文档介绍：2
2
2
2
2
2
2
2
2 2
安徽农业科学, Journal of Anhui Agri. Sci. 2009, 37 (23) : 11316 - 11317, 11329 责任编辑张明明责任校对张士敏
基于 BP神经网络的猪舍有害气体定量检测模型研究
1 1 2
俞守华,张洁芳,区晶莹(1. 华南农业大学信息学院,广东广州 510642; 2. 华南农业大学公共管理学院,广东广州 510642)
摘要为寻找适合猪舍混合有害气体浓度识别的神经网络模型,建立了基于误差反向传播(BP)神经网络的猪舍有害气体定量检测模
型,分别使用 trainbr函数、traingdm函数及 trainlm 函数训练该神经网络,对有害氨气和硫化氢组成的混合气体浓度进行识别,并利用
MATLAB软件的神经网络工具箱进行仿真。结果表明,采用 trainbr函数训练的网络对该混合气体的平均识别精度高,速度较快,对噪声
不敏感,适合猪舍有害气体的浓度识别。这为猪舍有害气体智能化监控提供了参考依据。
关键词 BP神经网络;猪舍;气体定量检测
中图分类号 S815. 9 文献标识码 A 文章编号 0517 - 6611 (2009) 23 - 11316 - 02
Quan titative D etection M odel of Pern ic ious Gases in P ig House Ba sed on BP Neura work
Y U Shou hua et a l (College of Informatics, South C2hina Agricultural University, Guangzhou, Guangdong 510642)
Abstract To find a work model suitable to identify the concentration of m ixed pernicious gases in pig house, the quantitative detec
tion model of pernicious∃gases in pig house was set up based on BP (Back p ropagation) work. The BP work was trained
separately by the three functions, trainbr, traingdm and trainlm, in order to identify the concentration of m ixed pernicious posed of he
2 2
patic gas and ammonia gas. The work toolbox in MATLAB software was used to simulate the detection. The results showed that the
work trained by trainbr function has high average identification accuracy and faster detection spe