Uncertainty Multi-source Information Fusion for Intelligent Flood Risk Analysis Based on Random Set Theory

Information fusion has been a hot topic currently, how to make information fusion for intelligent decision is a challenge. Although the applications of random set theory attract many researchers, the probability function distribution is still imprecise. In this paper, we give a new definition of probability distribution function (PDF) of random set theory, and propose an integration methodology for urban flood risk assessment by fusing multi-source information (e. g., remote sensing images, Digital Elevation Model (DEM) and rainstorm data) based on random set theory. The methodology analyzes and fuses the multi-source information, which overcomes the uncertainty of the decision makers of flood risk and generates precise estimates of the probability of flood risk. In our experiments, we take Wuhan city in China and three kinds of data sources information as an example to assess flood risk level. The experiments indicate that our algorithm not only provide precise estimates of the probability of flood risk but also give the bound of probability.