加工番茄早疫病病叶色素含量高光谱估测

Hyperspectral Estimation of Pigment Content of Process Tomato Leaves under Early Blight

早疫病对加工番茄的产量和品质有较为严重的影响。通过测定不同生育期自然发病的加工番茄早疫病病叶色素含量,把病片原始光谱反射率、一阶微分光谱反射率的光谱特征参数与相应病叶色素含量进行相关分析,采用线性和非线性单变量回归技术,建立色素含量估测模型并检验。研究结果表明:在传统的光谱特征参数和新建光谱特征参数中,R680、CCII、PSSR、PSNDc、FD664、DVI[FD686,FD664]为色素含量敏感光谱特征参数;R680、CCII、DVI[FD686,FD664]的线性单变量估测精度比较高;R680为变量的对数模型是估测Chl.a的最佳模型,DVI[FD686,FD664]为变量的对数模型是估测Chl.b和Chl.a+b的最佳模型。通过光谱响应特征对加工番茄早疫病病叶色素含量进行监测,能够为加工番茄早疫病监测预警和防治提供决策依据。

Process tomato early blight seriously affects the tomato yield and quality of the fruit.It is necessary to determine the leaves pigment content of process tomato early blight which is a natural disease at its different growing stages.Analyzing the correlation between the original spectral reflectance,first order differential spectral reflectance and corresponding pigment content,using the linear and nonlinear single variable regression technology,the pigment content estimation model is established and inspected.The results show that among the traditional and new spectrum feature parameters the R680、CCII、 PSSR、 PSNDc、FD664、 DVI are sensitive ones.The logarithm model of R680 for variable is the best model of estimation Chl.a,the logarithm model of DVIfor variable is the best model of estimation Chl.b and Chl.a＋b.The inspection R2 of model is 0.812、0.845、0.830 respectively.RMSE is less than 0.2.The results of the study show that it is feasible that hyperspectral technology estimates the pigment content of process tomato blight early and it is valuable that hyperspectral remote sensing can monitor the pigment content of crop diseases.