基于面向对象分类技术的景观信息提取研究

依据高分辨率遥感影像的特点,如何充分地利用影像的光谱信息和空间信息以及地学特征进行更为微观的遥感监测或大比例尺制图是高分辨率遥感研究的重要内容之一。本文以地形复杂的武夷山自然保护区为研究区域,以SPOT5原始影像为数据源,采用面向对象的多尺度分割方法,实现了不同尺度地物信息的分层提取。基于上层的分类结果对特定地物进行影像分割,选择合适的特征参数,并通过多次试验建立影像对象的隶属度函数,或利用最邻近分类法,逐级分层地提取了研究区的景观信息。研究结果显示：利用面向对象分类方法可以快速方便地对地形复杂条件下研究区的SPOT5遥感影像的景观信息提取,精度为76%,为高分辨率遥感影像的信息提取提供了更为快速、有效的技术途径。

Study on Extraction of Landscape Information Based on the Object-oriented Classification Techniques

For the high-resolution remote sensing images, how to use the information of spectrum and space to do more microscopic monitoring or large-scale remote sensing mapping is one of the important contents of high-resolution remote sensing. In this case study, the SPOT5 image of Wuyi Mountain natural reserve area was hierarchically classified with the objects-oriented method. Firstly, the image was segmented synthetically combined with the information of spectrum and space. And then, hierarchical classification was realized to extract the landscape information by means of the membership functions or nearest classification. The result showed that： the object-oriented method can quickly and easily exact the landscape information in POT5 images for the complex study area in topography and the accuracy reaches to 76%. This approach provides a new way for classification of high-resolution remote sensing data.