植物叶绿素含量的非接触式光谱检测有着测定范围大、易于实现自动检测等优点,对实施精准农业、搭建培养室内智能监控系统具有重要的意义。但叶片结构、叶片粗糙表面引起的光散射效应、环境光等干扰因素使其定量分析难度增加,本文采用一种结合等距映射(ISO-MAP)和偏最小二乘(PLS)的非线性降维建模方法,有效地提高了模型的预测精度。与波长参比和单纯PLS两种传统建模方法相比,ISOMAP—PLS模型的预测精度分别提高了56.3%和34.4%。实验结果表明,ISOMAP算法可以降低建模复杂度的同时提高预测精度,将其应用在非接触式叶绿素含量检测中是可行的,为实现植物多种生化参数的非接触检测以及相应仪器的开发奠定了基础。
Chlorophyll content noncontact spectral measurement has advantages of large measurement range,easy realization of automatic detection ,etc. It is of great significance for implementing precision agriculture and building the intelligent monitoring system in the cultivate room. However, interference factors such as leaf structures, light scattering effect caused by leaf rough surface,and environment light will increase the difficulty of quantitative analysis. In this paper, a nonlinear modeling method which combines isometric mapping(ISOMAP) with partial least squares(PLS) is applied to noncontact measurement of plant chlorophyll content. It effectively improves the predictive accuracy. Compared with Wavelength Reference and Single PLS modeling methods, the predictive accuracy of ISOMAPPLS model increases by 56.3% ,34.4% respectively. The experiment results indicate that the ISOMAP algorithm can reduce modeling com plexity and improve predictive accuracy at the same time. It is concluded that it is feasible to apply ISOMAP algorithm to chlorophyll content noncontact measurement. It lays the foundation for realizing noncontact measurement for varie ty of plant biochemical parameters, and for the development of the corresponding instruments in future.