摘要
提出了一种基于Gabor滤波器的坯布表面缺陷检测方法。该方法基于Gabor滤波方案,分别将原始图像与滤波器进行卷积操作,然后采用大津法阈值分割来获得坯布表面缺陷。为了优化Gabor滤波器的检测效果,研究不同的参数设定对于检测结果的影响,通过不同类型的无纺布缺陷的实验结果证明了该方法的有效性。最后通过对比常用的检测算法,突出该算法的实用价值。
A defect-detecting method of non-woven based on Gabor filter was proposed.According to this method,the original image was convoluted with the filter,and then be segmented by OTSU threshold to get the default of the non-woven.Also,the influence of different parameters was investigated through the experiments in order to optimize effectiveness of the filter,the result based on various types of defection verified the effectiveness of the proposed methods.At last,by comparing with the commonly used detection algorithm,highlights the value in application of the algorithm.
出处
《武汉理工大学学报》
CAS
CSCD
北大核心
2012年第7期129-133,共5页
Journal of Wuhan University of Technology
基金
国家自然科学基金(51005086)
青年教师基金(2012QN248)
