关键词: 异性纤维检测, 颜色空间模型, 连通区域分析, 电磁阀, 剔除时间补偿

Abstract: Aiming at the problem of foreign fibers in cotton spinning enterprise, a fast and efficient detecting and locating method for foreign fibers in cotton is proposed. In order to achieve fast segmentation of the fibers from cotton, a detecting model of foreign fibers and the model parameters offline learning method are established in RGB color space. Then morphological processing and connected area analysis algorithm are implemented for accurate location of the different foreign fibers in cotton. Furthermore, eliminating time compensation thought is put forward according to the test of electromagnetic valve response characteristics, which can guarantee the precision of elimination of the foreign fibers. The experimental results prove that this method can achieve to detect and eliminate most of the fibers in cotton.

Key words: foreign fibers detecting, color space model, connected area analysis, electromagnetic valve, eliminating time compensation