Vegetation is a sensitive indicator suitable for testing of ecological stresses and natural anomalies of the technogenic character. First, it is determined by the prompt response of photosynthetic apparatus to changes of environmental conditions, mainly by change of green pigment (chlorophyll) content in leaves. Second, the specific kind of a reflectance spectrum of leaves is due to chlorophyll presence in them, and the area in the range of 500-80 nm is extremely sensitive to variations of its pigment content. Thirdly, there are interesting results now concerning spectral properties of leaves and crops canopies obtaining with high-resolution spectroscopy. The data are high informative in relation to content of chlorophyll and some other biochemical constituents of a cell. The high resistance to various types of noises is inherent to methods developed on the basis of such spectral data. We have developed a method for chlorophyll estimation using the 1-st derivative plots of reflectance spectral curves. The method gives good results for plant-soil systems with both for 100% and
incomplete projective covering as our simulation models show. Field measurements of chlorophyll content in closed and open canopies crops confirm the results. A hardware-software complex has been produced by us for chlorophyll determining under field conditions. It consists of spectral and computing blocks. First of them is a two-beam spectrometer of high resolution supplied by a system to visualize of measured object. The irradiance and temperature sensors are included to the spectral block as well as GPS-receiver. The following technical characteristics are inherent to the block: spectral range 500-800 nm, band-pass 1.5 nm, field of view 16x16o, scanning time 0.1-1.0 s, dynamic range of signal 1:1024 (10 bit), signal/noise ratio 400, amount of pixels in image 1240, range of estimated chlorophyll concentrations 1.5-8.0 mg/dm2, supply voltage 12 V, weight 8 kg. Computing block is intended for spectral date processing to obtain chlorophyll estimations using our algorithm. The block is supplied by our original software WINCHL, which includes spectrum and algorithm
libraries and various mathematical tools. Accumulation of reflectance spectra of various plants together with data of environmental conditions at measurements gives a good possibility to use all of them for future scientific researches and developing other important parameters of canopy status.
It is shown that the spectral curve of reflectance of vegetation
contains the sufficient information to create a set of parameters
for effective monitoring of agricultural crops. Most of them are
based on the chlorophyll estimation or characteristics, which are
dependent on specific influence of inner structure of plant
tissues on leaf reflection in the region of chlorophyll
absorption. New chlorophyll indices are proposed for estimation of
chlorophyll content in leaves using the shape of leaf reflectance
curves. The ratio of two maxima in the 1-st derivative plot from
reflectance spectral curve in 680-750 nm region has been shown
to correlate with chlorophyll content in winter wheat leaves.
Independent component analysis of reflectance spectral curves has
been applied as well. An interrelation between the chlorophyll
concentration and vectors of principal components has been found.
The estimates of the chlorophyll content by using of these
parameters and regression equations gave suitable results.
Comparison of two approaches has been performed. Stability of both
approaches with regard to incomplete project covering have been
tested. Usage of physical and graphical models permits to estimate
stability in calculation results of chlorophyll concentration
influence of soil reflection. It has been shown that the ratio of
two maxima in the 1-st derivative plot was changed now more than
5% and 11% under 50 % and 25 % projective cover,
respectively, on a background of dark soil or sand. The
reflectance coefficient at 550 nm correlates with chlorophyll
content but it is highly sensitive to contribution of soil
reflectance. Therefore combination of chlorophyll estimates
obtained by red edge parameters and the reflectance coefficient at
550 nm gives possibility to estimate a projective covering. We
shown that principal components approach is resistant to
influence of project covering.
The problem of remote estimation of chlorophyll content in vegetation is considered. A lot of reflectance spectra have been recorded for winter wheat leaves with various chlorophyll content. The plots of the 1-st derivative of reflectance spectral curves have been computed and analyzed in respect interrelation with pigment content. The ratio of two maxima in the plots has been revealed as a correlating characteristic, which was used for chlorophyll estimation. To diminish the level of noise in 1-st derivative plots, producing by measuring system, the computing procedure have been applied by Savitzky and Golay formula using 2-d order polynomial estimation of 9-point convolution. Application of genetic algorithm to search of maximum positions in 1-st derivative plots has been tested with a positive result. Pair and multiple regression as well as neural net approach have been tested for estimation of chlorophyll content.
Using our experience in signal processing and optimization of complex systems we propose a new method to adaptive sensing of chemical content of vegetations. This framework is demonstrated for different agricultural plants using the neural network algorithm for classification of spectral curves and adaptive filtration. Utilization of characteristics of leaf reflectance spectrum, which are a relative characteristic of the light reflected from canopies, makes it possible to avoid the necessity of measuring the 100% reflectance standard and to provide the high resistance of the method to distorting factors in particular to soil reflectance contribution. For utilization of the method the numerical algorithms is proposed. Various estimation problems will be considered to illustrate the computational aspects of the proposed method. The software is based on digital filter, optimization approach and neural network algorithm for classification of chemical components. Supporting software for data management, storage, signal processing will be development. A concept of an intelligent sensor is considered.
In this paper we consider the problem of estimating chlorophyll content in vegetation using an experimental optical method from noisy spectral data. It is shown that the quantitative analysis of the spectral curves for the reflection of plant leaves may be the basis for development of new methods for interpretation of the data obtained by the remote measurement of plants. A mathematical model of vegetation reflectance is proposed to estimate the chlorophyll content from spectral data. Estimates are defined as minimizers of penalized cost functionals with sequential quadratic programming (SQR) methods. An estimation is related to the local scoring procedure for the generalized additive model. A deviation measurement in risk analysis of vegetation is considered. The role of deviation and risk measures in optimization is analyzed. Experimental and simulation results are shown for different agricultural plants using a functional-parametric representation of spectral curves.
In this paper we consider the problem of estimating chlorophyll content in vegetation using an experimental optical method from noisy spectral data. It is shown that the quantitative analysis of the spectral curves for the reflection of plant leaves may be the basis for development of new methods for interpretation of the data obtained by the remote measurement of plants. A mathematical model of vegetation reflectance is proposed to estimate the chlorophyll concentration from spectral data. Estimates are defined as minimizers of penalized cost functionals with sequential quadratic programming (SQP) methods. An estimation tool is related to the local scoring procedure for an generalized additive model. Experimental and simulation results are shown for different agricultural plants using a functional parametric fitting of spectral curves.