Application Of Evidential Reasoning In Urban Development

Traditional remote sensing image classification methods have been mature, especially the maximum likelihood technique based on statistical analysis methods. But traditional remote sensing image classification methods can't handle multiple source remotely sensed data. In order to make optimized decisions, better use must be made of all available information acquired from different sources. Evidential reasoning has been proposed as one of the most promising approaches for integrating multi-source information.Evidential theory is essentially concerned with reasoning in hypothesis space using Dempster's rule of combination,support committed to hypothesis space. From different pieces of evidence is aggregated into a combined "belief". The optimal hypothesis is decided based on the level of combined belief. The critical step in applying evidential reasoning is evaluating basic probability assignment (BPA), or establishing mass function. In this research, a mass function estimation was explored. A supervised knowledge derivation method is proposed. Several aspects are included in the thesis:1 Apply evidential theory to multi-source data integration.2 Explore the method of quantitative description of evidential support.3 Implement an evidential reasoning algorithm program.4 Using the information fusion based on evidence theory in urban development.5 Test the performance of the evidential reasoning algorithm using a multi-source spatial data set.Evidential reasoning has some advantages. It mustn't know about physical relation among multi-source data and mustn't build multi-source data models;Man-made transcendent knowledge is looked as information which endowmass function that take part in the operation of evidential combination.Look multi-source information as different evidence. Primary attempt using evidential theory in the urban development have been done. Produce Wuhan developing change map using multi-source data fusion. Multi-source information can supplement one another, thus remote sensed image information is more abundant and image interpretation effect is better.