3D microstructure reconstruction of nonwoven fabrics based on depth from focus

Abstract

The filtration properties of nonwoven fabrics mainly depend on the fiber structure and alignment, which is difficult to be determined by using traditional methods. It is necessary to develop some new imaging method to characterize the 3D microstructure of nonwovens instead of simple 2D imaging of fabric surface appearance. In this paper, a novel method based on depth from focus is introduced to reconstruct three-dimensional microstructure of nonwoven fabrics. Firstly, a self-developed micro imaging system is established to capture the image sequence of the nonwoven fabric specimen, to be used for further reconstruction of a 3D model. Secondly, a depth from focus algorithm is developed to generate the depth map from image sequences. Thirdly, each fiber segment is located and identified by regional growth and the missing parts caused by occlusion could be restored. Fourthly, central the axis of the fiber is extracted by a thinning algorithm and polynomial curve fitting. Finally, the fiber radius is calculated and 3D model reconstructed using a ball whose sphere center rolls along the central axis. Our experimental results show that the real three-dimensional microstructure of nonwovens can be reconstructed well by using this new depth from focus method, which is very useful for the accurate modeling and analysis of nonwoven fabrics.