Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики
Every year, the world faces new difficult challenges in maintaining global security. Compliance with food security principles is an important component of the global context of world development. Recent military conflicts have had a strong impact on the development of regions that provide food for m...
Збережено в:
| Дата: | 2023 |
|---|---|
| Автори: | , , , |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"
2023
|
| Теми: | |
| Онлайн доступ: | https://journal.iasa.kpi.ua/article/view/279511 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | System research and information technologies |
| Завантажити файл: |  |
Репозитарії
System research and information technologies| _version_ | 1866302897259544576 |
|---|---|
| author | Zgurovsky, Michael Yefremov, Kostiantyn Gapon, Sergii Pyshnograiev, Ivan |
| author_facet | Zgurovsky, Michael Yefremov, Kostiantyn Gapon, Sergii Pyshnograiev, Ivan |
| author_sort | Zgurovsky, Michael |
| baseUrl_str | http://journal.iasa.kpi.ua/oai |
| collection | OJS |
| datestamp_date | 2023-05-24T21:28:17Z |
| description | Every year, the world faces new difficult challenges in maintaining global security. Compliance with food security principles is an important component of the global context of world development. Recent military conflicts have had a strong impact on the development of regions that provide food for millions of people around the world. Ukraine plays a key role in providing agricultural products to the population of countries from different continents. The article is devoted to the study of the state of agricultural crops in a regional section during the period of active hostilities by means of geomatics, which allow one to assess the degree of transformation of sustainable farming quickly, determine the trend of the development of the industry, and calculate the likely scale of changes in the obtained products in the coming years. As a result, with the help of deep learning models integrated into geoinformation systems, the boundaries of agricultural fields in the Kherson and Zaporizhia regions were determined, the state of moisture and bioproductivity of agricultural crops was determined for three years, an analysis of changes has been made in the state of agricultural fields under the influence of new factors of conducting active hostilities during the first half of 2022, the next harvest productivity forecast was made in two southern regions of Ukraine. The study was carried out by the team of the World Data Center for Geoinformatics and Sustainable Development of the Igor Sikorsky Kyiv Polytechnic Institute. It was part of research on the analysis of the behavior of complex socio-economic systems and processes of sustainable development in the context of the quality and safety of people’s lives. |
| doi_str_mv | 10.20535/SRIT.2308-8893.2023.1.01 |
| first_indexed | 2025-07-17T10:28:06Z |
| format | Article |
| fulltext |
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev, 2023
Системні дослідження та інформаційні технології, 2023, № 1 7
TIДC
ТЕОРЕТИЧНІ ТА ПРИКЛАДНІ ПРОБЛЕМИ
І МЕТОДИ СИСТЕМНОГО АНАЛІЗУ
UDC 519.85 + 502.3/.7 + 004.9
DOI: 10.20535/SRIT.2308-8893.2023.1.01
RESEARCH OF FOOD SECURITY PROBLEMS OF THE WAR-
TORN REGIONS OF UKRAINE USING GEOMATICS METHODS
M. ZGUROVSKY, K. YEFREMOV, S. GAPON, I. PYSHNOGRAIEV
Abstract. Every year, the world faces new difficult challenges in maintaining global
security. Compliance with food security principles is an important component of the
global context of world development. Recent military conflicts have had a strong
impact on the development of regions that provide food for millions of people
around the world. Ukraine plays a key role in providing agricultural products to the
population of countries from different continents. The article is devoted to the study
of the state of agricultural crops in a regional section during the period of active hos-
tilities by means of geomatics, which allow one to assess the degree of transforma-
tion of sustainable farming quickly, determine the trend of the development of the
industry, and calculate the likely scale of changes in the obtained products in the
coming years. As a result, with the help of deep learning models integrated into
geoinformation systems, the boundaries of agricultural fields in the Kherson and
Zaporizhia regions were determined, the state of moisture and bioproductivity of ag-
ricultural crops was determined for three years, an analysis of changes has been
made in the state of agricultural fields under the influence of new factors of conduct-
ing active hostilities during the first half of 2022, the next harvest productivity fore-
cast was made in two southern regions of Ukraine. The study was carried out by the
team of the World Data Center for Geoinformatics and Sustainable Development of
the Igor Sikorsky Kyiv Polytechnic Institute. It was part of research on the analysis
of the behavior of complex socio-economic systems and processes of sustainable
development in the context of the quality and safety of people’s lives.
Keywords: food security, spatial data analysis, deep learning, agricultural fields,
mathematical modeling.
INTRODUCTION
Intensification of extreme weather conditions, climate change processes, coro-
navirus pandemic, etc. led to an aggravation of the food situation for many coun-
tries of the world [1]. Before Russia`s full-scale invasion of Ukraine, the world
was close to a global food crisis, but since February 24, 2022, the situation has
significantly worsened [2]. Ukraine is a supplier of a large number of agricultural
products to dozens of taps in the world. Ukraine has 15% of the global product
market (UBTA) for individual grain crops [3]. Some countries of the world de-
pend on certain types of agricultural products from Ukraine for more than 50%.
The full-scale war led to a significant reduction in the area of cultivated agricul-
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 8
tural areas, a decrease in the number of people and equipment involved in the cul-
tivation of agricultural crops. In addition, the structure of management in the field
of irrigation was transformed.
