ArticlePDF Available

APPLICATION OF BIOSENSORS FOR PLANTS MONITORING

Authors:

Abstract and Figures

Current methods of diagnostics of plant state need to conduct expensive and long-time physical-chemical and microbial analyses of soil and plant samples. The chlorophyll fluorescence induction method allows determining the functional state of plant in express mode without plant damage and it gives an opportunity to estimate the influence of stress factors on the plant state. In recent decades a number of researches of the chlorophyll fluorescence induction were significantly increased because of appearance of relatively inexpensive portable fluorometers. This paper represents results of testing biosensors developed at the V.M. Glushkov institute of Cybernetics of NAS of Ukraine on base of chlorophyll fluorescence induction method. It were developed appropriate software to facilitate data acquisition and processing. Analysis of some experimental results by means of neural networks is discussed.
Content may be subject to copyright.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
115
APPLICATION OF BIOSENSORS FOR PLANTS MONITORING
Oleksandr Palagin, Volodymyr Grusha, Hanna Antonova,
Oleksandra Kovyrova, Vasyl Lavrentyev
Abstract: Current methods of diagnostics of plant state need to conduct expensive and long-time
physical-chemical and microbial analyses of soil and plant samples. The chlorophyll fluorescence
induction method allows determining the functional state of plant in express mode without plant damage
and it gives an opportunity to estimate the influence of stress factors on the plant state. In recent
decades a number of researches of the chlorophyll fluorescence induction were significantly increased
because of appearance of relatively inexpensive portable fluorometers. This paper represents results of
testing biosensors developed at the V.M. Glushkov institute of Cybernetics of NAS of Ukraine on base
of chlorophyll fluorescence induction method. It were developed appropriate software to facilitate data
acquisition and processing. Analysis of some experimental results by means of neural networks is
discussed.
Keywords: fluorometer, biosensor, wireless sensor network, chlorophyll fluorescence induction, neural
network, information technology.
ACM Classification Keywords: H.4 Information system application
Introduction
Over the past decade portable devices of "Floratest" family were developed and manufactured in
V.M. Glushkov Institute of Cybernetics of NAS of Ukraine. The researchers of chlorophyll fluorescence
induction (CFI) effect encounter the problem to gain sufficient amount of data by means of autonomous
fluorometers. Besides, the time of a measurement of chlorophyll fluorescence induction varies from
several minutes to one hour, depending on environmental conditions, species of plants and experiment
specificity. The temperature and humidity of air and soil, illuminance can vary, that can influence on
reliability of measuring data. All this has to be taken into account during ecological and agro-ecological
monitoring. So, to overcome above-mentioned disadvantages, it was designed wireless biosensors that
are combined in wireless sensor network together with special network coordinators, and concentrator
[Palagin at al., 2017]. The biosensors were tested in laboratory and field conditions. This paper
represents some important results of that testing and data analysis.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
116
Work objectives
Work objectives are testing of developed biosensors and developing database, software and methods to
facilitate acquisition and processing of measured data.
Measurement of CFI and its parameters
The technique of laboratory or field experiment includes next:
1. Selecting plants. Planning and choosing testing plants. Goal of an experiment has to be taken into
consideration when experiment is planned and plants are selected. A chosen plant-indicator has to be
sensitive to stress factor [Guo and Tan, 2015].
2. Plants are grown in identical conditions in pots or on field with identical soil.
3. The grown plants are divided into few groups – control and experimental.
4. Experimental plants are put on influence of stressful factors of different degree in accordance with
testing program.
5. Network of biosensors measures chlorophyll fluorescence induction (CFI) of control and experimental
plants in accordance with testing program for type of stress and its degree in scheduled terms.
