论文部分内容阅读
根据元胞自动机的基本原理,结合云图发展、移动、衰减的特征,提出了一个新的基于元胞自动机的云图短时预测方法。该方法将网格化的云图像素点作为元胞空间的基本单元,邻居定义为摩尔型。通过制定云图像素点的移动、发展和衰减的转换规则,预测短时(12 h内)的云图。试验表明,这种新方法的预测云图与实测云图的基本特征比较相似,特别是预测的整体范围和变化趋势效果较好。
According to the basic principle of Cellular Automata (Cellular Automata), a new cloud-based short-term prediction method based on cellular automaton is proposed based on the characteristics of cloud development, movement and attenuation. The method takes the gridded cloud image pixel as the basic unit of the cell space, and the neighbor is defined as the mole type. Predict the short-term (within 12 h) of the cloud map by formulating the conversion rules of cloud pixel motion, development and attenuation. Experiments show that the basic features of the new method are similar to those of the actual cloud image, especially the predicted whole range and trend are better.