Combat actions and specific management decisions in the temporarily
occupied territories have significantly changed the irrigation system that has
existed for many years. The canal systems, which were fed from the Dnieper and
transported water hundreds of kilometers to the south and east of the country, are
under temporary occupation and their state of functioning is difficult to
investigate. At the same time, the region continues to be a supplier of a significant
amount of agricultural products, so it is extremely important to understand the
degree of transformation processes in order to assess possible losses and predict
the degree of the food crisis. Due to constant military operations, direct access to
the territory is extremely difficult, reliable statistical data on the volume and
condition of the harvest, irrigation of the territory, etc. are not collected, the only
possible methods of assessing the degree of transformation of agricultural fields
are methods of remote sensing of the Earth, and geomatics in general. The most
characteristic signs of the transformation of the water regime, especially for
irrigated areas, are signs of a sharp change in indicators of bioproductivity and
territory moisture. Such characteristics can be obtained with high accuracy based
on the analysis of satellite images of medium resolution, which is not a limiting
factor based on the realities of war.
This study is a continuation of the thematic research of the team of authors
on the study of sustainable development of communities and territories of Ukraine
and security processes in the regions of the state [4, 5] and research on the
development of the applications of geomatics methods [6, 7].
DATA FRAMEWORK
Two southern regions of Ukraine: Kherson and Zaporizhzhya, were chosen to
assess the impact of the processes associated with the occupation on the condition
of agricultural plots (Fig. 1).
Fig. 1. The study area (black borders) with the indicated averaged zones of temporary
occupation as of May 2022 (gray color)
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 9
For most of the first four months of the active phase of the war, these areas
have been partially occupied, and this period completely coincides with the period
of active agricultural work, in particular with the irrigation of the territory.
Remote monitoring of the state of moisture in agricultural fields, analysis of the
distribution of the Normalized Difference Moisture Index (NDMI) on the
territory, its dynamics over different years will allow to assess the degree of
changes in sustainable agricultural practices in the region and assess the state of
the potential harvest, which in the pre-war period for years provided food for the
population of Ukraine and residents of countries that import agricultural products.
The selected areas have a high share of plowed territory, both irrigated and non-
irrigated agricultural fields, and a dense and fairly even distribution of plots
throughout the territory. This makes it possible to distinguish the anthropogenic
influence of the occupation itself on the situation with the moisture of agricultural
fields from the general background climatic influence. The condition of plots
exclusively on traditionally irrigated lands can be analyzed separately. In addition
to the analysis of the differentiation of the moisture index, it is necessary to
investigate the change in the bioproductivity index: Normalized Difference
Vegetation Index (NDVI), which on the one hand directly correlates with the
moisture content of the territory, but allows to distinguish the vegetation state of
the vegetation itself, which affects the indices of the moisture index itself (there
may be weakly moistened territories, however, due to the high vegetation, give
the moisture index high values). The condition of plots exclusively on
traditionally irrigated lands can be analyzed separately. In addition to the analysis
of the differentiation of the moisture index, it is necessary to investigate the
change in the bio productivity index: NDVI, which on the one hand directly
correlates with the moisture content of the territory, but allows to distinguish the
vegetation state of the vegetation itself, which affects the indices of the moisture
index itself (there may be weakly moistened territories, however, due to the high
vegetation, it gives the moisture index of high values).
To assess the impact of the occupation, mainly data from remote sensing of
the Earth (DSR) [8], data on the administrative and territorial structure of the
country [9] and the borders of temporarily occupied territories from open online
sources [10] were used. For the analysis of moisture and bio productivity, data
were obtained from the Sentinel-2 mission satellite, platforms 2-A and 2-B and
product type S2MSI1C with a cloud cover of no more than 10% in the study area
[11]. The resolution of multispectral three-channel images is 10 m per pixel,
single-channel 20 m per pixel. Channels B08 and B11, bio productivity index:
B08 and B04 were used to calculate the moisture index.
NDMI = (B08 – B11)/(B08 + B11); (1)
NDVI = (B08 – B04)/(B08 + B04). (2)
Data on the boundaries of regions, districts and communities of the study
area were obtained from the official website of the support for the decentraliza-
tion reform [12].
The boundary of the line of contact of the troops is dynamic and has not yet
been determined, therefore the boundaries of the temporarily occupied territories
for the Kherson and Zaporizhzhia regions were drawn quite approximately based
on an integrated analysis of data on the boundaries of the occupation zones as of
the end of May 2022 according to publicly available web map data [13, 14].
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 10
RESEARCH FRAMEWORK
To analyze the impact of the occupation, mainly geomatics methods, systems
analysis methods and machine learning methods were used. In particular, the
overlay method and map algebra method were used in desktop GIS [15, 16] even
before the beginning of 2022 to calculate bio productivity and moisture indices.
The papers consider aspects of integration of temporal statistical
characteristics with spectral and textural characteristics extracted from high-
quality Sentinel-2 images using Random Forest classification [17]. The
performance and contribution of different combinations is evaluated based on
classification accuracy. The results show that the statistical analysis of time series
is an effective way of presenting information about the degree of soil moisture.