The using of few fluorometers or the developed network of wireless biosensors allows reducing the time
needed for measurements and it can provide data that are more adequate. The time can be calculated
according to formula:
s
N
iprmad
eN
ttt
t
)( ,
where e
t is a time to get experimental data; N is an amount of measurements; ad
t is a time of dark
adaptation of leaf; pr
t is a time needed to prepare the next measurement; m
t is a time of measurement
of chlorophyll fluorescence induction curve, s
Nis a number of sensors.
If the sensors are placed on a leaf under sunlight then the dark adaptation has to be not less than 20
minutes. If the plant (or its leaf) is placed during long period in a shadow then 5 minutes is enough for
the dark adaptation.
6. Measuring data of chlorophyll fluorescence induction, acquires by biosensors from control and
experimental plants.
7. It is useful to record the air and soil temperature and humidity during a measurement of chlorophyll
fluorescence induction. In addition, chemical and biological analysis of soil can be used for specific
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
117
biological researches. It is allows to take into consideration climatic effect as additional stress factor on
parameters of chlorophyll fluorescence induction.
8. The results of measurements are processed by means of graphical, statistical and correlation
analysis and machine learning technique. Before analysis, the measured data can be normalized.
9. The finish result of testing is detecting the sensibility of biosensors to influence of different stresses.
Typical curve of chlorophyll fluorescence induction is shown on figure 1. For analysis of measured
curves the researchers typically analyze special parameters of CFI curves such as: Fo (initial level of
chlorophyll fluorescence); Fm (maximum level of chlorophyll fluorescence); Fst (stationary level of
chlorophyll fluorescence); Fv = Fm – Fo (variable fluorescence); Fv/Fm ratio, Area (the area above the
fluorescence curve between Fo and Fm), Fj (fluorescence value at point J, t 2 ms); Fi (fluorescence
value at I, t 30 ms) and so on. Also the machine learning method is getting popular recent years
[Kalaji at al, 2017].
Figure 1. The typical curve of chlorophyll fluorescence induction
Development of software and database for work with chlorophyll fluorescence induction curves
Several activities have to be repeated during processing chlorophyll fluorescence induction curves (CFI)
by means of personal computer: opening file with measuring results, graph building for previous visual
estimation of dynamics of CFI curve, grouping different measurements, calculation of curves parameters
and so on. It gets a lot of time. The special software FAnalyzer was developed to simplify the processing
of chlorophyll fluorescence induction curves.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
118
Functions of the developed software are the following:
1) Receiving measuring data from biosensors and further data output in form of a graph.
2) Storing the received measuring data on hard disk and opening in form of graphs;
3) Opening and storing several curves of CFI in one file. The file can be opened later and
processed by means of program packages such as R, Excel, Matlab and so on.
4) Calculation and storing CFI curves parameters, that are frequently used to analyze the
measuring results (F
m
, F
o
, F
st
, Rfd, Area, F
i
, F
j
and other), and main statistical indicators for that
parameters.
The graphical user interface of program is shown on Figure 2.
Figure 2. Graphical user interface
The suggested software allows to reduce time for preparing data to analysis, to calculate main
parameters of CFI and proper statistical indicators. The calculation can be used for comparative
analysis of plant states in conditions of influence of stress factors and in normal conditions.
During using multiple biosensors simultaneously it is necessary to store, process and visualize a large
amount of measuring data. For convenience of users, the database and proper graphical user interface
were developed. They allow storing a large amount of measurements in one place for further data
analysis of measuring data by means of tools and methods, selected by user. During the database
development, a set of entities was defined to represent in the database. The last ones contain
information about: plant information; type of monitoring of plant state; measured curve of chlorophyll
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
119
fluorescence induction; information about soil, air and parts of plant (in case of chemical-biological
analysis);information about devices and sensors, used for measurements; weather
information; information about a person, conducting measurements; information about an organization
and a location, where measurement was conducted.
A database management system MySQL was used for database implementation. The database
diagram is shown on Figure 3.
Figure 3. The database diagram
Importing data to the database can be carried by means of special software.
