基于多分类器融合的遥感分类研究
首发时间:2006-12-20
摘要:为了提高遥感图象分类的精度,弥补传统分类方法通过增加单个分类器结构的复杂度来提高分类精度通常不能满足问题的需求的缺陷,提出了一种基于多分类器融合的遥感分类方法,将多个结构较为简单的分类器进行不同策略的融合来提高整体的分类精度。实验结果证明,该方法在遥感分类精度上优于传统分类方法。
关键词: 遥感分类 多分类器融合 模糊积分 模糊密度 遗传算法
For information in English, please click here
A study on classification of multi-spectral remotely sensed image based on Multi-classifier Fusion
Abstract:The traditional way of increasing the complexity of a single classifier can’t satisfy the classification accuracy requirements. We proposed a method based on multi-classifier fusion to improve the accuracy of classification of remote sensing image. We use different strategies to fuse several simple-structure classifiers in order to improve the whole classification accuracy. The result shows the method is better than the traditional ways on the classification of remote sensing image.
Keywords: remote sensing image, multi-classifier fusion, fuzzy integral, fuzzy measure, genetic algorithm
基金:
论文图表:
引用
No.1041890410116659****
同行评议
共计0人参与
勘误表
基于多分类器融合的遥感分类研究
评论
全部评论0/1000