摘要
纺织品起球客观评级是纺织品性能测试数字化的重要组成部分.毛球分割是纺织品起球客观评级中的关键步骤,分割结果直接影响评级的精度.目前方法大多利用图像的灰度信息,易受图像亮度和对比度变化以及光照不均匀等因素影响.针对这些问题,研究了基于相位一致性检测的分割算法.该算法基于LogGa-bor滤波器计算相位一致性,对检测结果进行加权阈值分割.实验结果表明,该方法能获得令人满意的分割结果,提高了织物起球客观评级系统的实用性.
Pilling segmentation is a key step in fabric pilling objective evaluation, which is the important component of textile performance test digitization. The segmentation results affect the evaluation precision directly. Existed image processing based on pilling segmentation methods often use intensity information, which can be affected easily by brightness/contrast and un-uniform illumination. To deal with such problems, pilling segmentation based on phase congruency （PC） algorithm is proposed, which adopts Log Gabor filter to calculate PC and uses weighing threshold method to segment the PC image. Experiment results show that the segmentation effect is satisfactory. It can improve the practicability of fabric pilling objective evaluation system.
出处
《天津工业大学学报》
CAS
北大核心
2009年第5期60-63,共4页
Journal of Tiangong University
基金
国家自然科学基金项目(60602036)
作者简介
崔琴（1986-），女，硕士研究生；
肖志涛（1971-），男，教授，硕士生导师．E-mail：xiaozhitao@tjpu．edu.cn
