一种基于分组式蛇模型的GIS矢量边界更新方法

A Method of Vector Edge Updating Based on Grouping Snake Model in GIS

提出一种基于分组式蛇模型的GIS矢量边界自动更新方法。以GIS已有矢量数据作为初始位置提供先验知识,以遥感影像RGB值的梯度信息作为轮廓演化时的外部能量,可以使蛇模型将遥感影像与GIS数据相结合,在二者交互作用下实现GIS矢量边界更新。将传统蛇模型与贪婪算法的特点相结合,设计一种折中模型——分组式蛇模型。仿真实验表明,分组式蛇模型改善蛇点局部定位准确性并提高了蛇点之间的关联性,较传统蛇模型和贪婪算法具有更好的更新准确率与多边形相似度。

An automatic approach of updating vector edge in GIS based on Grouping Snake is proposed in this paper. The vector data existing in GIS are refined using the snake model in terms of information in remote sensing images, during the whole process of which, vector data oct as both the updated objects, and the prior knowledge that conducts change detection in remote sensing images. There are two typical algorithms about snake model opposite in the computational complexity the original snake model and the greedy algorithm. A trade-off Grouping Snake is put forward, attempting to group control nodes, in order to decrease the poor localization and the weak neighbour-node relationship, respectively caused by the original snake and the greedy algorithm. A few experiments are performed on vector edges and TM remote sensing image of Zhalong Wetland in China to validate Grouping Snake in vector edge updating.