摘要
利用便携式光谱分析仪测量生菜叶片的光谱反射率,并对其进行对数变换。通过变量筛选得到725、1 075、1 272、1 450、1 640和1 958 nm波长处的光谱反射率与生菜干基含水率呈极显著相关。为克服多重共线性影响,分别采用多元线性回归分析、主成分回归分析、偏最小二乘回归分析及偏最小二乘-人工神经网络回归分析4种方法建立了叶片干基含水率的定量分析模型。结果表明4种算法预测值与实测值相关系数分别为0.485 0、0.899 2、0.917 4和0.947 0,偏最小二乘-人工神经网络模型的预测能力优于其他模型。
Spectral reflectance of lettuce leaves in growing status was measured using the ASD FieldSpec3,and logarithmic transformation was also obtained.By variable selection,it was found that the linear relationships between dry-basis moisture content of lettuce leaves and spectral reflectance data in 725 nm,1 075 nm,1 272 nm,1 450 nm,1 640 nm and 1 958 nm were very notable.In order to overcome the impact of multicollinearity,quantitative analysis models of dry-blade＇s moisture content have been established respectively with methods of multiple linear regression analysis,principal component regression analysis,partial least squares regression analysis and PLS-neural network analysis.The result showed that the correlation coefficient R of measured and predictive values from the four algorithms were 0.485 0,0.899 2,0.917 4 and 0.947 0 respectively,which showed better predictive performance of the model based on PLS-neural network analysis than the others.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2011年第5期166-170,共5页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家高技术研究发展计划(863计划)资助项目(2008AA10Z204)
国家自然科学基金资助项目(61075036)
江苏省农业装备与智能化高技术研究重点实验室项目(BM2009703)
江苏省"333人才工程"科研项目
关键词
水分胁迫
生菜
叶片
含水率
光谱特征
数学模型
Water stress
Lettuce
Leaves
Water content
Spectral characteristics
Mathematical model
