Pulp Fibre Recognition Based on Curvelet Transform and Skeleton Tracing Algorithm

In the analysis of pulp fibre, the characteristics of pulp fibre are important parameters in papermaking industry. In this paper, the curvelet transform and skeleton tracing method based on preferential direction were introduced to analyze the paper pulp fibre images. The curvelet transform was developed based on the Fourier transform, wavelet transform and radon transform. It has overcome some limitations of wavelet in representing orientations of edges in images. The theory and implementation of curvelet transform were summarized. In this paper, the curvelet transform was used to analyze the pulp fibre images, and the different characteristics of each fibre were extracted to disjoin the intersect fibres; the crossed or furcated fibres were distinguished by the width information. Experiment results showed that the proposed algorithm obtained perfect effects in the analysis of pulp fibre images.