A Method for Detection of Foreign Body in Cotton Based on Threshold Segment

摘要

In order to extract foreign body from the complex channel background and cotton layers, a detection method which combined improved Otsu threshold with background estimation threshold is presented. Firstly, the original image which containing multiple foreign fibers is divided into two new images which containing only two substances by Otsu threshold. And then using the estimated value of the means and the standard deviations of the two new images a background estimation threshold was determined. The foreign fibers are extracted by the estimation threshold. Simulation results show that this method can overcome the effect which caused by the channel background interference and diversity of the foreign fibers in the actual working environment and can extract foreign bodies quickly and effectively.