欢迎访问林业科学,今天是

›› 2006, Vol. 42 ›› Issue (zk): 155-158.

• 论文及研究报告 • 上一篇    下一篇

基于VLBP神经网络的林火预测研究

胡林 冯仲科 聂玉藻   

  1. 北京林业大学省部共建森林培育与保护教育部重点实验室,北京100083
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2006-09-25 发布日期:2006-09-25

Forest Fire Prediction Research Based on VLBP Neural Network

Hu Lin,Feng Zhongke,Nie Yuzao   

  1. The Key Laboratory for Silviculture and Conservation of Ministry of Education,Beijing Forestry University Beijing 100083
  • Received:1900-01-01 Revised:1900-01-01 Online:2006-09-25 Published:2006-09-25

摘要: 以北京市房山区的林火为研究现象,用神经网络方法研究林火发生及蔓延与环境因子间的关系。用GPS采集林火火场中心数据,测量火场面积;用GIS做空间分析,提取火场空间环境因子;将数据在Matlab中用VLBP神经网络进行训练,建立林火风险的预测方程。风险模型由林火发生的风险模型和林火的蔓延风险模型组成,由于该模型采用的方法不依赖于先验统计模型,所采用的神经网络方法对非线性的数学关系有很强的预测能力,因而该方法具有很强的适用性。

关键词: 林火, 人工神经网络, 风险模型, ArcView GIS

Abstract: In this research,the Fangshan district as the research district,the relation between the fire risk and its environment factors were studied using the artificial neural network methods. The procedure framework is as the fellow: the fire regimes' center and its' area is surveyed with GPS,and the data was put into the GIS for spatial analysis. The data of the environment factors affecting the forest fire were withdrawn using the GIS,training the data in Matlab with the VLBP artificial neural network,and establishing the forest fire risk forecast equation,this risk model is composed with the fire ignition model and the forest fire spread model. Because this model uses the artificial neural network which does not rely on fire data's probability distribution,and the nerve network method has the very strong forecast ability to the non-linear mathematical relation,thus this model has the very strong versatility.

Key words: forest fire, artificial neural network, risk model, ArcView GIS