基于BP神经网络识别毛竹冻融循环次数  

Recognizing the Number of Freeze-Thaw Cycles of Phyllostachys edulis Based on BP Neural Network

作　　者：吴志勇 WU Zhi-yong(Fujian Forestry Prospect And Design Institute,Fuzhou 350002,Fujian,P.R.China)

机构地区：[1]福建省林业勘察设计院,福建福州350002

出　　处：《森林防火》2022年第2期93-96,共4页JOURNAL OF WILDLAND FIRE SCIENCE

基　　金：福建省林业厅科技项目(闽林[2013]5号)。

摘　　要：毛竹内部在经历不同次数冻融后会发生不同程度损伤,严重的内部结构破坏会影响毛竹材性。利用105组毛竹傅里叶红光谱数据,通过BP神经网络建立了毛竹冻融次数预测模型,无损预测毛竹内部冻融损伤。将105组毛竹平均分为4组,每组依次通过0、1、2、3次冻融循环处理,采集每个毛竹的红外光谱,带入BP神经网络模型中进行训练,发现较优识别准确率可达100%。结果显示:利用傅里叶红外光谱结合BP神经网络,可有效预测毛竹冻融次数,能为毛竹材料选取提供可行方法。After being subjected to different times of freezing and thawing,the inside of bamboo will be damaged to different degrees,and serious internal structural damage will affect the wood properties of Phyllostachys edulis.Using 105 groups of Fourier red spectrum data o£Phyllostachys edulis,a prediction model of freeze­thaw times of Phyllostachys edulis was established through BP neural network to predict the internal freeze­thaw damage of Phyllostachys edulis non-destructively.Firstly,105 groups of bamboos are divided into 4 groups.Each group is processed through 0,1,2,and 3 freeze-thaw cycles.Then,the infrared spectrum of each bamboo is collected and brought into the BP neural network model for training,and the better is found The recognition accuracy can reach 100%.The results show:The freezing and thawing times of Phyllostachys edulis can be effectively predicted by using Fourier infrared spectroscopy combined with BP neural network,which can provide a feasible method for the selection of Phyllostachys edulis materials.

关 键 词：BP神经网络 识别 冻融循环处理 傅里叶红外光谱 毛竹 

分 类 号：S762[农业科学—森林保护学]