| 摘要: |
| 为提高海上溢油轮廓SAR提取精度, 验证了FCM(Fuzzy C-Means Algorithm)与DRLSE(Distance Regularized Level Set Evolution)模型结合的方法提取SAR溢油信息的有效性; 鉴于其无法避免细小噪音的影响以及薄油膜提取效果不好的问题, 提出了阈值和DRLSE模型结合的溢油信息提取方法, 通过阈值构建溢油区域初始轮廓, 克服了图像细小噪声对溢油提取的影响, 更有利于提取薄油膜信息, 溢油提取精度优于H/A/alpha-Wishart非监督分类方法和FCM与DRLSE模型结合的方法。 |
| 关键词: DRLSE 模型 SAR 溢油提取 阈值 |
| DOI：10.11759/hykx20171011006 |
| 分类号: |
| 基金项目:国家重点研发计划项目(2017YFC1405600); 国家自然科学基金(41706208, 41776182); 山东省自然科学基金(ZR2016DM16) |
|
| SAR oil-spill extraction method based on DRLSE model |
|
LIU Shan-wei,WANG Wan-di,LI Xiao,CHEN Yan-long,ZHANG Ting
|
| Abstract: |
| In this study, we evaluated the SAR information extraction of oil spilled at sea and the effectiveness of combining the fuzzy C-means (FCM) and distance regularized level set evolution (DRLSE) models to extract SAR oil-spill information. In light of the inability of this approach to prevent small-noise effects and its poor thin-oilfilm extraction performance, we propose a method for extracting oil-spill information that combines threshold data and the DRLSE model. With this method, the initial contour of the oil-spill region is constructed based on the threshold, which overcomes the influence of small noises on the oil extraction, and the extraction of thin-oil-film information is facilitated. Our method demonstrates better oil-extraction precision than the H/A/alpha-Wishart unsupervised classification method and the combined FCM and DRLSE models. |
| Key words: DRLSE model SAR oil spill extraction threshold |
