Research And Application For Flood Forecasting Model Of Reservoir Based On GIS And RS

Reservoir is the important secure base of water conservancy project and water resource, which plays an important role in protecting the security of life and property of people situated downstream and ensuring water supply for industrial and agricultural production and urban people's life. To make the reservoir hold as much water as possible under the condition of its safety, aiming at making full use of flood resource and ensuring the need of water supply, a set of accurate and calculable flood forecasting model is needed to be built to forecast flood scientifically.The rapid development of remote sensing technology provides a sufficient data source for reservoir flood forecasting model. The application of remote sensing data in flood forecasting is the main line of this paper. We studied the problem of extracting the information of watershed underlying surface with remote sensing image using the basic principle of artificial immune system immunity learning; We studied the technological process of extracting the information of watershed underlying surface using the data source of ETM+ data; We also studied the problem of building flow networks using the data source of underlying surface information and DEM; According to the basic theory of distributed hydrological model and research findings mentioned above, we built a reservoir flood forecasting model. We built the experiment system of reservoir flood forecasting and applied it in Xueye Reservoir. The result of the reseach is of great reference value to improving the level of flood forecast, The main contribution of this paper:(l)An algorithm of Classification of Remote Sensing Image based on immune learning was presented. The algorithm used immune learning theory of artificial immune system, which first constructed a linear classifier used for the learning of the characteristics of categories, and studied separately the antigen population of each category to reduce the algorithm's convergence time; In the process of extracting, it distinguished by a variety of different ways and introduced artificial priori knowledge to increase the precision and efficiency of extracting.The experiment showed that:the results of algorithm have high classification accuracy.(2)This paper defined the technological process of extracting the information of watershed underlying surface. Based on the modeling needs of flood forecasting and the characteristics of Landsat-7 ETM+data source, by synthesizing means of band combination, calculating characteristic index and calculating the texture characteristics according to the extracting object, we organized feature space appropriately to extract underlying surface information effectively.(3)The algorithm of building flow networks was developed in this paper. Based on understanding the characteristics of reservoir confluence sufficiently, we extracted runoff grid by the data source of DEM and the information of underlying surface; we did topological processing to the extracted grid and the flow relationship, and determined the computing sequence of runoff grid; and also we carried out the building of flow networks.(4)The reservoir flood forecasting model based on GIS and RS was constructed. According to the basic theory of distributed hydrological model, the model took the extracted information of land surface as the basis to calculate runoff and used the flow networks to compute unit by unit. The experiment results in Xueye Reservoir showed that:The model can effectively improve the forecasting precision.(5)This thesis achieved the experiment system of reservoir flood forecasting based on GIS and RS and applied it in Xueye Reservoir.