摘要
以渤海ENVISATASAR图像为例,采用基于灰度共生矩阵的纹理特征分析方法,计算油膜和2种类油膜(近海自然表面膜和低风速区)的纹理特征量.结果表明:对比度、方差、熵、相关性、相异性5种纹理特征量可用于识别油膜和类油膜,但对于不同类型的类油膜,用于区分它们和油膜的纹理特征量是不同的.因此,类油膜的分类对区分油膜与类油膜是十分必要的.
Taking the ENVIST ASAR image over the Bohai Sea as an example,the texture characteristics of oil films and two kinds of look-alikes（coastal natural surface film and low wind velocity） were studied using the texture feature analysis method based on the grey level co-occurrence matrix.The results show that the contrast,variance,entropy,correlation,and dissimilarity can be used to distinguish oil films from look-alikes.For different look-alikes,the texture features that are used to distinguish them from oil films are different.Therefore,the classification of look-alikes is very necessary.
出处
《河海大学学报（自然科学版）》
CAS
CSCD
北大核心
2011年第5期569-574,共6页
Journal of Hohai University(Natural Sciences)
基金
国家自然科学基金(41006108)
卫星海洋环境动力学国家重点实验室开放基金(SOED1104)
关键词
图像识别
合成孔径雷达
溢油
纹理特征
灰度共生矩阵
pattern recognition
synthetic aperture radar（SAR）
oil spill
texture characteristics
grey level co-occurrence matrix
