径向基函数网络与WebGIS融合的苹果病虫害预测  

Forecast of Disease and Pest in Apple Orchards Based on Fusion of Radial Basis Function Neural Network and WebGIS

作　　者：李丽[1] 李道亮[1] 周志坚[2] 张银[3] 傅泽田[1] 

机构地区：[1]中国农业大学教育部现代精细农业系统集成研究重点实验室 [2]中国农业大学理学院 [3]中国农业大学党政办公室

出　　处：《农业机械学报》2008年第3期116-119,153,共5页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家“863”高技术研究发展计划资助项目(项目编号:2002AA243031)

摘　　要：以ArcIMS为二次开发平台,开发了基于Java技术和MVC架构的苹果病虫害预测预报系统平台,在建立了20余种苹果病虫害预测模型的基础上,采用动态数据交换技术(DDE)实现了GIS分析功能与病虫害预测模型的集成;具有空间数据和属性数据的分析处理、苹果病虫害预测预报、WebGIS信息发布等功能;可以通过地图或专题图、数据表格、图形或文字等多种形式预测预报苹果病虫害发生时间和发生程度。以苹果山楂叶螨为例,详细阐述了系统中径向基网络模型的构建,该模型的测试准确率为87.5%,高于BP神经网络的75%。A forecast system of diseases and pests for apple orchards was developed by using MVC structure with Java support based on ArclMS. The forecast model, which could predicate 20 kinds of the diseases and pests, was integrated with spatial analyses of GIS by dynamic data exchange （DDE）. The system can be applied to analyze spatially and temporally, forecast the diseases and pests, and promulgate the information with WebGIS. The forecasting of the regions and infection tendency of diseases and pests can be revealed with several patterns, such as map filled with different colors, topic graph, data table, text, or figure. Additionally, for instance of Tetranychus Viennensis in apple orchards, the predicting model constructed with radial basis function neural network （RBFNN） was purposed in this paper, and the predicting results were compared with those of back-propagation （BP） neural network model.

关 键 词：苹果 病虫害 预测 径向基函数网络 WEBGIS 

分 类 号：TP399[自动化与计算机技术—计算机应用技术] S436.611[自动化与计算机技术—计算机科学与技术]