| 摘要: |
| 针对利用传统瞬时水边线提取方法处理高分辨率遥感影像存在提取结果不连续、效率不高和无法同时提取多片水域等问题, 提出了一种顾及轮廓信息和距离正则化水平集演化(distance regularized level set evolution, DRLSE)模型的遥感影像瞬时水边线快速提取方法, 并将其应用于福建泉州附近海域瞬时水边线提取。首先,使用DRLSE模型提取地物轮廓信息, 以解决经典阈值方法水边线提取结果不连续问题; 其次, 利用DRLSE模型的初始矩形轮廓中心位置和周长信息, 对噪声点等轮廓进行自动剔除, 并提取多片水域, 以提高瞬时水边线提取后处理效率。研究结果表明:相较于泛洪算法、Canny算子和CV(Chan-Vese)模型, 应用本方法进行大陆海岸线瞬时水边线提取更高效, 且提取结果连续、精度更高。 |
| 关键词: 瞬时水边线 DRLSE模型 初始轮廓信息 泛洪算法 Canny算子 Chan-Vese模型 |
| DOI：10.11759/hykx20221102001 |
| 分类号:P229; P748 |
| 基金项目:中国科学院战略性先导科技专项(A类)子课题“祁连山冻土水文变化及其对供水安全的影响”(XDA20100103);中国地质调查局地调项目“海洋地质信息系统建设及产品开发”(DD20191006) |
|
| Fast extraction of instantaneous waterline from remote sensing based on contour information and DRLSE model |
|
YANG Zhi-hai1, WANG Guang-jun1, GAO Zhi-xiong2, FENG Lei3, LIANG Si-hai4
|
|
1.School of Land Science and Technology, China University of Geosciences(Beijing), Beijing 100083, China;2.CNNC 7th Design and Research Institute Co., LTD., Taiyuan 030012, China;3.Aero Geophysical Survey and Remote Sensing Center for Natural Resources, China Geological Survey, Beijing 100083, China;4.School of Water Resources and Environment, China University of Geosciences(Beijing), Beijing 100083, China
|
| Abstract: |
| To solve the problems of the discontinuous extraction, low efficiency and inability to extract multiple water areas simultaneously in traditional methods for extracting the instantaneous water lines from high-resolution remote sensing images, a fast extraction method of the instantaneous waterline is proposed, which considers contour information and distance regularization level set evolution (DRLSE) model, and the method is applied to extract instantaneous waterline of the sea area near Quanzhou City, Fujian Province. Firstly, DRLSE model is used to extract ground contour information to solve the problem of discontinuous waterline by classical threshold methods; Then, the center position and perimeter information of the initial rectangle of DRLSE model is used with the noise points and other undesired contours are automatically eliminated, and multiple water areas are extracted, thus the post-processing efficiency of instantaneous water lines extraction is improved. The research results show that this method is more efficient in extracting instantaneous waterline of continental coastline compared with Flood fill algorithm, Canny algorithm and CV model, and the extraction results are continuous and more accurate. |
| Key words: instantaneous waterline DRLSE model initial contour information flood fill algorithm Canny algorithm Chan-Vese model |
