaining the simplified extraction process
Solution Approach 2:
The patent introduces a new dimension of analysis by calculating path lengths and branching point information in addition to basic binarization. This dimensional enrichment allows the system to distinguish between merged particles based on their geometric paths, enabling accurate counting without complicating the extraction process
2Ease of manufacture
If conventional binarization is used to extract carbide shapes, then the extraction process is simple, but the structure cannot properly identify individual particles with linear characteristics
Solution Approach 1:
The separation unit segments carbide structures by detecting branching points and calculating path lengths from these points. This segmentation approach preserves the linear characteristics of carbide particles while enabling accurate identification of individual particles, resolving the contradiction between simple extraction and precise identification
Solution Approach 2:
The patent applies local quality by analyzing different regions of the carbide structure with different methods. Branching points are detected in specific regions, and path lengths are calculated along different trajectories, allowing the system to adapt its analysis to the local characteristics of each carbide particle
Data Source
AI summary
There is provided an apparatus for detecting a carbide particle capable of properly identifying carbide particles in a steel material, a method for detecting the carbide particle capable of properly identifying the carbide particles in the steel material, or a program for detecting a carbide particle capable of properly identifying the carbide particles in the steel material. An apparatus for detecting a carbide particle according to an embodiment of the present invention includes an extraction unit for binarizing image data of a microscope image of a steel material to extract shapes of carbides, and a separation unit for separating particles of the carbides from the binarized image data based on watersheds or the shapes of the carbides.