Hardness Tester Image Processing for Accurate Vertex Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hardness testing methods face reliability issues due to dependence on noise reduction accuracy and misrecognition of impression shapes, especially when dealing with multiple samples or scratches parallel to the impression region, leading to inaccurate results.
Innovation Solution
A hardness tester and method that extracts impression regions and vertices using binarized image data with reduction and expansion processing, followed by distance conversion, allowing for accurate calculation of hardness without relying on noise reduction accuracy and enabling the detection of multiple samples.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional techniques use noise reduction processing on binary images to determine impression boundaries, then the processing can be simplified, but the reliability of vertex position determination deteriorates due to dependence on noise reduction accuracy
Solution Approach 1:
The patent applies preliminary action by performing reduction processing and expansion processing on the binary image before boundary detection. This pre-processing removes noise and fills gaps in the impression boundaries, creating a cleaner image that leads to more reliable vertex position determination without requiring complex noise reduction algorithms later in the process.
2Adaptability or versatility
If conventional techniques estimate impression regions using approximated multidimensional curves, then the method can handle complex shapes, but the measurement precision of vertex positions deteriorates
Solution Approach 1:
The patent applies inversion by reversing the conventional approach: instead of estimating curves and finding their intersections, it directly detects boundary points from the pre-processed binary image and uses these actual boundary points to determine vertex positions. This inversion from estimation to direct detection significantly improves measurement precision while maintaining the ability to handle complex impression shapes.
3Device complexity
If conventional techniques assume only one impression in the captured image, then the processing can be simplified, but the adaptability to test multiple samples deteriorates
Solution Approach 1:
The patent applies universality by designing an image processing system that can handle both single and multiple impressions using the same core algorithms. The boundary detection and vertex determination methods work universally for any number of impressions in the image, allowing the system to adapt to different test scenarios (single sample or multiple samples) without requiring different processing approaches.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A hardness tester (100) includes an image capture control means (61) for obtaining an image data of the surface of the sample by controlling and making an image capture means (12) capture an image of the surface; an impression region extracting means (61) that binarizes the image data, applies reduction/expansion processing to the binarized image data, applies a distance conversion processing to the reduction/expansion processed image data, and extracts a closed region corresponding to a contour of the indenter by using the distance-converted image data; an impression vertex extracting means (61) that estimates the vertex for measuring the impression based on a profile of the closed region and extracts a point in the binarized image data, as the vertex for measuring the impression, that agrees with a predetermined condition; and a hardness calculating means (61) for calculating hardness of the sample based on the vertexes for measuring the impression.