Predicting Chlorophyll Content Of Greenhouse Tomato With Ground-Based Remote Sensing
Ding Yong-Jun,Li Min-Zan,Li Shu-Qiang,An Deng-Kui
DOI: https://doi.org/10.1117/12.866205
2010-01-01
Abstract:NIR spectroscopy can be used in analysis of plant chlorophyll content on a large scale area. This offers the opportunity to use spectral reflectance as a non-destructive method for analyzing photosynthetic pigment status in plant. This research studied the variation of the chlorophyll content and spectral response at different growth stages of greenhouse tomato. Leaf spectral measurements from each treatment (4 N-levels: 0%, 33.3%, 66.6%, 100%) were taken in the greenhouse using an ASD FieldSpec HH spectrophotometer. Chlorophyll content of tomato leaves were measured by alcoholic-acetone extraction in lab. It was found that chlorophyll content of tomato leaf was increasing continuously to the maximum 50 days after the transplantation, while red edge moved to the NIR bands (long wave), and green peak position moved to the Blue bands (short wave) and green peak amplitude decreased. The chlorophyll content would decrease after fruiting stage, while red edge, green peak position and amplitude moved to the opposite direction. Regarding quantitative analysis the relationship between chlorophyll content and spectral response, red edge parameters (S-red(area of red edge), D-red (amplitude of red edge) and Pred (position of red edge)) in the first derivative of reflectance curve were obtained at bands of 680-760 nm. Similarly, blue edge, green peak and red valley parameters were defined to reflect spectral character. Vegetation indices were used extensively to estimate the vegetation growth status. Thus, the following wavelengths were used for developing RVI, NDVI and ARVI indices: lambda 440nm, lambda 500nm, lambda 550nm, lambda 680nm, lambda 770nm, P-blue(position of blue edge), P-yellow(position of yellow edge), P-red(position of red edge), P-greenpeak(position of green peak), P-redvalley(position of red valley). Seven optimal spectral parameters were chosen with the method of Karhunen-Loeve from the above-mentioned 68 self-defined property parameters. Stepwise multiple regression (SMLR), principal component regression (PCR), ridge regression (RR) and partial least squares regression (PLSR) were used to develop the prediction models of the chlorophyll content of tomato leaf. The best model was obtained by RR. Root MSE was 0.406 and R2 was 0.839.