Research of change of chlorophyll fluorescence induction under influence of copper
To research the influence of heavy metals on plants it is reasonable to select the goose-foot plant.
Goose-foot has a wide natural habitat, grows in a different environmental conditions. It was studied the
influence of different doses of toxicant, copper sulphate (CuSO
4
), on the test plants Plants were
cultivated in 12 pots, three-four plants per pot. The plants were divided into 4 groups. Different
concentration of CuSO
4
were dissolved in water and brought into the soil of these four groups.
Group 1 (V1) – control group without CuSO
4
.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
120
Group 2 (V2) – 1 g of CuSO4 / 1 kg of soil.
Group 3 (V3) – 3 g of CuSO4 / 1 kg of soil.
Group 4 (V4) – 6 g of CuSO4 / 1 kg of soil.
The experiment was conducted during 13 days. At the beginning of the experiment the chlorophyll
fluorescence induction was measured in all groups of plants (Figure 4). The same day the water solution
of CuSO4 was brought into soil of test plants.
Figure 4. The intensity of chlorophyll fluorescence of goose-foot plant before of toxicant bringing in
The chlorophyll fluorescence intensity of test plants changed under the influence of toxicant. Figures 5
and 6 show graphs of the chlorophyll fluorescence on the second and third days of the impact of copper
sulphate. It can be easily seen, that on the third day the maximum level of the chlorophyll fluorescence
induction parameters (Fst, Fm) considerably decreased for plants that had been treated by toxicant.
Figure 5. The intensity of chlorophyll fluorescence of goose-foot plant on the second day of toxicant
influence
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
121
Figure 6. The intensity of chlorophyll fluorescence of goose-foot plant on the third day of toxicant
influence
It should be noted, that on the sixth day of toxicant influence only one group of plants V2 remained, two
other groups of plants V3 and V4 perished.
Analysis of parameter Fv/Fm provides information about the photochemical reactions, which are most
sensitive to environmental factors. The maximum difference of the parameter Fv/Fm between the control
group and the group V4, which received the maximum dose of copper sulphate, equals 38 %. At the
beginning of the experiment this parameter had almost the same value in the three groups V1, V2, V3,
V4 – 0,906 on the average. In the control group parameter Fv/Fm decreased by 5,8 % in comparison
with the first day of measurement. In the group of plants V2 on the fifth day of toxicant influence the
parameter Fv/Fm decreased by 5 % and the overall decrease equaled 4,6 % in comparison with the first
day of experiment. The value of parameter in the group V3 decreased on the six day of the influence of
copper sulphate. In the group V4 this parameter decreased by 39 % on the third day of the influence of
toxicant. Figure 7 shows changes of the parameter during experiment.
Analysis of results shows, that different doses of copper sulphate influenced on the photosynthetic
apparatus of plants in different ways. Thus, the dose of 6 grams of copper sulphate was critical for plant
of group V4. Also, the dose of 3 g of copper sulphate is critical for plants of group V3 and causes
irreversible changes in the plants. Photosynthetic apparatus of plants, treated by 3 g of CuSO4, stops to
function on eighth day of toxicant influence. The dose of 6 grams breaks the photosynthetic processes
in plants on the third day. However, it should be noted, that the dose of 1 gram of CuSO4 does not
cause any serious changes in plants and also does not break the photosynthesis of plants.
For developing methodical support for wireless biosensors the experiment was conducted to research
the influence of heavy metals on plants. It allowed to estimate the dose of copper sulphate, that is
critical for plants, and to determine informative parameters of chlorophyll fluorescence induction curves.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
122
During application of industrial methods the obtained results will be used to detect the presence of
heavy metals in plants and estimate their impact on ecological state of certain territories.
Figure 7. Changes of the parameter Fv/Fm during the testing
Using neural networks for determination of plants under stress
Nowadays neural networks and widely used for the analysis of biological and agricultural data in viral
diseases of plants, pest determination, water consumption estimation, plant quality estimation etc.
