The research of pulp fiber cross separation processing

摘要

The purpose of the research is to solve the current problems of inaccurate separation of overlapping objects, operation of the influence on the measurement of single target object and the error is large .Due to the properties of pulp fibers, this paper provides an accurate and automatic method of detecting cross fibers and isolating fibers from the pulp fiber under noises. In this paper, a off-the-shelf scanner is used to generate a grayscale image of multiple fibers followed by customized image processing algorithms that compute the length of each fiber in the image. So an automatic method to separate cross fiber in pulp fiber accurately, analyse the geometrical characteristics of the intersection or cross section is put forward. In this method finding and marking the intersection or cross section, and then using the fiber edge and skeleton near the intersection or cross section of the continuity and differentiability come ture.Finally calculating the average slope will be done. Experiments in VC++6.0 show that the proposed method is efficient in separating the corss fibers accurately.