Identification of Animal Fiber Based on Scale Shape

Scale structure pattern of animal fiber is different and that is a major reference distinguishing them from each other. There are four main scale parameters, including fiber diameters, scale interval, scale perimeter and scale area, can be used to describe their basic shape. In this paper, cashmere and fine wool fiber sample are check up under light microscope with a magnification of 40× for objective and their images are captured by CCD camera. After a series of operations are performed, a simple skeletonized binary representation only having one pixel wide and showing only fiber and scale edge details can be obtained. Four basic shape parameters described above are measured from these images and a database composed of numerical data of four relative indexes is established. A LVQ neural network classification model, including four input nodes, sixteen hidden nodes and two output nodes, are developed on them. The simulation results show that whether on training set or testing set, the model can always distinguish cashmere from fine wool (70s) effectively and the average classification accuracy are higher than 91 percent.