G.N. Zholtkevych1,
G.Yu. Bespalov1,
K.V. Nosov1,
E.V. Visotskaya2,
A.I. Pecherskaya2
1 – V. N. Karazin Kharkiv National University, Ukraine
2 – Kharkiv National University of Radioelectronics, Ukraine
Discrete models of dynamical systems
of relationships between spectral characteristics of grass for remote sensing
of effects disclosing locust crowds
The outbreaks of
locust create heavy biosafety problems related to the damage for an agricultural
sector of economy in different regions of the Planet. These problems are
growing in connection with the global warming. In this regard, detection of
locust crowds on vast areas becomes the task of great practical importance. In
some cases these areas may be hard-to-reach. In such a situation, the role of remote methods is constantly intensifying.
Such methods can use relatively low-cost techniques for registering locust
crowds on the field, e. g., digital photography from light-weight unmanned
aerial vehicle (UAV). The presence of locust's protective coloration masks locust
crowds on the ground vegetation and requires special processing methods for
revealing systemic effects, which can disclose the crowds. One of the methods,
that solves the assigned task, can be built on information technology, which
uses the new, developed with the authors' participation, class of mathematical models [1-2] called discrete
modeling of dynamical systems
(DMDS).
The aim of the
present work is a formalized description of systemic effects characterizing the
dynamics of spectral characteristics of grassland vegetation communities with the
help of the mentioned discrete dynamical models. The character of the systemic
effects can serve as an indicator of presence or absence of crowds of locust,
having protective coloring painting, and can help to disclose such crowds.
As the source data
for modeling the digital photos of locust crowds occurred in summer 2012 in
Astrakhan region (Russia) [3] was taken. For studying systemic effects the
dynamic models based on the concept of limiting factors according to Liebig's
Law of the Minimum [2] were built.
Using the photos the intensities of three colors of RGB color model were
calculated (by averaging through sites under investigation). As components of a
dynamical system the following spectral characteristics of digital photos,
received on the basis of the three color intensities, were used: G/B , R/G, (R+B)/B, where R, G, B is the average intensities of red,
green, and blue colors. To this list of components a latent component was added.
It is assumed, that correlation between the latent component and others
components is to be 0. According to the model, it is deemed that all the
components compose the dynamical system, and separate photographed areas
represent different steps of a single cycle. Using benefits inherent in the
very nature of the DMDS concept, it is possible to restore a dynamics of
changes of these components on the base of separate photos, taken at different
moments. So the long-term monitoring of changes in a grassland community is not
required. Restored dynamics and relationships between components can help us to
understand the system effects facilitating the identification of locust crowds.
The system of
relationships between components obtained by the DMDS allows to build idealized
trajectories of the dynamical system reflecting the typical features of a cycle
of changes of components values for a grassland community in the cases with and
without locust crowds. If there is no crowds on a site, the maximums of the G/B
and R/G values alternate, and that fact can be explained as an indicator of
alternation of "young" and
"old" phases of grassland community's development. Dynamics of the
latent component allows us to interpret it as a performance indicator of
reducers incorporating into trophic chain of dead organic matter, which are
accumulated on "old" phases of plant communities' development.
Locust crowds with protective coloring disturb the
above mentioned cycle of components changes. Instead of the previous cycle an
another cycle appears, which requires different interpretation (this
interpretation is not of interest from
the point of view of the targeted task of disclosing the locust crowds; for
solving this task, of interest is the remote registration of absence of the
cycle generic for a grassland community without locust crowds).
The results obtained
in the work confirm possibilities for remote detection of systemic effects
disclosing locust crowds on grass with the use of the DMDS models and
relatively simple and low-cost technical means. These capabilities can be fully
implemented by information technologies providing calculation of idealized
trajectories by the DMDS methods, that reflects cycles of changes of remotely
recorded parameters (in particular, by digital photos from light-weight UAV)
with follow-up storage of those idealized trajectories in databases (obtained
for various landscapes, weather conditions, etc.). The effectiveness of such
information technologies can be improved at the expense of expert opinions based on the expert's
interpretation of dynamics of latent component's relationship with other
components, by follow-up identification of latent component with some factors
of life activity of ecological systems and biological objects located in the
region of interest and having specific physical or biological significance.
So one of the most
interesting prospect consists on the use of the DMDS models in information
technologies enhancing the capabilities of remote registration of dangerous and
harmful organisms on landscapes with their own well-established communities of
plants and animals.
REFERENCES
1.
G.N. Zholtkevych,
G.Yu. Bespalov, K.V. Nosov,
Mahalakshmi
Abhishek Discrete Modeling of
Dynamics of Zooplankton Community at the Different Stages of an Antropogeneous
Eutrophication //Acta Biotheoretica, 2013. - ¹ 8. – Ñ.
48 – 53.
3.
Æèðíîâà Ò. À.,
Óòàëèåâà À. À., Áðóìøòåéí Þ. Ì. Ñàðàí÷à â àñòðàõàíñêîé îáëàñòè: âîñïðèßòèå
íàñåëåíèåì è ìåòîäû áîðüáû // Àñòðàõàíñêèé âåñòíèê ýêîëîãè÷åñêîãî îáðàçîâàíèÿ.
Âûïóñê ¹ 2 (24). - 2013.