证据推理应用于多源信息融合分析

The Research of Data Fusion from Multiple Sources by Evidential Reasoning

运用证据理论合并来自不同数据源的证据 ,以实现各种数据所含信息融合的原理和方法。该方法的最大特点是将数据源中存在的不确定性引入数据分析过程。另外 ,利用证据推理 ,可以在数据分析过程中方便地引入专家解译图像的经验和知识。最后通过利用ERSSAR数据和TM影像进行融合分析 ,证明该方法在遥感图像自动分类中有很好的应用前景。

The principle and method of evidential reasoning are described in the paper, which is used for combining the information from multiple data sources so that various information can be fused in remote sensing image analysis. What is most important in the method is that the uncertainty in data sources can be propagated in the process of data analysis, and the expert knowledge can also be used in the method. At the end, it is demonstrated that evidential reasoning has extensive applications in the classification of remote sensing images through the fusion analysis of ERS SAR and TM image.