The method uses clear pixels from dense, low-quality images to derive NDVI
statistics, thus reducing the influence of random factors such as weather conditions.
Approaches to developing a linear mixed effects (LME) model for poorly
calculated areas using time series of Sentinel 1A and 1B images and ground
measurements of soil moisture are considered [18]. The model assumes a linear
relationship that varies in both time and space between soil moisture and
backscatter coefficient. The LSE model can be effectively applied to estimate soil
moisture from multi-temporal Sentinel-1 images, which is useful for flood and
drought monitoring and improving runoff forecasting.
Techniques for mapping soil moisture and irrigation at the scale of
agricultural fields based on the synergistic interpretation of multitemporal optical
and synthetic aperture radar (SAR) data (Sentinel-2 and Sentinel-1) were also
presented [19]. The resulting irrigation maps were validated using reference fields
in the study area. The best results were obtained with classifications based only
on soil moisture indicators, with an accuracy of 77%.
An important aspect of the study was the separation of data on the NDVI and
NDMI indices exclusively for the territory of the agricultural fields of the two
regions, without considering the surrounding roads, settlements, water bodies,
forested areas, etc. For this, the model was trained using the Image Analyst
module, and the machine learning method integrated with desktop GIS was used:
Detect Objects Using Deep Learning [20]. For training the model, the Non-
Maximum Suppression parameter was used to detect and remove duplicate
objects (Fig. 2).
Fig. 2. Reference set of polygons for training the model for identifying the boundaries of
agricultural fields
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 11
The model tool processes the input geospatial images that are in the extent of
the project map. The following approaches were used to train the model:
MaskRCNN Object detection (Fig. 3) and Single Shot Detector (SSD) (Fig. 4) [21].
The Mask R-CNN is obtainded rablacing the RoI pool by RoIAlign in Faster
R-CNN. It helps to preserve spatial information which gets misaligned in case of
RoI pool. RoIAlign uses binary interpolation to create a feature map that is of
fixed size for e.g. 7 x 7. The output from RoIAlign layer is then fed into Mask
head, which consists of two convolution layers. It generates mask for each RoI,
thus segmenting an image in pixel-to-pixel manner.
During training our models we need to outline which default boxes corre-
spond to a ground truth detection and train the network accordingly. To achieve
this, we need to determine properly objective loss function for SSD model. The
SSD training objective is gotten from [24]. Let 0,1i p
jx be an indicator for
matching the i-th default box to the j-th ground truth box of category p. In the
matching strategy were 1 p
iJix . The overall objective loss function is a
weighted sum of the localization loss (loc) and the confidence loss (conf):
)),,,(),((
1
),,,( glxLcxL
N
glcxL locconf
Fig. 3. The Mask R-CNN framework for instance segmentation [22]
Fig. 4. SSD architecture [23]
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 12
where N is the number of matched default boxes. If N = 0, the loss is set to 0. The
localization loss is a Smooth L1 loss between the predicted box (l) and the ground
truth box (g) parameters. It is regressed to offsets for the center (cx, cy) of the de-
fault bounding box (d) and for its width (w) and height (h)
),ˆ( ),,( 1
},,,{
m
j
m
iL
k
ij
hwcycxm
N
Posi
loc glsmoothxglxL
, ˆ
w
i
cx
i
cx
jcx
j
d
dg
g
,
)(
ˆ
h
i
cy
i
cy
jcy
j
d
dg
g
, logˆ
w
i
w
jw
j
d
g
g
h
i
h
jh
j
d
g
g logˆ .
The confidence loss is the softmax loss over multiple classes confidences (c):
)ˆ(log)ˆ(log),( 0
i
Negi
p
i
p
ij
N
Posi
conf ccxcxL
,
where
)(exp
)(exp
ˆ
p
ip
p
ip
i
c
c
c
and the weight term α is set to 1 by cross validation.
MaskRCNN and SSD are used for segmentation and precise delineation of
object boundaries on a space image.
All calculations of index values were carried out for territories that were
within the boundaries of the identified plots. Based on the fact that in the resulting
geospatial layers (GSP) there were several million individual values regarding the
characteristics of moisture and bio productivity of agricultural fields, the process-
ing results were processed in the R software environment.
Using the capabilities of the Copernicus Open Access Hub [25], images
were uploaded to the territory of Kherson and Zaporizhzhia regions for the period
May-June 2019–2022. Each image had to be covered by clouds no more than
10%. Due to the presence of many wet atmospheric fronts in the specified period
of the year, such a wide permissible time period was chosen for uploading
images, where priority was given to space images that were taken in the first half
of June (70% of the received images). Due to the unsatisfactory state of cloud
coverage of the images, the period of the end of May (12% of the images) and the
second half of June (18% of the images) was chosen for the rest of the scenes. For
each year, a minimum of 8 separate photo scenes were uploaded, which covered
at least 97% of the territory of the selected regions (Fig. 5).