[Samborska at al.].
Researches of influence of herbicide on chlorophyll fluorescence were conducted at the V.M. Glushkov
Institute of Cybernetics and the enough amounts of data were gained for using neural networks.
Herbicide Roundup (glyphosate) was used for experiments. Roundup is a broad-spectrum systemic
herbicide. The plants of Datura stramonium (weed) were divided into three groups. One, control group
was not treated and two others were sprayed with different doses of herbicide.
Two-layer feed-forward network was chosen for classification of curves. Neural network has 89 inputs
and 3 outputs (every measured curve consist of 89 points). Second, output layer consist of three
neurons (three variants of curves). The required number of neurons of hidden layer was determined by
conducting series of experiments. The performance (P) of the training was evaluated using means
square error.
There were trained neural networks with different number of hidden layer neurons (from 1 to 364). The
training of every network was repeated 30 times and the results were averaged and combined in vector
Pmean (Figure 8). Thus, the neural network works most efficiently with not more than 70 neurons in the
hidden layer. A neural network with 25 neurons in the hidden layer was chosen for further use. The
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
123
network uses the sigmoid transfer function for hidden neurons and the softmax function for output
neurons.
Figure 8. Dependence of the mean square error of the neural network training on the number of
neurons
The neural network was trained on data measured in different days. The data measured in different
days were used separately for training of the network. The results of the training are presented in
Table 1.
As seen from Table 1, the smallest errors of recognition were received with data in 7 and 11 days. It is
known that Roundup breaks the synthesis of the amino acids on 5-6 day and plants wade and discolor
after two weeks. But after two weeks the curves of chlorophyll fluorescence of the treated leaves had
serious difference even on one plant, therefore the neural network recognition is unsatisfactory. On the
contrary the Student’s test confirmed the difference between curves of plants of different groups at the
end of second week.
Thus we showed that neural networks can be trained for stress recognition of plants using curves
measured by sensors developed at V.M. Glushkov Institute of Cybernetics of NAS of Ukraine.
It is useful to use neural network for creating methods for evaluation of the state of plants in the city and
the farm. It can be used at the stage of making decision (start watering, give fertilizer, etc.). Neural
network training needs a representative set of data to make valid managerial decision, so the wireless
biosensor networks allow to receive enough number of fluorescence induction curves.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
124
Table 1. Results of the neural network training using data measured in different days, where E is an error of
training, Ev is an error of validation, Et is an error of testing, Em is a mean calculated from three
previous errors.
The number of curves E, % Ev,% Et,% Em,% Notes
40 64,3 66,7 33,3 60,0 Before treatment of the herbicide
43 16,1 33,3 66,7 25,6 Before treatment of the herbicide
41 0 33,3 50,0 12,2 The third day after treatment
43 80,6 66,7 83,3 79,1 The fifth day
43 0 0 33.3 4.7 The seventh day
30 0 0 20 3.2 The eleventh
43 19,4 16,7 66,7 25,7 The thirteen
21 3,2 0 66,7 11,6 The twentieth
Using neural networks for determination of plant species
CFI curves of different plant species have some significant difference, thus they can be used for
determination of specie of plant that are shown in [Kirova at al, 2009] by means of OJIP curve (CFI
curve received during nearly 10 seconds) and neural network. With aim of testing the developed
sensors for this task, a set of plants was measured. The set includes 176 curves from 6 species. The
curves were measured during 5 minutes (full curve of CFI) and 10 seconds (OJIP curve) for next plants:
soybean, goosefoot, ficus elastic, ficus benjamina, euphorbia, and zinnia.
Two-layer feed-forward network with 89 inputs and 25 neurons in the hidden layer was chosen. The
network uses the sigmoid transfer function for hidden neurons and the softmax function for output
neurons as in previous experiment. The output layer consists of 6 neurons.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
125
The results of testing of the neural network present in Table 2. The neural network was trained 100
times and errors of testing were averaged after.
