A novel image analysis method for the fiber orientation measurement of nonwovens based on multi-focus image fusion technology

A novel multi-wavelet image fusion algorithm is proposed and the fiber orientation of the nonwoven fabric is determined based on the fused image. Different fusion methods suitable for high and low frequencies after image decomposition are developed to solve the problem where fiber information is unable to not be clearly identified using only single-focus image. Based on the fused images, one set of image analysis algorithm is developed for feature extraction, including the binarization of a double threshold segmentation of a genetic algorithm, median filter, and corrosion. The fiber orientation and pore can be measured using Hough transform and geometrical analysis. Experiments show that fiber orientation can be clearly displayed and can be used as the foundation for research.