For each year, a new mosaic of both an integral image in the visible range
(RGB) and a new mosaic of individual spectral channels (B04, B08, B11) was
created from a series of separate images to calculate the moisture and bio produc-
tivity indices. Three-spectral rasters in the visible range and single-spectral rasters
in the index range were obtained at the output. A raster in the visible range allows
you to visually examine the area for artifacts of space images, and, if necessary,
replace individual scenes with those that meet the requirements for the visibility
of black and white fields. From the new mosaics of individual spectral channels,
integral rasters of indices were calculated for each year (Fig. 6).
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 13
The new image mosaics must be separated from the two areas to avoid ana-
lyzing areas that are outside the study ones. To do this, a process of raster extrac-
tion along the contours of the Kherson and Zaporizhzhia regions was carried out
for all new raster mosaics obtained. The resulting GPS included both those neces-
sary for the analysis of the territory of agricultural fields, as well as external water
bodies, urbanized areas, infrastructure facilities, etc. For their illumination and
selection of exclusively rural areas, training of a machine learning model inte-
grated into the capabilities of the geographic information system (GIS) was car-
ried out.
To carry out the model training process, it was necessary to manually high-
light the boundaries of several thousand agricultural fields on space images for
the selected period. The fields were vectorized evenly over the territory of the two
regions with important identification of the borders of both irrigated and non-
irrigated areas (Fig. 7). Often, irrigated areas have a rather specific contour of a
regular circle, which, with insufficient training of the model on these fields, can
lead to incorrect identification of boundaries.
Fig. 5. Coverage of the research area with space images
Fig. 6. A fragment of the territory moisture index raster for 2022
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 14
The training process took place in a GIS environment where image tiles
were first created as input layers to train the model. Image tile size was 448 pixels
with metadata format: PASCAL Visual Object Classes and RCNN Masks. The
image batch size type was 8, the model was run for 100 iterations (epochs). The
additional pixel border around each field is 2 pixels. The maximum overlap of the
resulting boundaries is 0.1, the minimum reliability of the selected boundaries is
60% (0.6). In total, the model consisted of 12 iterations of corrections and
additional training.
As a result, the GIS was obtained with about 370,000 identified agricultural
fields for the territory of two regions: 210.000 for the Kherson region and 160.000
for the Zaporizhia region (Fig. 8). The average reliability of the selected limits
was 75%.
Further improvement in model accuracy can be performed to achieve other
applied goals in agriculture. For assessing the degree of transformation of the
moisture regime and bio productivity of the fields, the obtained reliability is
considered completely satisfying.
The extraction of new raster mosaics of the moisture index and bio produc-
tivity was carried out based on the obtained field mask. The final rasters began to
calculate data exclusively for agricultural plots (Fig. 9).
Fig. 7. Characteristic boundaries of irrigated (1) and non-irrigated (2) fields
Fig. 8. Identified borders of agricultural fields for Kherson and Zaporizhzhia regions
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 15
To assess the state of agricultural fields within individual regions, districts
and communities, in the context of temporarily occupied and government-
controlled territories, it is necessary to convert raster images into vector format
and supplement the GIS with attributive data of the layers of the administrative-
territorial system (ATU). After converting the data into a vector format, the area
of each cell of the new GPS was calculated to obtain the ratio of the areas of areas
with different humidity for each unit of ATU. The resulting layers consisted of
more than 10 million records, the calculation of which was extremely difficult
exclusively with GIS tools, so the corresponding statistical processing was carried
out in the R software environment.
THE MOISTURE LEVEL ANALYSIS FOR THE IDENTIFIED FIELDS
Based on the rasters of the thematic indexes of agricultural fields, it is possible to
make a preliminary integrated analysis for two regions, without dividing the re-
gions into districts, communities, and the zone of occupation. Using the methods
of zonal statistics, data were obtained on the average values of the moisture index
rasters for 2019–2022 (Fig. 10). It can be seen from the graphs that over the past 4
years there has been a rather strong spread of index values: from 0.02 in 2020 to
0.07 in 2019. Moreover, the structure of the distribution of values is not uniform:
normal or lognormal distribution in 2019 and 2020, and a distribution with two
peaks in 2021 and 2022, indicating dry periods with a strong influence of irriga-
tion systems in particular.
A normal or lognormal distribution indicates the classic case in which the
average values of the wetness index cover a larger area. The values for 2019 cor-
respond to this distribution, with a certain local peak on the graph for values that
characterize low, poorly moistened vegetation. Most of the territory, according to
the distribution, is covered with medium-low vegetation with low water stress.
The log-normal distribution for 2020 is characterized by a large peak in the plot
for coverage for low and dry vegetation, resulting in an overall lowest 4-year
mean wetness index value. This situation strongly correlates with climate indica-
tors, according to which 2020 was the driest year in terms of precipitation, which
Fig. 9. A fragment of the moisture index raster for agricultural fields in 2022
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 16
caused a sharp drop in water levels in natural and anthropogenic reservoirs in the
region. For 2021, two peaks on the graph are characteristic: for low and poorly
moistened vegetation, and for medium-high, medium with low water stress. This
distribution is caused by the contrasting weather conditions in May–June 2021,
when the dry period ended with intense precipitation combined with a strong con-
trast in the wetness index for irrigated and non-irrigated areas. Where the high
value of the index is characteristic mainly of irrigated fields, which occupy sig-
nificant areas in the Kherson region. This statement is supported by further re-
search. The year 2022 is also characterized by two peaks on the graph, among
which the larger peak is responsible for poorly watered areas, and the smaller one
is for vegetation with low water stress.