Table 2. The results of determination of plant species
Duration of measurement of CFI 5 minutes 10 seconds
minimal testing error, % 0 0
mean testing error,% 6,52 9,80
So, the curves of developed sensors can be used for taxonomic determination of plants. The curves
measured during 5 minutes are more appropriate for this task. There are raised the issue of
determination of plants with large amount of curves of very close species. The approach to solve it is
described in [Kirova at al., 2009].
Conclusion
It was conducted the series of experiments for the testing biosensors developed at the V.M Glushkov
Institute of Cybernetics of NAS of Ukraine to determine the sensitivity of biosensors to influence of
stressful factors of different nature on experimental plants. The suitable software and database were
developed to facilitate data processing. As result of using neural network, it can be concluded that
neural network can recognize the different dose of fertilizer before changing of leaves appears and a 5
minutes measurement of CFI is more informative for determination of plant species then 10 seconds
measurement.
Bibliography
[Guo and Tan, 2015] Y. Guo, J. Tan. Recent advances in the application of chlorophyll fluorescence
from photosystem II. Photochemistry and Photobiology Vol. 91, Issue 1, Wiley, 2015. pp. 1-15
[Kalaji at al., 2017] H.M. Kalaji, G. Schansker, M. Brestic at al. Frequently asked question about
chlorophyll fluorescence, the sequel. Photosynthesis Research. Vol. 132, Issue 1, Springer, 2017. pp
13-66.
[Kirova at al., 2009] M. Kirova, G, Ceppi, P. Chernev, V. Goltsev, R. Strasser Using artificial neural
networks for plant taxonomic determination based on chlorophyll fluorescence induction curves.
International Journal “Information Theories and Applications”, Vol. 24, Number 2, © 2017
126
Biotechnology & Biotechnological Equipment. Vol. 23, Issue sup1: XI Anniversary scientific
conference, 2009 pp. 941-945. ISSN: 1310-2818 http://dx.doi.org/10.1080/13102818.2009.10818577
[Palagin at al., 2017] O. Palagin, V. Romanov, I. Galelyuka, O. Voronenko, Y. Brayko, R. Imamutdinova.
Wireless sensor network for precision farming and environmental protection. I.Tech 2017
[Samborska at al., 2014] I. Samborska, V. Alexandrov, L Sieszko, B. Kornatowska, V. Goltsev, M.D.
Cetner, H.M. Kalaji. Artificial neural networks and their application in biological and agricultural
research. Signpost Open Access Journal NanoPhotoBioSciences, Vol. 02, 2014. pp. 14-30. ISSN:
2347-7342 http://signpostejournals.com/ejournals/portals/12/v22.pdf
Authors' Information
Oleksandr Palagin – Depute-director of V.M.Glushkov Institute of Cybernetics of
National Academy of Sciences of Ukraine, Academician of National Academy of
Sciences of Ukraine, Doctor of technical sciences, professor; Prospect Akademika
Glushkova 40, Kiev, 03187, Ukraine; e-mail: palagin_a@ukr.net
Volodymyr Grusha – research fellow of V.M.Glushkov Institute of Cybernetics of
National Academy of Sciences of Ukraine; Prospect Akademika Glushkova 40, Kiev,
03187, Ukraine; e-mail: vhrusha@gmai.com; website: http://www.dasd.com.ua
Oleksandra Kovyrova– research fellow of V.M. Glushkov Institute of Cybernetics of
National Academy of Sciences of Ukraine; Prospect Akademika Glushkova 40, Kiev,
03187, Ukraine; e-mail: kovyrova.oleksandra@gmail.com; website:
http://www.dasd.com.ua
Antonova Hanna – engineer of V.M. Glushkov Institute of Cybernetics of National
Academy of Sciences of Ukraine; Prospect Akademika Glushkova 40, Kiev, 03187,
Ukraine; e-mail: annat7806@gmail.com; website: http://www.dasd.com.ua
Vasyl Lavrentyev – research fellow of V.M. Glushkov Institute of Cybernetics of
National Academy of Sciences of Ukraine; Prospect Akademika Glushkova 40, Kiev,
03187, Ukraine; e-mail: vaslavr@i.ua; website: http://www.dasd.com.ua
... The term "biological detector" is applied to animals (Peñaranda & Asemsio, 2008) and plants (Volkov & Ranatunga, 2006;Palagin, Grusha, Antonova, Kovyrova & Lavrentyev, 2017) that can be trained, conditioned, or genetically modified to detect key molecules in the environment. Detecting "target odors" plays a key role for a variety of purposes within the forensic setting (Castillo, 2016). ...
