摘要
为有效提高基于遥感技术的海岛调查中海岛提取的效率,该文提出了基于多窄带模型(Multi Narrow Band Model,MNBN)初始化的Grab Cut分割算法对高分辨率遥感图像中的海岛进行提取。首先基于海面与陆地提取进行海岛初提取,再利用初提取结果建立多窄带模型对Grab Cut进行初始化来提高分割算法的速度。分析表明,该算法可有效完成海岛提取,与其他自动海岛提取算法相比具有速度快、鲁棒性高等优点。
Remote sensing technology is widely applied in island investigation in recent years.To improve the efficiency of island extraction,a new extraction method was proposed,which includes two main stages：global initial extraction based on color feature and local re-extraction based on grab cut algorithm.Speed and robustness of island extraction is improved by the proposed method compared with other island automatic extraction methods.
出处
《遥感信息》
CSCD
北大核心
2015年第6期63-66,70,共5页
Remote Sensing Information
基金
海洋公益性行业科研专项经费项目资助(201405028)
关键词
海岛提取
高分辨率遥感图像
GRAB
CUT
多窄带模型
多尺度分割
island extraction
high resolution remote sensing
Grab Cut
multi-narrow-band model
multi-scale segmentation
