基于小波变换及Gabor滤波的起毛起球图像分割  

Fabric pilling image segmentation based on wavelet transform and Gabor filter

作　　者：汪亚明[1] 崔新辉[1] 韩永华[1] 

机构地区：[1]浙江理工大学信息学院,杭州310018

出　　处：《丝绸》2016年第3期37-40,共4页Journal of Silk

基　　金：国家自然科学基金项目(61272311);浙江省自然科学基金重点项目(LZ15F020004);机械工程浙江省高校重中之重学科和浙江理工大学重点实验室优秀青年人才培养基金项目(ZSTUME01B17);计算机应用技术重点学科研究生创新研究项目(XDY14003);浙江理工大学科研启动基金项目(13032156-Y)

摘　　要：针对目前织物起毛起球等级评定方法不能有效实现毛球、纹理有效分离,或虽能分离但对光照不匀敏感的问题,提出小波变换和Gabor滤波相结合的方法进行图像分割。根据织物起毛起球图像噪声特点,利用小波变换的多分辨率特性去除光照不匀、织物不平产生的低频噪声和织物绒毛等引起的高频噪声;然后根据织物纹理的周期性和方向性,通过Gabor滤波去除织物纹理噪声,并使用Otsu算法对去噪后图像进行二值分割。实验结果表明,该方法能够有效实现织物起毛起球图像的去噪处理,得到准确的织物毛球分割图像。Current pilling grading methods can not realize effective separation of pilling and texture,even though separation can be achieved,it is sensitive to uneven illumination. Thus,this paper proposes the combination of wavelet transform and Gabor filter for image segmentation. According to the characteristics of fabric pilling image noise,this paper utilizes multiresolution feature of wavelet transform to eliminate low-frequency noise caused by uneven illumination and uneven fabric and high-frequency noise caused by fabric fluff. Then,according to periodicity and direction of fabric texture,Gabor filter is applied to eliminate fabric texture noise. Meanwhile,Otsu algorithm is used for denoised binary image segmentation. The results show that this method can effectively achieve de-noising treatment of fabric pilling,and the accurate pilling segmentation image is gained.

关 键 词：起毛起球 图像分割 小波变换 GABOR滤波 阈值分割 织物 

分 类 号：TS101.91[轻工技术与工程—纺织工程]