遥感影像亚像元制图方法研究进展综述

遥感影像混合像元的普遍存在给遥感影像解译造成困扰。有效处理混合像元问题,细化分类结果,获得更为精细的地物细节信息就需要进行亚像元绘图。目前亚像元制图方法主要包括3个步骤:① 混合像元分解；② 提取软信息；③ 亚像元制图。总结归纳了近年来遥感影像亚像元绘图领域的研究进展和成果,详细阐述了亚像元制图的步骤及涉及的研究方法。依据辅助信息的类型将亚像元绘图方法大致划分为:基于空间相关性、基于空间结构信息、基于神经网络、基于像元交换途径的4类亚像元分类方法,并分别对各种方法的优缺点进行了分析对比。最后,评述了亚像元制图的发展趋势。

Progress on Sub-pixel Mapping Methods for Remotely Sensed Images

Mixed pixels commonly exist in remote sensing images.They may cause difficulties to process and analyze the remote sensing images.Sub-pixel mapping method is used to solve the problem of mixed pixels.It can obtain more accurate results for image classification and more details of distribution of land cover classes.There are always three steps to preform the sub-pixel mapping:① unmixing the mixed pixel; ② obtaining the soft information;③ mapping at sub-pixel scale.This paper described the details about the sub-pixel mapping method in recent years and reviewed the development of these sub-pixel mapping methods.These methods can be classified into four types:① Spatial correlation-based sub-pixel mapping;② spatial construct information-based sub-pixel mapping;③ article neural network -based sub-pixel mapping;④ pixel swapping-based sub-pixel mapping.Then these methods are contrasted and analyzed as well.Finally,a trend for development of sub-pixel mapping is given in the conclusion