The territorial analysis of the moisture content of agricultural fields for 4
years showed a decrease in plots with medium-high vegetation cover and low wa-
ter stress in two regions in 2022 (Fig. 11).
Fig. 10. Distribution of the humidity index for 2019–2022 in the Kherson and Zaporizhia
regions
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 17
Even compared to 2020, when a dry period was observed in both regions, the
area of fields with low water stress in 2022 decreased by 8% in Kherson region
and by 24% in Zaporizhzhya region. The Kherson region has a larger number of
irrigated areas, which, most likely, continued to be actively irrigated in 2022,
which quite strongly smoothed out the drop in the index value. Compared to the
previous year 2021, for 2022 the drop was 60% for both Kherson and Za-
porizhzhia regions. Such a rapid decrease in well-watered agricultural fields can
be explained by the transformation in the conduct of irrigation and other proc-
esses in the agriculture of the region, which, in turn, is most likely directly related
to the temporary occupation of the region. This statement requires further re-
search in the context of expanding the experimental period of observations of the
dynamics of the territory’s moisture index. Most likely, there is a certain shift in
the phases of irrigation, which leads to a shift in the vegetation phases of agricul-
tural plants. From the graph of the dynamics of the index, it can be seen that in
2022 there is a significant increase in the area of fields with medium-low vegeta-
tion cover and low water stress compared to all other years: an increase of 119%
in Kherson region and 93% in Zaporizhia region compared to 2021 However,
Fig. 11. Dynamics of the humidity index of the territory
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 18
compared to 2019, there are almost no changes in areas, which most likely indi-
cates a significant influence of climatic factors on the value of the indicator.
The influence of climatic factors, namely a rather cold and wet spring, also
affected the distribution of areas with dry and very low vegetation cover (Fig. 12).
In 2022, there is a drop in this indicator compared to other years. This is es-
pecially characteristic of the Kherson region, where favorable synoptic conditions
combined with artificial irrigation processes significantly reduced the index with
low values.
An analysis was conducted exclusively for irrigated fields. For this purpose,
those that are not reached by the system of irrigation canals were illuminated from
the entire array of plots. The situation with the irrigated area is quite dynamic, so
the average indicator of irrigated areas over 4 years was chosen. The Kherson re-
gion has a much higher density of canals and irrigated areas in general (Fig. 13).
Therefore, fields with low water stress occupy much more area in the Kherson
region, where there are almost no irrigated fields with dry vegetation (less than 4%).
Territorial analysis of the influence of occupation, in contrast to the temporal
distribution, did not show any differences in the condition of agricultural fields in
temporarily occupied and unoccupied areas of the region. Intense hostilities affect
the entire territory of the regions, where the demarcation zone dynamically
changes many times a month. Moreover, the most cultivated and irrigated parts of
both the Kherson and Zaporizhzhia regions have been under temporary occupa-
tion since the first days of the war, the unoccupied territory consists of fields that
are not reached by canal branches. This condition leads to a slight proportional
increase in the area of poorly moistened areas precisely in the unoccupied territo-
ries (Fig. 14). However, all other index values remain proportionally almost the
same.
Fig. 12. Percentage distribution of the territory's humidity index values
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 19
CONCLUSIONS
1. The temporary occupation of the part of the Ukraine’s southern regions
leads to significant transformations in the structure of agriculture. Even though
almost all the capacities of the irrigation systems of the Kherson and Zaporizhia
regions was under the control of the occupiers, there is a significant decrease in
Fig. 13. Correlation of moisture index values for irrigated areas in 2022
Fig. 14. Differentiation of the humidity index in temporarily occupied and unoccupied
territories
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 20
agricultural fields in a well-watered state (up to 60% compared to previous years).
This may indicate the imperfect use of existing irrigation facilities and the reluc-
tance of local farmers to cooperate with the occupation authorities.
2. However, at the same time, the catastrophic impact of temporary occupa-
tion on the state of agricultural land is not observed. Most of them are in a
satisfactory condition, which makes it possible to predict a slight decrease in the
amount of potentially harvested agricultural products. Active military actions do
not allow agricultural works to be carried out fully in the unoccupied part of the
regions, which affects the state of agricultural crops. However, even in such a dif-
ficult situation, the state of moistening of the fields of most of the unoccupied ter-
ritory remains satisfactory. Which also indicates a favorable forecast for the col-
lection of agricultural crops. However, the factor of conducting hostilities, which
can significantly worsen the situation with the state of agricultural fields in both
occupied and non-occupied territory, remains unpredictable.
3. The prepared machine learning model for identifying the boundaries of
agricultural plots significantly improved the accuracy of the estimates made by
illuminating all extraneous territories. In individual communities of the region,
without considering the results of the model, the indicators of the state of hydra-
tion differed by 10–15% compared to the indicators of the indexes calculated ex-
clusively within the boundaries of the plots.
4. The developed machine learning model can be applied to other regions of
Ukraine, which will make it possible to assess the impact of military operations
and/or temporary occupation for all affected regions. It is urgent to expand the
research area in the following research to Mykolaiv region, which was also sig-
nificantly affected by the temporary occupation of its territory.
