Research Of Fuzzy Direction Relation Based On Raster Data

Direction relation, which is one of the important spatial relations in spatialdatabase, is used widely in many domains, including GIS, multimedia system andimage database. But the direction relation is fuzzy because of the fuzziness ofgeographical objects themselves, the imprecision of geographical informationprocessing and the vagueness of mankind cognition about spatial direction relations,uncertainty is inherent property of direction relations. At present, the research ondirection relation is mainly based on qualitative method, while there is relatively littleresearch on fuzziness. Based on these, the description on fuzzy direction relationsbased on raster data are studied and explored.First of all, the fuzzy direction relation model based on raster data is extended,the fuzzy direction relation model based on raster data is proposed by analyzing theproperty of raster data. At the same time, based on this model, the membership ofdirection relations based on raster data is calculated, the calculation result canrepresent as the target objects that pertain to each direction, and this kind ofcalculation result matches people’s cognitive habits precisely.Secondly, as direction relation research process is easy to overlook the problemof discrete data, one kind of discrete data retrieval algorithm is produced.Discrete data retrieval algorithm assign a dynamic weight value for each discrete dataand continuously update the value of the raster data in the iteration algorithm to findthe discrete data. The algorithm is proved convergent by using a formula derivation.Using the geographical information data test discrete data retrieval algorithm,experiment results show that this algorithm is improved in optimizing the precision.Finally, an experiment is performed to the description model based on the spatialdirection relations. From basic membership function and basic membership functionsof the choice of parameters and fuzzy objects, the present thesis makes a deep study tothe function of this model. Experimental results show that the model can reflect thefuzzy characteristics of direction relations in real world and is fit for the habit of the human mind’s cognition.