论文部分内容阅读
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。
The ubiquitous presence of remote sensing image mixed pixels disturbs interpretation of remote sensing images. Effectively deal with the problem of mixed pixels, refine the classification results, to obtain more detailed information of objects need to sub-pixel drawing. At present, the sub-pixel mapping method mainly includes three steps: ① mixed pixel decomposition; ② extract soft information; ③ sub-pixel mapping. Summarizes the research progress and achievements in the field of remote sensing sub-pixel mapping in recent years, and elaborates on the steps of the sub-pixel mapping and the research methods involved. According to the types of auxiliary information, sub-pixel drawing methods are roughly divided into four types of sub-pixel classification methods based on spatial correlation, spatial structure information, neural network based on pixel exchange approach, The advantages and disadvantages of the analysis and comparison. Finally, the development trend of sub-pixel cartography is reviewed.