Book
Full-text available
This book is part of a broad investigation, where the authors declare that they have no conflicts of interest; in addition, the research was funded by the National Police of Colombia and the Institute for Forensic Science of Texas Tech University (TTU), with regard to the projects: residual odor of corpses and their detection by canine teams and Analyzing the Effects of Body Wrapping on Rate of Decomposition, as a collaborative work between the candidate doctoral students in agro-sciences at the Universidad de La Salle and the director and members of the research group: “sniffing knowledge” of the PONAL-ESGAC School for Canine Guidance and Training, GrupLAC: COL0064351, in order to contribute to best practices and better performance of police work canines, as a strategy to reduce crimes associated with the murder and disappearance of people and urban impact and rural security in the post-armed conflict period.
... 2. Проведення вимірювань. Розташування сенсорів на листях дерев згідно розробленої методики [9]; замір висоти розміщення сенсорів; замір відстані сенсорів від координатора; перевірка приєднання всіх сенсорів до мережі, налаштування параметрів вимірювань у меню координатора; запуск вимірювань; перевірка наявності даних у сенсорі; перевірка передачі повідомлень від сенсора до координатора. 3. Реєстрація даних. ...
Article
Full-text available
The authors reviewed the types of network testing. The most common are simulation modeling, mock-up modeling, and full-scale experiments. It was examined existing software environments for simulation and debugging kits for mock testing. The main definitions and terms on the theory of experiment planning are given. According to the theory of experiment planning, the authors developed a plan for conducting a full-scale experiment and defined an algorithm for testing a wireless sensor network for express diagnostics of plants' state. Network testing parameters are the quality of network communication; network formation time; network communication distance; and battery life of sensors. The preparation and process of conducting a full-scale experiment for testing a WSN sample is described in detail. Wireless sensor network testing was carried out by different clusters, at different distances in five stages. During the experiment on testing the WSN, the operation of individual network nodes was checked, and the operation of the network as a whole. During the testing period, no abnormal operation of the sensors and the coordinator was revealed. 82 measurements were made in just five stages. The communication quality of the wireless sensor network has been checked. When testing the network, the transmission of data packets from the sensors to the coordinator was mostly successful. The integral estimate of unsuccessful data transmission sessions in the network was calculated. The communication range of the network at a distance of 20, 30, 40 and 60 m was checked. A graph of the dependence of the sensor signal power on the distance was built. Statistics were obtained on the decrease in the battery charge for each sensor. Based on the results of a full-scale experiment, the operation of a wireless sensor network for express diagnostics of the state of plants is considered successful. Keywords: Zigbee, wireless sensor network, sensor, full-scale experiment.
... В прихованому шарі використовується сигмоїдна функція активації, а на вихідномунормалізована експоненційна функція (softmax function). В роботі [6] показано, що криві розроблених сенсорів ІФХ можна використовувати для таксономічного розрізнення рослин, 5-хвилинні криві дещо краще підходять для даної задачі. В даній статті пропонується дослідити можливість мінімізації входів нейронної мережі. ...
