基于对象和特征组合的高分辨率遥感影像分类研究

针对高分辨率影像的空间分辨率高而光谱分辨率相对不足的特点,提出一种基于多尺度分割和特征组合的对象级分类方法。试验结果表明,面向对象的分类方法可有效抑制＂椒盐现象＂,显著提高分类精度,更加适合于几何和纹理信息丰富的高分辨率影像的自动识别分类。

Classification of High-Resolution Remote Sensing Image Based on Object Oriented and Feature Combination

A classification method for high-resolution image based on object oriented and feature combination is presented in this paper.The whole image is firstly multi-scale segmented to get objects.Then,the features of objects,such as spectral,geometrical and topological characteristics,are measured.Finally,the classification is carried out by using multi-features and the rules.The research shows that the object oriented method is more efficient than the pixels-based methods,and reduce the ＇Pepper and Salt＇ pheomenon effectively.