TM和SAR遥感图像的不同层次融合分类比较  

Comparison of TM and SAR Remote Sensing Image Different Level Fusion Classification Effect

作　　者：于秀兰[1] 钱国蕙[1] 

机构地区：[1]哈尔滨工业大学航天电子与光电工程系,哈尔滨150001

出　　处：《遥感技术与应用》1999年第3期38-43,共6页Remote Sensing Technology and Application

基　　金：国防基金

摘　　要：多传感器遥感图像为分类技术提供了更多的地物特征信息，有助于提高分类精度，增强计算机自动解译的能力，减少遥感图像的后处理时间。给出了ＴＭ 和ＳＡＲ遥感图像根据信息处理所在的不同层次融合分类特征的选取及分类方法，并从融合分类的精度和计算时间消耗上对哈尔滨市附近地区５１２×５１２ 大小的ＴＭ和ＪＥＲＳ１ 的ＳＡＲ图像的不同层次融合的分类进行比较。Multi sensor remote sensing image fusion technique provides more information of ground objects to classification than single sensor image. And it is helpful to improve classified accuracy, enhance computer automatic interpreting capability, and reduce the remote sensing image postprocessing time. The characters selection and classified method of TM and SAR remote sensing image different level fusion are given in this paper, and the different level fusion classified results of 512×512 TM and JERS 1 SAR images near Haerbin are compared as well in classified accuracy and cost time.And character level fusion is recognized the most suitable level in the fusion classification according to the classification accuracy and computing cost.

关 键 词：遥感图像 TM SAR 图像处理 图像融合 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]