基于羽毛图像纹理分割的毛杆提取方法  

An Approach to Image Segmentation Based on Texture

作　　者：刘洪江[1] 汪仁煌[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510090

出　　处：《广东工业大学学报》2010年第4期42-45,57,共5页Journal of Guangdong University of Technology

摘　　要：为了有效地从羽毛图像中提取毛杆,在对比各种分割方法的基础上,提出基于小波纹理分割方法.考虑羽毛的特点,为去噪和加深纹理特征,对图像进行恢复和增强.经过二层小波分解后提取各细节子图的特征,形成特征向量,并由聚类法进行分类分割.实验结果表明,该方法能较好地提取毛杆.To extract effectively the stick in feathers,it proposes a new method of image segmentation,based on the wavelet decomposition,by analyzing the textural features of image and comparing some methods of texture segmentation.The characteristics of feathers were considered,and the image was restored and enhanced to deepen the texture features.The features of vectors were obtained by the wavelet coefficient after wavelet transformation.Then,clustering was used to segment the texture image directly.The experimental results verify the effectiveness of the method.

关 键 词：纹理分割 图像恢复 频域增强 二维小波分解 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]