高分二号影像数据地物信息提取分割尺度研究

面向对象技术是提取高分辨率影像中地物信息的主流方法,而多尺度分割是面向对象技术的基础与关键,分割尺度的选择将直接影响最终信息提取的精度与质量。借助于eCognition平台,选用高分二号影像数据作为研究对象,采用多尺度分割的面向对象分类方法,充分利用遥感影像几何、纹理、光谱等信息,确定不同地物类别的最优化分割尺度,建立最佳分类规则,较好的提取了目标地物,更在一定程度上提高了分类精度,为震后灾害评估、震中道路快速提取提供技术支持。

Research on Segmentation Scale of Information Extraction Based on GF-2 Image

Object-oriented technology is the main method of information extraction based on high resolution image.Image segmentation is the basis of Object-oriented technology and key,its segmentation quality is good or bad,will directly affect the accuracy of information extraction.With eCognition software,this paper uses object-oriented classification technology by multi-scale segmenting to extract information based of GF-2 data.Combining the geometry,texture,spectral characterization and so on,determining the optimal segmentation scale of different ground objects,building classification rule to extract the target objects better,and to improve the classification accuracy to a certain extent.Provide technical support for post-earthquake disaster assessment and rapid extraction of epicenter roads.