Segmentation Method for Enhancing the Continuity and Integrality of Microcracks in Concrete Fracture XCT Image

Abstract

The integrality and connectivity of detected cracks are critical to quantitatively analyzing the fracture patterns of cementitious materials using X-ray computed tomography (XCT). Owing to noise, uneven gray values, pores, and other additions with gray values similar to those of the cracks in the raw images, image segmentation results of microcracks always have poor continuity and integrality. In this paper, a sophisticated image processing method is proposed to detect integral microcracks in XCT images accurately, and a reconnecting algorithm is proposed to enhance the continuity and integrality of microcracks in the concrete fracture XCT image. The analysis of XCT images of fractured rubber concrete samples verified that the method could visualize and quantitatively analyze the fracture pattern, demonstrating the method’s robustness.