5. Similar methods can be applied to those regions that have not undergone
occupation, in the context of temporal and territorial analysis of the condition of
agricultural fields under conditions of climate change, etc.
FUNDING INFORMATION
The proposed research is carried out as part of the systematic research of the In-
formation and Analytical Situational Center of Igor Sikorsky Kyiv Polytechnic
Institute on the projects “Scenario modeling based on satellite data of critical
changes in the ecological and economic state of temporarily occupied territories
as a factor of national security of Ukraine” (0122U001437).
REFERENCES
1. P. Hebinck and H. Oostindie, “Performing food and nutritional security in Europe:
claims, promises and limitations,” Food Sec., vol. 10, pp. 1311–1324, 2018. doi:
10.1007/s12571-018-0853-9.
2. M. Tanchum, The Russia-Ukraine War has Turned Egypt’s Food Crisis into an Exis-
tential Threat to the Economy. Middle East Institute, MEI, 2022. CID:
20.500.12592/0h424d.
3. Agricultural Production. [Online]. Available: https://ourworldindata.org/ agricul-
tural-production
4. Michael Zgurovsky, Andrii Boldak, Kostiantyn Yefremov, and Ivan Pyshnograiev,
“Modeling and investigating the behavior of complex socio-economic systems,”
Research of food security problems of the war-torn regions of Ukraine using geomatics methods
Системні дослідження та інформаційні технології, 2023, № 1 21
2017 IEEE First Ukraine Conference on Electrical and Computer Engineering
(UKRCON), pp. 1113–1116, 2017. doi: 10.1109/UKRCON.2017.8100400.
5. M. Zgurovsky, K. Yefremov, I. Pyshnograiev, A. Boldak and I. Dzhygyrey, “Quality
and Security of Life: A Cross-Country Analysis,” 2022 IEEE 3rd International
Conference on System Analysis & Intelligent Computing (SAIC), 2022, pp. 1–5.
doi: 10.1109/SAIC57818.2022.9923006.
6. M. Zgurovsky et al., “Parameterization of Sustainable Development Components
Using Nightlight Indicators in Ukraine,” 2018 IEEE First International Conference
on System Analysis & Intelligent Computing (SAIC), 2018, pp. 1–5. doi:
10.1109/SAIC.2018.8516726.
7. M. Zgurovsky, K. Yefremov, S. Gapon and I. Pyshnograiev, “Modeling of Potential
Flooding Zones with Geomatics Tools,” 2022 IEEE 3rd International Confer-
ence on System Analysis & Intelligent Computing (SAIC), 2022, pp. 1–4. doi:
10.1109/SAIC57818.2022.99230.
8. Heike Schmidt and Arnon Karnieli, “Remote sensing of the seasonal variability of
vegetation in a semi-arid environment,” Journal of Arid Environments, vol. 45, iss.
1, pp. 43–59, 2000. doi: 10.1006/jare.1999.0607.
9. Administrative boundaries of Ukraine. [Online]. Available: https://data.humdata.org/
dataset/ukraine-administrative-boundaries
10. Interactive Map: Russia’s Invasion of Ukraine. [Online]. Available:
https://www.understandingwar.org/
11. European Union’s Earth observation programme. [Online]. Available:
https://www.copernicus.eu/en
12. Support site for Ukraine’s decentralization reform. [Online]. Available:
https://atu.decentralization.gov.ua/
13. DeepState Map: Russia’s Invasion. [Online]. Available: https://deepstatemap.live/
14. Map tells story about war in Ukraine and Eastern Europe. [Online]. Available:
https://liveuamap.com/
15. M. Wu et al., “Monitoring cotton root rot by synthetic Sentinel-2 NDVI time series
using improved spatial and temporal data fusion,” Sci. Rep., vol. 8, 2016 (2018). doi:
10.1038/s41598-018-20156-z.
16. M.S. Boori, K. Choudhary, R. Paringer, A.K. Sharma, A. Kupriyanov, and S. Cor-
gne, “Monitoring Crop Phenology Using NDVI Time Series from Sentinel 2 Satel-
lite Data,” 2019 5th International Conference on Frontiers of Signal Processing
(ICFSP), 2019, pp. 62–66. doi: 10.1109/ICFSP48124.2019.8938078.
17. C. Huang et al., “Land Cover Mapping in Cloud-Prone Tropical Areas Using Senti-
nel-2 Data: Integrating Spectral Features with Ndvi Temporal Dynamics,” Remote
Sens., 12(7): 1163, 2020. doi: 10.3390/rs12071163.
18. Y. Zhang, J. Gong, K. Sun, J. Yin, and X. Chen, “Estimation of Soil Moisture Index
Using Multi-Temporal Sentinel-1 Images over Poyang Lake Ungauged Zone,” Re-
mote Sens., 10(1), 12, 2018. doi: 10.3390/rs10010012.
19. S. Bousbih et al., “Soil Moisture and Irrigation Mapping in a Semi-Arid Region,
Based on the Synergetic Use of Sentinel-1 and Sentinel-2 Data,” Remote Sens.,
10(12), 1953, 2018. doi: 10.3390/rs10121953.