Article
Full-text available
Описано декілька підходів до нормалізації даних хлорофіл-флуорометрів та зменшення їхньої розмірності. The paper describes few approaches to normalization and dimensionality reduction of chlorophyll fluorometer data. Описаны несколько подходов к нормализации данных хлорофилл-флуорометров и уменьшение их размерности.
Article
Full-text available
Using chlorophyll (Chl) a fluorescence many aspects of the photosynthetic apparatus can be studied, both in vitro and, noninvasively, in vivo. Complementary techniques can help to interpret changes in the Chl a fluorescence kinetics. Kalaji et al. (Photosynth Res 122:121-158, 2014a) addressed several questions about instruments, methods and applications based on Chl a fluorescence. Here, additional Chl a fluorescence-related topics are discussed again in a question and answer format. Examples are the effect of connectivity on photochemical quenching, the correction of F V /F M values for PSI fluorescence, the energy partitioning concept, the interpretation of the complementary area, probing the donor side of PSII, the assignment of bands of 77 K fluorescence emission spectra to fluorescence emitters, the relationship between prompt and delayed fluorescence, potential problems when sampling tree canopies, the use of fluorescence parameters in QTL studies, the use of Chl a fluorescence in biosensor applications and the application of neural network approaches for the analysis of fluorescence measurements. The answers draw on knowledge from different Chl a fluorescence analysis domains, yielding in several cases new insights.
Article
Full-text available
In the photosynthesis process, part of the absorbed photon energy in photosystem II (PSII) may be re-emitted as chlorophyll a fluorescence (ChlF). Environmental and plant physiological changes affect the emission of ChlF, which makes ChlF a potentially useful tool to sense these changes. Volumes of research on ChlF are published each year and some published work has been controversial. In order to facilitate the application of ChlF, it is important to have a holistic picture of the field. This review summarizes ChlF applications published in the last 10 years.This article is protected by copyright. All rights reserved.
Article
Full-text available
Photosynthetic apparatus is a highly conservative element of the plant cell and its characteristics are comparatively similar no matter where a plant stands in the taxonomic classification. A problem of great interest is whether a highly informative method can be applied for taxonomic classification of plants based on their photosynthetic characteristics. We used as such one method artificial neural networks and as a source of information about the photosynthetic apparatus, chlorophyll fluorescence induction curves. We used feedforward neural network with back-propagation of error signals through the connections of the network. The designed network has two layers, five hidden neurons, and logarithmic sigmoidal transfer function. As input the network receives the numerical equivalent of the induction curves. In our studies we used induction curves from 35 different species which belonged to 19 different families from 17 orders divided into 11 superorders, all of which were from one of the two major groups of flowering plants (Monocotyledones, Dicotyledones). All together we used 2016 curves (61 curves per plant for most of the species). When a neural network is trained with the whole data set, the error is very high (over 50 %). That’s why we developed a system of neural networks each of which divides a taxonomic rank to its subgroups. This step-by-step approach starts with analysis of the whole data set and ends with classification of each family into its species used in the studies. This system allows taxonomic classification with high accuracy (the error is under 5 %). Considering the written above, we can conclude that photosynthetic apparatus contains information about the genotype of the plants and can be used for their taxonomic classification.
Issue sup1: XI Anniversary scientific conference
Biotechnology & Biotechnological Equipment. Vol. 23, Issue sup1: XI Anniversary scientific conference, 2009 pp. 941-945. ISSN: 1310-2818 http://dx.doi.org/10.1080/13102818.2009.10818577
Wireless sensor network for precision farming and environmental protection. I
  • O Palagin
  • V Romanov
  • I Galelyuka
  • O Voronenko
  • Y Brayko
  • R Imamutdinova
[Palagin at al., 2017] O. Palagin, V. Romanov, I. Galelyuka, O. Voronenko, Y. Brayko, R. Imamutdinova. Wireless sensor network for precision farming and environmental protection. I.Tech 2017