【作者】 赵灵军；

【导师】 陈圣波；

【作者基本信息】 吉林大学 ， 地图制图学与地理信息工程， 2005， 硕士

【摘要】 预测是对研究对象的未来状态进行预计和推测。预测模型的建立是为了体现事物发展的趋势。在充分研究灰色系统理论的基础上,利用统一建模语言UML 进行了空间信息灰色预测模型的整体设计;利用“线-基本线-线”的分解合成过程构造了灰色预测核的二元参数预测;利用“网格相交”和“栅格技术+”, 以MapObjects 为例,在灰色预测核的三元参数预测的基础上进行矢量数据的预测;以BMP 图像为例,进行栅格数据的预测。以吉林市市区发展预测为例,验证了预测模型的正确性。更多还原

【Abstract】 Prediction is to forecast the future state of the research object, and the result can tell us the execution of the forecasting. Based on graphics and image, this paper designs a spatial information-based grey prediction model, which can be used in many scientific fields. Using the object-oriented (OO) techniques as the guiding principle, this paper takes advantage of Unified Modeling Language (UML) to design a prediction model of spatial data. Based on object-oriented four principles (abstract, information hiding, encapsulation and modular), the prediction model, including the prediction core, the grid data prediction, the vector data prediction, and the user-defined prediction, are established. Each module is independent, and each module has the same interface standard. Grey Model (GM) is a basic model in the grey systematic theory and gray control theory. Based on grey module, differential coefficient fitting method is used in this GM. GM possesses little information for modeling, unnecessary to know the original data distributing, turning disordered orientation to ordered orientation, high precision of modeling, and so on. Therefore, in this study, grey prediction model, GM (1, 1), as follows: According to the array dimension of historical data and their spatial shapes, grey prediction in divided into three sections: one-dimensional parameter prediction, two-dimensional parameter prediction and three-dimensional parameter prediction. One-dimensional parameter prediction is built based on grey prediction, which can predict discrete data. The two-dimensional parameter prediction is visually realized based on the one-dimensional parameter prediction, which can forecast raster image. The three-dimensional prediction is built based on the two-dimensional prediction, which is mainly used to predict vector data. For the two-dimensional prediction, the procedure of “line-basic line-line”is proposed. For three-dimensional prediction, the procedure of “polygon-line-basic line-line-polygon”is defined, by which the polygon is turned into line, and the line is turned into basic line as a minimal unit for prediction. A new concept “calculation zone”is defined during the process更多还原