摘要
为了更好地提取影像阴影部分的信息,本文采用一种顾及地物间空间关系的水系提取方法,将地物之间的空间关系与面向对象的方法相结合。通过多尺度分割算法生成同质的影像对象,利用决策树、隶属度函数和阴影与建筑物的邻接关系对水体进行提取。以昆明某区域影像为实验数据进行了试验,试验结果表明,所提出的方法可以有效解决阴影对水系提取造成的影响,大大提高了水系提取的精度。
In order to extract the intormation ofimages shadow well, this paper used a water extraction method taking the relation- ship between space and features into account, and it combined with tile objeet-oriented method. The water information was extracted through the homogeneous images generated from multi-scale segmentation algorithm, the decision tree, the membership function and the adjacent relationship between shadows and buildings. A studied area in Kunming was taken into experiment, the results showed that the proposed method could effectiw^ly remove the impact of shadows on water, and improve greatly the extraction accuracy of the water.
出处
《测绘科学》
CSCD
北大核心
2013年第3期163-165,共3页
Science of Surveying and Mapping
基金
国家自然科学基金项目(41061043)
云南省教育厅科学研究基金项目(2011J075)
关键词
遥感影像
面向对象
地物空间关系
多尺度分割
remote sensing ,image
object-oriented
spatial and feature relationship
multi-scale segmentation
