摘要
基于灰度共生矩阵(GLCM)开展辽东湾ENVISAT ASAR海冰影像的纹理特征分析研究。在纹理分析过程中,研究GLCM的方向、位移量和灰度量化级等参数对SAR海冰影像纹理特征的影响,得到适合辽东湾海冰检测的灰度共生矩阵参数的值及纹理统计特征量。利用BP神经网络分类法,建立辽东湾SAR影像的海冰检测方法,并结合MODIS影像和TM影像对该海冰检测方法的结果进行比较,验证其有效性。
Texture analysis and study of ENVISAT ASAR sea ice imagery for the Liaodong Bay are carried out using gray level co-occurrence matrices （GLCM）. In the analysis process, the effects given to the sea ice textural features by the parameters such as the GLCM orientation, the displacement, and the quantified gray level are studied in order to determine the values of the GLCM parameters and to represent the textural features, which are suitable for the sea ice detection of the Liaodong Bay. Then the BP neural network classification method is used to derive an approach for sea ice detection from SAR images of the Liaodong Bay. Finally the results of this texture analysis are compared with those of the MODIS imagery and the TM imagery to verify the reliability of this method.
出处
《海洋科学进展》
CAS
CSCD
北大核心
2008年第3期386-393,共8页
Advances in Marine Science
基金
国家海洋局第一海洋研究所基本科研业务费专项资金项目--基于遥感的辽东湾海冰冰情预测方法(2007B06)
国家海洋局青年基金--SAR辽东湾海冰类型识别研究(2008402)
