果树施药仿形喷雾的神经网络统一模型

Unified Model of Neural Net Model for Pesticide Modeling Spray to Fruit Trees

在果树施药仿形喷雾研究中,为建立反映果树施药仿形喷雾过程参量与分布质量系数之间映射关系的神经网络模型,必须通过不断的尝试,获得合适的网络结构、权值和阈值。然而,由于果树仿形喷雾的影响因素多且重复性差,势必增加计算的复杂度和成本。为此,构建了一种基于面向对象的神经网络统一模型。研究结果表明,该模型具有良好的扩展性和复用性。在系统设计时,运用了不同的设计模式,提高了研究的效率,且代码具有可复用性和重构性。

In order to research Neural Net Model for Profile Modeling Spray of chemical to solution application to Fruit Trees, a neural net model has been constructed to map the relation ship of process variables and distributing quality of Profile Modeling Spray of the Fruit Trees, which do a lot of experiment to find the net-structure, power and threshold. However, there are many affect in the Profile Modeling Spray so that it is hard to repetition, and cost more computing complex. In this paper, Design Pattern based Unified Model of Neural Net Model was constructed. The result shows that, the model has expansibility and duplicate; during design the system, more design patterns were used, which improve the efficiency of research and the code has duplicate and reconstructed.