Polynomial Regression, Area and Length Based Filtering to Remove Misclassified Pixels Acquired in the Crack Segmentation Process of 2D X-Ray CT Images of Tested Plaster Specimens

This work presents an effective and robust technique to remove misclassified pixels acquired in the crack segmentation process of 2D X-ray CT images of tested plaster specimens. Cracks have distinct properties, such as they are fairly piece-wise linear, and they have certain area and length ratios, which can be used to remove misclassified pixels from cracks segments. In this paper, a combination of polynomial regression and area-based, length-based filtering scheme is applied to remove undesired pixels from the 2D CT images of plaster specimen. With the help of experimental results the effectiveness and robustness of the proposed technique are verified.