Studies On Passive Sensing Of Plant Chlorophyll Fluorescence And Application Of Stress Detection

The chlorophyll fluorescence (CF) signature emitted from vegetation provides abundance of information of photosynthesis activity, and has been used as a powerful tool to obtain physiological information of plant leaves in a non-invasive way. Remote sensing (RS) can provide Earth's surface information in large area and temporal period with different scales. So far RS has been widely applied in many fields in agriculture. The combination of CF and RS may take respective advantages. In this study, wheat (Triticum aestivum L.) was selected as the main material. First, the methods and algorithms were studied to separate CF from apparent reflectance spectra at leaf and canopy levels respectively. Then, CF from apparent reflectance spectra was analyzed with physiological and biochemical parameters such as CF parameters measured by modulated chlorophyll fluorometer, chlorophyll, etc. The main purpose was to study the chacteristics in fluorescence spectra at stress conditions (such as water and nitrogen deficiency, rust infected), as well as the influences on CF and other physiological and biochemical parameters made by stress factors. Finally, the potential of passive sensing of CF applied in crop stress detection was discussed. The results were summarized as below.1. Chlorophyll fluorescence spectra were successfully obtained from apparent reflectance spectra by special methods and algorithms. The leaves apparent spectra were measured under illuminations with and without specially designed filters, using a LI-COR 1800 integrating sphere coupled with an ASD spectrometer. Two kinds of filters (3 for long wave pass edge and 2 for short wave pass edge) were tested respectively. The results showed that fluorescence spectra could be derived from reflectance difference spectra under illuminations with and without long pass filter. Similarly, fluorescence spectra were derived from reflectance under illumination with short pass filters directly. Obvious differences in fluorescence spectra were founded in different kinds of plant. Additionally, CF was also detected from the transmittance spectrum and the relative scale was approximately close with that of the reflectance. The results confirmed the conclusion that CF was contributed to apparent reflectance to some extent and it could be separated quantificationally.2. The CF indices that closely related to apparent reflectance spectra at canopy scale were preliminary illustrated. The solar-induced CF can reflect real plant physiological status. Diurnal change experiments were conducted to measure canopy reflectance in wheat and Parthenocissus tricuspidata respectively, at the same time CF parameters were measured. The results indicated that CUR (curvature index), Dmax/D702 (the ratio of the first derivative at two bands) and PRI (physiological reflectance index) had the same tendency with Fv/Fm, which showed a characteristic of 'V type. Then, a method based on FLD (Fraunhofer line discriminator) was introduced to calculate solar induced fluorescence according to the molecular oxygen absorption by the terrestrial atmosphere at 688 and 760 nm. CF derived from FLD was sensitive to PAR, and had a notably negative relation with fluorescence kinetic parameter Fv/Fm (R~2>0.9). The paper provide a new method to detect vegetation fluorescence signals by radiance mode using field spectrometer, which extend the application scales of the instruments.3. The characteristics of fluorescence spectra at leaf or canopy levels exhibited differencesunder different N or water treatments. The ratio of the reflectance difference at 685nm and 740nm Dif685/Dif740 at leaf level was linear correlated with Fv/Fm and it could be an indicator to reflect leaf water or nitrogen content. At the canopy level PRI, Dmax/D702, FLD fluorescence present high correlationships with plant water content or canopy nitrogen density (CND). These indices have the potential to diagnose the water or nitrogen status in the fields, but the influence factors especially canopy structure should not be neglected.4. A preliminary study on CF for detecting stripe rust in wheat was conducted. The results showed that physiological parameters SPAD, photosynthetic rate decreased rapidly as leaf severity increases. CF parameters measured by modulated fluorometer showed that the maximal photochemical efficiency Fv/Fm and yield of quantum efficiency decreased gradually, whereas non-phochemical quenching co-effficient qN increased. CF intensity calculated from leaf apparent reflectance had an increase tendency at 685nm while decreased at 740nm. At the canopy level, CF at 688 nm was positively correlated with the disease index (DI) whereas CF at 760 nm was negatively correlated with DI. Ratios of CF at 688 to that at 760 nm, which found to be a suitable indicator of stress events in plants, were positively related to DI.5. The applicability of passive sensing of CF to the detection of crop stress was accessed. Spectral data and agricultural sampling data were not well linked together. On the basis of research on CF characteristics at leaf level, it is necessary to develop a model to simulate CF from leaf to canopy level, taking account of the influence of viewing geometry, atmospheric characteristics and the environmental factors. Additionally, CF technique should integrate with reflectance spectrum and crop cultivation characteristics in order to detect crop condition timely, rapidly and accurately, and to realize precision management under optimal regulation and control.