Determination of lettuce nitrogen content using spectroscopy with efficient wavelength selection and extreme learning machine

Abstract

Measuring the nitrogen content of plants is useful for nitrogen fertilizer management. The aim of this work is to explore the use of spectroscopy for estimating lettuce leaf nitrogen content. Leaf reflectance spectra were measured using a spectroradiometer with a range of 350-2500 nm, and 160 fresh lettuce leaves given five different nitrogen treatments were used for spectra acquisition and total nitrogen determination. Interval partial least squares (iPLS), synergy interval partial least squares (siPLS), and siPLS combined with genetic algorithms (GA-siPLS) were used to select the optimal intervals or variables. Partial least squares (PLS) and extreme learning machine (ELM) methods were used to develop calibration models. One characteristic wavelength region (470-590 nm) using 120 variables was selected by iPLS; four characteristic wavelength regions (440-530, 530-620, 620-710 and 890-980 nm) using 362 variables were selected by siPLS, and 56 wavelength variables were selected by GA-siPLS. Six different regression models were established for nitrogen content by PLS and ELM based on optimal intervals or variables. The results imply that GA-siPLS is an efficient variable selection algorithm and ELM is a successful nonlinear regression tool. Furthermore, GA-siPLS combined with ELM is a feasible method for measuring nitrogen content in lettuce, as it performed better than other models. The optimal results achieved a root mean square error of prediction (RMSEP) = 0.2890% and correlation coefficient Rp = 0.9218.

Application of Hyperspectral Imaging to Detect Sclerotinia sclerotiorum on Oilseed Rape Stems

“…Selecting optimal wavelengths with great universality and repeatability among different samples was essential for these purposes. According to previous studies, optimal wavelengths selected by different methods were different [ 45 , 46 ]. However, some of the optimal wavelength selection methods were based on performances of discriminant models.…”

Rapid detection of talcum powder in tea using FT-IR spectroscopy coupled with chemometrics

“…In recent years, techniques for building nonlinear models coupled with spectroscopic techniques have also been successfully applied in many areas 32 33 34 . Extreme learning machine (ELM) is one of learning neural algorithms, which has been successfully applied in nonlinear regression problems 35 36 . Comparing with traditional learning algorithms, ELM not only reaches the smallest training error but also the least amount of output 37 .…”

“…During modeling, full cross-validation was used to validate the quality of the models and to prevent over-fitting of the calibration. And the performance of models was evaluated by correlation and root-mean square error of calibration, cross validation and prediction (R C , R CV , R P , RMSEC, RMSECV, RMSEP) 38 39 .…”

Hyperspectral assessment of plant responses to multi‐stress environments: Prospects for managing protected agrosystems

“…For example, Woodhouse, Heeb, Berry, Hoshizaki, and Wood () showed the potential of this technique for monitoring health conditions of excised leaves of lettuce grown hydroponically under different stress conditions (i.e., copper, zinc, nitrogen, phosphorus, potassium, drought), applied singularly. More recently, studies have focused on the determination of nitrogen content of lettuce leaves and canopy by spectra using multivariate modeling methods (Gao, Mao, & Zhang, ; Itoh et al, ; Mao, Gao, Zhang, & Kumi, ; Sun et al, ). A similar approach has been used to predict pigment (chlorophyll, carotenoid, anthocyanin) content in lettuce based on VIS‐NIR spectroscopy (Neto et al, ), as well as using spectral indices (Gazula, Kleinhenz, Scheerens, & Ling, ; Lopes et al, ; Xue & Yang, ).…”