摘要
利用2009年12月-2010年5月塑料大棚内外观测的气象数据,构建了基于BP神经网络的杨梅生产大棚内的最高、最低气温预测模型,根据逐时转化系数计算出棚内相应的逐时气温,达到逐时预报大棚内气温的目的。通过模拟回代和对独立试验数据的验证,基于BP神经网络模型对大棚内日最低气温、日最高气温和逐时气温预测值与实际值的回归估计标准误差(RMSE)分别为0.8℃、1.4℃和0.7℃,精度明显高于同时利用逐步回归法建立的模型。该模型所需参数少,实用性强,模拟精度高,可为设施杨梅气象服务和环境调控提供依据。
The minimum and maximum temperature prediction model inside greenhouse planted Myica rubra was established based on BP neural network,by using meteorological data both inside and outside the greenhouse from December 2009 to June 2010 in Wenzhou of Zhejiang province.Using the independent experimental data and simulation back generations to verify the model,the results indicated that the root mean square error（RMSE） between the predicted value and measured value based on 1∶ 1 line for the minimum and maximum and hourly inside air temperature were 0.8℃,1.4℃ and 0.7℃,respectively.The precision of BP neural network model was higher than that of the stepwise regression model obviously.The model,with few parameters,could predict the greenhouse temperature more accurately,which could provide scientific basis for facility meteorological service and environment regulation of greenhouse Myrica rubra cultivation.
出处
《中国农业气象》
CSCD
2011年第3期362-367,共6页
Chinese Journal of Agrometeorology
基金
公益性行业(气象)科研专项"设施农业及特色农产品气象保障关键技术研究"(GYHY200906023)
科技部农业科技成果转化资金项目"基于GIS的特色经济作物种植潜力分析与示范"(2008GB24160442)
浙江省科技厅科技计划"浙江杨梅气候生态研究"(2005C33050)
关键词
神经网络
气温
模拟模型
设施杨梅栽培
Neural network
Air temperature
Simulation model
Greenhouse planted Myrica rubra
