区域生长与GVF Snake模型相结合的水域边界提取

Water Area Boundary Extraction Based on GVF Snake Model and Region Growth

利用梯度矢量流GVF(Gradient Vector Flow)Snake模型实现了全色遥感影像上水域边界的智能提取。对于形状特别复杂的水域,单独利用GVF Snake模型提取其边界,存在着人机交互繁琐、时间消耗大、提取精度有待提高等问题。为了解决这些问题,提出了一种自动初始化的GVF Snake,利用区域生长法自动提供GVF Snake的初始值。从而简化了人机交互,降低了时间消耗,提高了提取精度。

Intelligent extraction of water area on panchromatic remote sensing image has been realized using GVF Snake model. However, when it came to water area with quite complicated shapes, there were several problems such as inconvenient man machine interaction, large time consuming and low extraction accuracy. To solve these problems, a novel GVF Snake with automatic initialization was put forward. GVF Snake was initialized automatically by region growth algorithm. It was proved that convenient man-machine interaction, less time consuming and better accuracy were obtained by this method.