Application Of Near Infrared Spectroscopy Analyzing Technique For Quality Measurement Of Tomato Leaves

Because Near Infrared Spectroscopy (NIR) analysis technology had so many advantages, such as simpleness, rapidity, high accuracy, no pollution, good repeatability, and easy online analysis, it was widely applied in some areas with rapid development and had a good potential.However, at present, the NIR analysis was only mainly applied in petrochemical, grain and medical industry. To broaden the area of application of NIR analysis and find a rapid and nondestructive method, studies herein were carried out to obtain some chemical components that have a relationship with tomato growth information, are focused in this study: one is how to measure the content of water, chlorophyll in tomato leaves by standard chemical analysis technique and the other is to set up a predicted model of water, chlorophyll in tomato leaves by the NIR analysis technique and chemometrics.In this study, some information about content of water, chlorophyll in tomato leaves were obtained by standard chemical analysis technique. Diffuse reflection spectra were obtained by using high powered NEXUS spectrometer, Spectra acquisition and analysis methods, such as sample treatment, scan condition, spectra preprocessing, modeling method, sample elimination method and so on, were investigated in this study.For water content, the best modeling results were obtained by using original spectra based on optic fiber accessory and by using preprocessed spectra (first derivative+5 points smoothing +MSC) based on updrift accessory.The results are as follows:R=0.98083,RMSEC=0.0150, RMSEP=0.0147, f=9R=0.95600,RMSEC=0.0118, RMSEP=0.0273, f=7For chlorophyll content, the best modeling results were obtained by using spectra with Standard Normal Variate (SNV) correction based on optic fiber accessory and by using preprocessed spectra (second derivative+SNV) based on updrift accessory.The results are as follows:R=0.95098, RMSEC =0.335,RMSEP=0.595, f=9R=0.99020, RMSEC=0.567, RMSEP=0.366, f=8For SPAD, the best modeling results were obtained by using spectra with MSC correction based on optic fiber accessory and by using preprocessed spectra (first derivative+5 points smoothing +MSC) based on updrift accessory.The results are as follows:R=0.98043, RMSEC=4.89, RMSEP=5.36, f=10R=0.95468, RMSEC=5.39, RMSEP=4.05, f=10All models were established by using partial least square (PLS) regression method. The modeling results indicated that NIR technique was accurate enough to replace standard chemical analysis methods for rapid detection of water content, chlorophyll content and SPAD value in tomato leaves.