遥感图像分类方法评析与展望  

Expectation and Evaluation of the Methods on RemoteSensing Image Classifications

作　　者：杜凤兰[1] 田庆久[1] 夏学齐[1] 

机构地区：[1]南京大学国际地球系统科学研究所,江苏南京210093

出　　处：《遥感技术与应用》2004年第6期521-525,共5页Remote Sensing Technology and Application

摘　　要：从遥感图像分类的基本概念和特征出发,在对各算法进行划分基础上,分析评价了主要遥感图像分类方法的原理、特点,优势与限制,并阐明了各主要方法作用及其意义,指出目前存在的问题,展望了遥感图像分类方法的发展与应用前景。This paper is based on the concept and characteristic of Remote Sensing classifications, combining the predecessors' fruits, then sums up all the methods (including traditional and new classifications, such as: object-oriented, sub-pixel classification etc.), detailed analyzes and evaluates each of them. Afterwards, it attempts to analyze the methods and the arithmetic of classifications, such as principles, effects, functions and limited conditions. It also compares the excellence, the defects and adapting conditions of each method, and summarizes all of the methods in different views. Finally, summary and prospect are described, in order to give guidance to practical work.

关 键 词：遥感 图像 分类 算法 像元 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]