Đang chuẩn bị liên kết để tải về tài liệu:
Automatic detection of surface water bodies from Sentinel-1 SAR images using Valley-Emphasis method

The results of proposal method were compared to those of widely used Otsu method and the reference data (e.g. Lansat 7, 8) gave the highest Completeness (User accuracy), Correctness (Producer accuracy) and Quality (Overall accuracies) at 98.8%, 90.7 % and 89.7 %, respectively. The employed method is straightforward, easy to implement and may be applied for other areas even at regional or global scales. | Vietnam Journal of Earth Sciences 37 (2015) 328-343 Vietnam Academy of Science and Technology Vietnam Journal of Earth Sciences (VAST) Website: http://www.vjs.ac.vn/index.php/jse Automatic detection of surface water bodies from Sentinel-1 SAR images using Valley-Emphasis method Nguyen Ba Duy University of Mining and Geology Accepted 25 December 2015 ABSTRACT Surface water resource plays as an important role in human daily life and in the eco-environment. In the study Valley-Emphasis method of automatic water extraction was employed to identify surface water bodies at three study areas, having different landscapes and covers, using Sentinel-1A IW images widely used automated Otsu method was performed for extracting surface water bodies to compare proposed method. The results of proposal method were compared to those of widely used Otsu method and the reference data (e.g. Lansat 7, 8) gave the highest Completeness (User accuracy), Correctness (Producer accuracy) and Quality (Overall accuracies) at 98.8%, 90.7 % and 89.7 %, respectively. The employed method is straightforward, easy to implement and may be applied for other areas even at regional or global scales. The method also improves automatic identification level of surface water bodies, providing essential information for flood disaster research. Keyword: Surface Water body, Valley-Emphasis Algorithm, Sentinel-1, SAR. ©2015 Vietnam Academy of Science and Technology 1. Introduction The ability to map open surface freshwater (i.e. lakes, ponds, reservoirs) is integral to many hydrologic and agricultural models, wildlife management programs, and recreational and natural resource studies (Henderson, 1995). Concomitantly, the visibility and appearance of open surface water bodies should have supplemental value in radar geoscience research addressing the relationships among radar wavelength, specular reflectance, smooth/rough signal return criteria, and terrain features, especially when spatial variables are .