20. Z.-Q. Zhao, P. Zheng, S.-T. Xu and X. Wu, “Object Detection With Deep Learning:
A Review,” in IEEE Transactions on Neural Networks and Learning Systems, vol.
30, no. 11, pp. 3212–3232, Nov. 2019. doi: 10.1109/TNNLS.2018.2876865.
21. E.A. Alburshaid and M.A. Mangoud, “Palm Trees Detection Using the Integration
between GIS and Deep Learning,” 2021 International Symposium on Networks,
Computers and Communications (ISNCC), 2021, pp. 1–6. doi: 10.1109/ISNCC52172.
2021.9615721.
22. Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick, Mask R-CNN. 2017.
Available: http://arxiv.org/abs/1703.06870
M. Zgurovsky, K. Yefremov, S. Gapon, I. Pyshnograiev
ISSN 1681–6048 System Research & Information Technologies, 2023, № 1 22
23. J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified,
real-time object detection,” in CVPR, 2016.
24. W. Liu et al., “SSD: Single Shot MultiBox Detector,” in B. Leibe, J. Matas, N. Sebe,
M. Welling (eds) Computer Vision – ECCV 2016. Lecture Notes in Computer Sci-
ence, vol. 9905. Springer, Cham, 2016. doi: 10.1007/978-3-319-46448-0_2.
25. Copernicus Open Access Hub. [Online]. Available: https://scihub.copernicus.eu/
dhus/#/home
Received 12.02.2023
INFORMATION ON THE ARTICLE
Michael Z. Zgurovsky, ORCID: 0000-0001-5896-7466, National Technical University of
Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail: zgurovsm@hotmail.com
Kostiantyn V. Yefremov, ORCID: 0000-0003-3495-6417, National Technical University
of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail:
k.yefremov@wdc.org.ua
Sergii V. Gapon, ORCID: 0000-0002-8834-5825, National Technical University of
Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail: gapon@wdc.org.ua
Ivan O. Pyshnograiev, ORCID: 0000-0002-3346-8318, National Technical University of
Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail: pyshno-
graiev@gmail.com
ДОСЛІДЖЕННЯ ПРОБЛЕМ ПРОДОВОЛЬЧОЇ БЕЗПЕКИ, ОХОПЛЕНИХ
ВІЙНОЮ РЕГІОНІВ УКРАЇНИ МЕТОДАМИ ГЕОМАТИКИ / М.З. Згуровський,
К.В. Єфремов, С.В. Гапон, І.О. Пишнограєв
Анотація. Світ з кожним роком наражається на нові важкі виклики щодо
підтримання глобальної безпеки. Важливою складовою глобального контек-
сту світового розвитку є дотримання принципів продовольчої безпеки. Новітні
військові конфлікти сильно впливають на стан розвитку регіонів, які забезпе-
чують мільйони людей по всьому світу продовольством. Україна відіграє клю-
чову роль у глобальних процесах забезпечення продукцією сільського госпо-
дарства населення країн з різних континентів. Присвячено дослідженню станів
сільськогосподарських культур у регіональному розрізі у період ведення ак-
тивних бойових дій засобами геоматики, що дозволяє швидко оцінити ступінь
трансформації сталого господарювання, визначити тренд розвитку галузі, об-
числити ймовірні масштаби зміни отриманої про-дукції у найближчі роки. У
результаті за допомогою інтегрованих у геоінформаційні системи моделей
глибинного навчання визначено межі сільськогосподарських полів
Херсонської та Запорізької областей, стан зволоженості та біопродуктивності
сільськогосподарських культур за три роки, проаналізовано зміни станів
сільськогосподарських полів під впливом нових факторів ведення активних
бойових дій за першу половину 2022 р., зроблено прогноз продуктивності на-
ступного врожаю у двох південних областях України. Наведене дослідження
виконано командою Світового центру даних «Геоінформатика та сталий роз-
виток» КПІ ім. Ігоря Сікорського і є частиною досліджень з аналізу поведінки
складних соціально-економічних систем та процесів сталого розвитку в
контексті якості та безпеки життя людей.
Ключові слова: продовольча безпека, просторовий аналіз даних, глибинне
навчання, сільськогосподарські поля, математичне моделювання.
