摘要: |
基于遥感图像涡区域检测边缘一般是由近似椭圆的分段圆弧曲线所构成的特点,提出一种“由粗到细”的海洋遥感图像中尺度涡计算机自动检测方法。首先对检测边缘利用曲线拟合和局部区域Hough变换的方法产生候选子区域,然后仅仅针对候选子区域利用局部区域图像灰度分割确定出涡的区域。实验结果表明所提出的方法可以完成处于成熟期、形态特征比较明显的中尺度涡自动检测。 |
关键词: 海洋遥感图像,中尺度涡,自动检测 |
DOI: |
分类号: |
基金项目:国家自然科学基金资助项目, 69972019号 |
|
AN AUTOMATIC DETECTING METHOD OF THE MARINE MESO-SCALE EDDY IN REMOTE SENSING IMAGE |
JI Guang-Rong, CHEN Xia, HUO Yu-Zhen, JIA Tong-Jun
|
Department of Electrical Engineering, Ocean University of Qingdao
|
Abstract: |
Because of the complexity of the marine remote sensing image,it is very difficult to automatic detect the meso-scale eddy from the image directly. Based onthe searching strategy of “from course to fine” in computer vision and that the detected edges of the eddy were consisted of sectional curves, the ellipse curve fitting was employed to roughly detect the eddy’s candidate sub-regions. Then the Hough transform was applied to refine the candidate sub-regions only in the local area of the candidate sub-regions. The image segmentation was applied to separate eddy from the background and depict it with an ellipse approximately in the refined sub-regions. Results show that this method is suitable for the mature meso-scale eddy detection of the marine remote sensing image. |
Key words: Marine remote sensing image, Meso-scale eddy, Automatic detection |