Pulp fiber corner detection arithmetic based on local curvature function

The key problem of fiber shape detection is the orientation of its corner; corner is the pivotal point which determines the shape feature of fiber curve. In order to describe the bent character of pulp fiber image, pulp fiber corner detection arithmetic based on local curvature function was proposed. After extracting edge by Canny operator, the local curvature function of contour was calculated, corners are the peaks of the function and detected by threshold. Experimental results show that the arithmetic is precisely, fast and effectively during the corner detection of pulp fiber. The arithmetic is very stable and robust.