遥感影像分类方法研究进展

A Review of Classification Methods of Remote Sensing Imagery

遥感影像分类是遥感信息提取的重要手段,是目前遥感技术中的热点研究内容。分类方法是遥感影像分类的重要内容,有效地选择合适的分类方法是提高分类精度的关键。随着遥感技术的发展,传统的非参数分类方法已经难以满足分类精度需求,基于智能算法的非参数分类方法得到了迅速发展,并在遥感影像分类中发挥着重要作用。近年来,组合分类器由于能够利用单一分类器的互补信息,成为了遥感影像分类的一个新热点。本文综合分析了各种分类方法的特点和优势,及分类方法的发展趋势,为遥感影像分类技术的发展提供科学的参考依据。

Remote sensing data classification is an important way of information extraction and a hot research topic of remote sensing technique. Classification method of remote sensing data is an important issue, and effective selection of appropriate classifier is especially significant for improving classification accuracy. Along with the development of remote sensing technique, traditional parametric classifier is difficult to meet accuracy requirement, leading to the rapid development of intelligent algorithm based non-parametric classifiers. Recently, combined classifiers become a new hot topic for its ability of utilizing complement information of single classifier. In the present paper, characters and advantages of different classifiers as well as the research prospect are analyzed. The paper provides a scientific reference for the development of remote sensing data classification technique.