语义网络引导的遥感图像分类  

Semantic Network Guided Remote Sensing Image Classification

作　　者：张建廷[1] 张立民[1] 徐涛[1] 

机构地区：[1]海军航空工程学院,山东烟台264001

出　　处：《遥感信息》2016年第1期38-42,共5页Remote Sensing Information

基　　金：国家自然科学基金(61102167)

摘　　要：为将领域知识与基于对象的图像分析范式有效地结合起来,以解决遥感图像分析中的语义鸿沟问题,提出了一种语义网络引导的遥感图像分类方法。该方法包含分割、预分类、结果修正3个步骤。采用语义网络引导的分类方法进行预分类,使用定性匹配度进行图像对象与概念匹配判定,定义了一种定量匹配度,并将其用于对预分类结果的修正。方法减小了图像对象与概念进行匹配的搜索空间,降低了运算复杂度。实验结果证明了该方法的有效性。In this study,a new procedure for remote sensing image classification based on semantic network was proposed to combine domain knowledge with object-based image analysis （OBIA） ,in order to deal with the semantic gap problems in remote sensing image analysis. Initial classification was guided by semantic network using qualitative matching degree. A quantitative matching degree was defined for the modification of the initial classification. The proposed method reduced the search space for the matching between image objects and concepts,and had lower computational complexity. It was validated by an urban image classification experiment.

关 键 词：基于对象图像分析 遥感图像分类 语义网络 匹配度 图像分割 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]