The Study of Sub-Pixel Edge Detection Algorithm Based on the Function Curve Fitting

Shang, Ya-ceng; Chen, Jing; Tian, Junwei

doi:10.1109/iciecs.2010.5678279

Cited by 6 publications

(1 citation statement)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For this, sub‐pixel edge localisation algorithms can be used, categorised in the literature into three categories. The first category is the fitting methods, which uses a continuous function to fit the gradient of the greyscale value of the image, such as B‐spline [19], Gaussian function [20], Hyperbolic Tangent [21], and the Erf function [22]. The second category is the interpolation methods, which interpolate the greyscale values to a finer grid of pixels, such as Nearest Neighbour, Bilinear [23, 24], and Bicubic interpolations [25].…”

Section: Image Processing Algorithm For Grain Correlationmentioning

confidence: 99%

Image correlation technique for strain measurement of polycrystalline microstructures

Hafiz

Stachurski

Kalyanasundaram

2022

Strain

View full text Add to dashboard Cite

An image processing technique is proposed to measure the deformation of polycrystalline materials based on correlating the grains in reference and deformed SEM images. The advantage of this technique compared to the conventional subset‐based Digital Image Correlation (DIC) is that it can be applied when speckle patterning is not efficient or when studying boundary‐related mechanics is the objective. The technique is based on correlating grains by defining their boundaries rather than just subsets of image pixels. It reveals the anisotropy inherent in the polycrystals since it allows the analysis to specify each grain separately without averaging the results. The technique is applied by detecting the approximate grain boundaries edges and then refining their location with high accuracy. The correlation is performed between points calculated from each grain in the reference and deformed images as a Point Set Registration (PSR) problem. Finally, the displacements and strains are calculated from the resulting transformation matrix. A benchmark problem was developed to discuss the error over a strain range of 0.02 to 0.2 and showed that the resulting strains are reasonably accurate. Also, an in situ experiment was conducted to demonstrate the implementation of the technique using a specimen with fine‐grained Zirconia polycrystals. The technique successfully revealed the crack tip plastic zone, and strain mismatch between grains.

show abstract