利用高分辨率影象提取土地覆盖信息的关键技术在于如何利用丰富的纹理信息来弥补光谱信息的不足。面向对象的图像分类技术改变了传统的面向像素的分类技术:(1)用来解译图像的信息并不在单个像元中,而是在图像对象和其相互关系中;采用多分辨率对象分割方法生成图像对象,提高了分类信息的信噪比;基于对象的分类技术不同于纯粹的光谱信息分类,图像对象还包含了许多的可用于分类的一些其他特征:形状、纹理、相互关系、上下关系等信息。面向对象的土地覆盖分类结果与传统分类方法相比,其特征提取算子更加地适合于几何信息和结构信息丰富的高分辨率图像的自动识别分类。 The key technique to extract land cover information from high-resolution images is to use rich texture information to make up for the lack of spectral information. Object-oriented image classification technology has changed the traditional pixel-oriented classification technology: (1) the information used to interpret the image is not in a single cell, but in image objects and their relationship; the use of multi-resolution object segmentation Method to generate image objects and improve the signal-to-noise ratio of classification information. Object-based classification technology is different from pure spectral information classification. Image objects also contain a number of other features that can be used for classification: shape, texture, Relationship and other information. Compared with the traditional classification methods, the object-oriented land cover classification results are more suitable for the automatic identification classification of high-resolution images with rich geometric information and structural information.