遥感分类信息不确定性的可视化表达方法

Visualizing Uncertainty in Classified Remote Sensing Images

影像分类是遥感技术的重要应用领域之一,而遥感分类信息中则存在着不确定性,如果仅仅对其进行不确定性程度上的数学度量,而忽略其在空间域和时间域上的分布特征,则很难完整准确地描述和理解遥感信息中的不确定性。从视觉感知角度表现不确定性信息是不确定性建模的重要组成部分,而利用现有的计算机可视化技术来表达遥感分类信息的不确定性则是一种直观有效的方法,可以帮助用户更加完整地理解遥感数据及其分类信息中不确定性的大小、分布、空间结构和趋势。我们将不确定性可视化技术分为静态可视化变量、动态可视化变量和特征可视化三种类型。并以我国黄河三角洲地区的Landsat 5 TM遥感影像的分类数据为例,对上述不确定性可视化技术进行讨论及分析。

The inherent uncertainties in the real world and the limitation of human cognition are the primary factors of the abundance of uncertainty existing in the judgment of sciences to objectivities. An amount of uncertainties exist in the spatial data describing the spatial entity, including temporal-relative location data and attribute data, some of which may seriously distort the reliability of products. The remotely sensed imagery contains uncertainties caused by unexpected factors during the acquiring process. If only measuring the degree of uncertainty but ignoring the spatial-temporal distribution features, it is not easy to describe accurately and completely the uncertainty in remote sensing information. Representing uncertainty information from the point of view of visual sensation is one of the important sections of modeling uncertainty. The uncertainty visualizing techniques will help users to explore uncertainty in raw data and size, distribution, special structure and tendency of uncertainty. Using existing computer visualization technologies to express uncertainties is an intuitive and effective way. In this paper, these visualization techniques include three types of visualization： static visualization, dynamic visualization and feature visualization. Based on the Landsat-5 TM remote sensing image covering the Chinese Yellow River delta area and its classified data, we investigated several techniques discussed above.