|
| id | journaliasakpiua-article-279511 |
| institution | System research and information technologies |
| keywords_txt_mv | keywords |
| language | English |
| last_indexed | 2025-07-17T10:28:06Z |
| publishDate | 2023 |
| publisher | The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" |
| record_format | ojs |
| resource_txt_mv | journaliasakpiua/ef/b668bc367c4ed9b5432a0bbaf50e15ef.pdf |
| spelling | journaliasakpiua-article-2795112023-05-24T21:28:17Z Research of food security problems of the war-torn regions of Ukraine using geomatics methods Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики Zgurovsky, Michael Yefremov, Kostiantyn Gapon, Sergii Pyshnograiev, Ivan продовольча безпека просторовий аналіз даних глибинне навчання сільськогосподарські поля математичне моделювання food security spatial data analysis deep learning agricultural fields mathematical modeling Every year, the world faces new difficult challenges in maintaining global security. Compliance with food security principles is an important component of the global context of world development. Recent military conflicts have had a strong impact on the development of regions that provide food for millions of people around the world. Ukraine plays a key role in providing agricultural products to the population of countries from different continents. The article is devoted to the study of the state of agricultural crops in a regional section during the period of active hostilities by means of geomatics, which allow one to assess the degree of transformation of sustainable farming quickly, determine the trend of the development of the industry, and calculate the likely scale of changes in the obtained products in the coming years. As a result, with the help of deep learning models integrated into geoinformation systems, the boundaries of agricultural fields in the Kherson and Zaporizhia regions were determined, the state of moisture and bioproductivity of agricultural crops was determined for three years, an analysis of changes has been made in the state of agricultural fields under the influence of new factors of conducting active hostilities during the first half of 2022, the next harvest productivity forecast was made in two southern regions of Ukraine. The study was carried out by the team of the World Data Center for Geoinformatics and Sustainable Development of the Igor Sikorsky Kyiv Polytechnic Institute. It was part of research on the analysis of the behavior of complex socio-economic systems and processes of sustainable development in the context of the quality and safety of people’s lives. Світ з кожним роком наражається на нові важкі виклики щодо підтримання глобальної безпеки. Важливою складовою глобального контексту світового розвитку є дотримання принципів продовольчої безпеки. Новітні військові конфлікти сильно впливають на стан розвитку регіонів, які забезпечують мільйони людей по всьому світу продовольством. Україна відіграє ключову роль у глобальних процесах забезпечення продукцією сільського господарства населення країн з різних континентів. Присвячено дослідженню станів сільськогосподарських культур у регіональному розрізі у період ведення активних бойових дій засобами геоматики, що дозволяє швидко оцінити ступінь трансформації сталого господарювання, визначити тренд розвитку галузі, обчислити ймовірні масштаби зміни отриманої продукції у найближчі роки. У результаті за допомогою інтегрованих у геоінформаційні системи моделей глибинного навчання визначено межі сільськогосподарських полів Херсонської та Запорізької областей, стан зволоженості та біопродуктивності сільськогосподарських культур за три роки, проаналізовано зміни станів сільськогосподарських полів під впливом нових факторів ведення активних бойових дій за першу половину 2022 р., зроблено прогноз продуктивності наступного врожаю у двох південних областях України. Наведене дослідження виконано командою Світового центру даних "Геоінформатика та сталий розвиток" КПІ ім. Ігоря Сікорського і є частиною досліджень з аналізу поведінки складних соціально-економічних систем та процесів сталого розвитку в контексті якості та безпеки життя людей. The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" 2023-03-30 Article Article application/pdf https://journal.iasa.kpi.ua/article/view/279511 10.20535/SRIT.2308-8893.2023.1.01 System research and information technologies; No. 1 (2023); 7-22 Системные исследования и информационные технологии; № 1 (2023); 7-22 Системні дослідження та інформаційні технології; № 1 (2023); 7-22 2308-8893 1681-6048 en https://journal.iasa.kpi.ua/article/view/279511/274104 |
| spellingShingle | продовольча безпека просторовий аналіз даних глибинне навчання сільськогосподарські поля математичне моделювання Zgurovsky, Michael Yefremov, Kostiantyn Gapon, Sergii Pyshnograiev, Ivan Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики |
| title | Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики |
| title_alt | Research of food security problems of the war-torn regions of Ukraine using geomatics methods |
| title_full | Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики |
| title_fullStr | Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики |
| title_full_unstemmed | Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики |
| title_short | Дослідження проблем продовольчої безпеки, охоплених війною регіонів України методами геоматики |
| title_sort | дослідження проблем продовольчої безпеки, охоплених війною регіонів україни методами геоматики |
| topic | продовольча безпека просторовий аналіз даних глибинне навчання сільськогосподарські поля математичне моделювання |
| topic_facet | продовольча безпека просторовий аналіз даних глибинне навчання сільськогосподарські поля математичне моделювання food security spatial data analysis deep learning agricultural fields mathematical modeling |
| url | https://journal.iasa.kpi.ua/article/view/279511 |
| work_keys_str_mv | AT zgurovskymichael researchoffoodsecurityproblemsofthewartornregionsofukraineusinggeomaticsmethods AT yefremovkostiantyn researchoffoodsecurityproblemsofthewartornregionsofukraineusinggeomaticsmethods AT gaponsergii researchoffoodsecurityproblemsofthewartornregionsofukraineusinggeomaticsmethods AT pyshnograievivan researchoffoodsecurityproblemsofthewartornregionsofukraineusinggeomaticsmethods AT zgurovskymichael doslídžennâproblemprodovolʹčoíbezpekiohoplenihvíjnoûregíonívukraínimetodamigeomatiki AT yefremovkostiantyn doslídžennâproblemprodovolʹčoíbezpekiohoplenihvíjnoûregíonívukraínimetodamigeomatiki AT gaponsergii doslídžennâproblemprodovolʹčoíbezpekiohoplenihvíjnoûregíonívukraínimetodamigeomatiki AT pyshnograievivan doslídžennâproblemprodovolʹčoíbezpekiohoplenihvíjnoûregíonívukraínimetodamigeomatiki |