摘要
针对海洋SAR图像的特点,采用基于灰度共生矩阵的纹理分析方法,提出适用于海洋溢油SAR图像分类的纹理特征量。并讨论了纹理特征量的筛选和纹理窗口大小的确定等问题。最后采用人工神经网络方法验证了SAR图象分类效果。
According to the characteristics of ocean synthetic aperture radar（SAR） images, a texture analysis method based on grey level coccurrence matrix is used, and the texture characteristic quantities suitable for the classification of ocean oil spill SAR images are suggested. The problems with the screening of texture characteristic quantities and the determination of texture window size are discussed, and the artificial neural network method is used to verify the classification results of SAR images.
出处
《海洋科学进展》
CAS
CSCD
北大核心
2007年第3期346-354,共9页
Advances in Marine Science
基金
国家海洋局青年海洋科学基金--溢油SAR遥感信息系统关键技术研究(2006401)
关键词
SAR
纹理分析
灰度共生矩阵
人工神经网络
synthetic aperture radar（SAR）
grey level co-occurrence matrix
artificial neural network
