A Project Report

On

SPATIO-TEMPORAL DYNAMICS OF LAND COVER OF

VISAKHAPATNAM USING CLOUD BASED ARTIFICIAL INTELLIGENCE
TECHNIQUES

Submitted in partial fulfillment for the award of

BACHELOR OF ENGINEERING

IN

CIVIL ENGINEERING BY

BANDARPALLY ANJALI 2451-20-732-014

MIDDE AKSHAY 2451-20-732-032
PERI VENKAT VISHWANATH 2451-20-732-043
SARIGARI NITHIN SAI 2451-20-732-044
Under the guidance of

Dr. V. SHIVA CHANDRA Mrs. K. SWETHA

Assistant professor Assistant professor
Department of Civil Engineering

Maruri Venkata Subba Rao Engineering College, Hyderabad.

DEPARTMENT OF CIVIL ENGINEERING

MATURI VENKATA SUBBA RAO ENGINEERING COLLEGE

(Affiliated to Osmania University & Approved by AICTE) Nadergul, Hyderabad-

501 510
 DEPARTMENT OF CIVIL ENGINEERING

MATURI VENKATA SUBBA RAO ENGINEERING COLLEGE

(Affiliated to Osmania University & Approved by AICTE)
Nadergul, Hyderabad-501 510

CERTIFICATE
This is to certify that the project work entitled “SPATIO-TEMPORAL DYNAMICS
OF LAND COVER OF VISAKHAPATNAM USING CLOUD BASED
ARTIFICIAL INTELLIGENCE TECHNIQUES” is a Bonafide work carried out by

BANDARPALLY ANJALI 2451-20-732-014

MIDDE AKSHAY 2451-20-732-032
PERI VENKAT VISHWANATH 2451-20-732-043
SARIGARI NITHIN SAI 2451-20-732-044

in partial fulfillment for the award of BACHELOR OF ENGINEERING IN CIVIL

ENGINEERING by the OSMANIA UNIVERSITY, Hyderabad, under our guidance
and supervision.
The results embodied in this report have not been submitted to any other university or
institute for the award of any degree or diploma.

Dr. V. Shiva Chandra Mrs. K. Swetha Dr. M. Kameswara Rao

Assistant Professor Assistant professor Head of the Department
Department of Civil Department of Civil Department of Civil
Engineering, MVSREC Engineering, MVSREC Engineering, MVSREC

i
 DECLARATION

This is to certify that the work reported in the present project entitled “SPATIO-
TEMPORAL DYNAMICS OF LAND COVER OF VISAKHAPATNAM USING
CLOUD BASED ARTIFICIAL INTELLIGENCE TECHNIQUES” is a record of
work done by us in the Department of Civil Engineering, MVSR Engineering College,
Nadergul, Hyderabad-510 (Affiliated to Osmania University). The reports are based on
the project work done entirely by us and not copied from any other source.

BANDARPALLY ANJALI 2451-20-732-014

MIDDE AKSHAY 2451-20-732-032
PERI VENKAT VISHWANATH 2451-20-732-043
SARIGARI NITHIN SAI 2451-20-732-044

ii
 ACKNOWLEDGEMENT

I am profoundly grateful to MVSR Engineering College and the Civil Engineering

Department for providing me with the opportunity to work on this project. The
successful completion of this project would not have been possible without the support
and guidance of many individuals.

First and foremost, I would like to express my deepest gratitude to my project guides,
Dr. V. Shiva Chandra and Mrs. K. Swetha, for their invaluable guidance, constant
support, and encouragement throughout the course of this project. Their expertise and
insightful feedback were instrumental in shaping the direction and quality of this work.

I would also like to extend my sincere thanks to my professors and my Head of the
Department for their constructive criticism and helpful suggestions during the progress
reviews. Their feedback significantly enhanced the depth and quality of my research.

I am also grateful to my teammates for their support, collaboration, and the stimulating
discussions that enriched my understanding of the subject matter.

Finally, I would like to thank my family and friends for their unwavering support and
encouragement throughout this endeavor. Their belief in my abilities gave me the
strength and motivation to persevere and complete this project.

Thank you all for your contributions to this project.

Sincerely,

BANDARPALLY ANJALI 2451-20-732-014

MIDDE AKSHAY 2451-20-732-032
PERI VENKAT VISHWANATH 2451-20-732-043
SARIGARI NITHIN SAI 2451-20-732-044

iii
 TABLE OF CONTENTS

PAGE NO.

Certificate i
Declaration ii
Acknowledgement iii
List of Figures v
List of Tables v
Abstract vi
CHAPTER 1 INTRODUCTION 1–3
1.1 General 1
1.2 Impact of Urbanization on LULC 1
CHAPTER 2 LITERATURE REVIEW 4-7
CHAPTER 3 STUDY AREA 8
CHAPTER 4 METHODOLOGY 9 – 13
4.1 Overview 9
4.2 Data Collection 9
4.3 Data Preprocessing 10
4.4 Supervised Image Classification 10
4.5 Post Processing 12
CHAPTER 5 RESULT AND ANALYSIS 14 – 15
CHAPTER 6 CONCLUION 16
REFERENCES 17 - 18

iv
 LIST OF FIGURES

Figure 1.1 Trend of Urbanization in India and the World

Figure 4.1 Filter Data
Figure 4.2 Visualization Parameters
Figure 4.3 Creation of Training Data
Figure 4.4 Training the Classifier
Figure 4.5 Applying the Classifier
Figure 4.6 Accuracy Assessment
Figure 4.7 Exporting The Data
Figure 4.8 Google Earth Engine interface with 2024 sentinel image
Figure 5.1 LULC of the year 2019
Figure 5.2 LULC of the year 2024

LIST OF TABLES

Table 5.1 Data Collection

Table 6.1 Land use changes from 2019 to 2024

v
 ABSTRACT

Land use and land cover (LULC) classification is a type of geospatial data in a visual
format that categorizes the Earth’s surface into distinct classes like forest, Water body,
built up and more. Utilizing remote sensing and geospatial data for land cover change
detection helps in evaluating dynamics in the way that land is used or to assess how
climate change alters the Earth’s surface within specific regions. Remote sensing image
classification is crucial for a wide range of applications, including land cover mapping,
environmental monitoring, urban planning, disaster assessment, and resource
management.

Using the Google Earth Engine (GEE) cloud platform, the purpose of this research is
to analyze the changes in Land Use and Land Cover (LULC) transformations that have
taken place in Visakhapatnam District of Andhra Pradesh over the past five years (2019
–2024). Change detection of an LULC map is a method that examines shifts in LULC
throughout time. Sentinel 2A satellite image collections were utilized in this study. The
study area was divided into five LULC categories using the most likely classified
approach to quantify the changes over the aforementioned period. The results indicated
that between 2019 and 2024, Vegetation in the study area increased by about 27.36%.
Built-up and urban areas increased by 46.85% and Water bodies area has decreased by
47.22%. Similarly, there is a small decrease in the Agricultural area. These results
indicate fast urbanization in the area within a period of 5 years. With the help of this
study, decision-makers will be able to make choices that are appropriate in a given
situation. The findings emphasize the value of satellite in finding the Spatio-temporal
dynamics of land cover of Visakhapatnam using cloud based artificial intelligence
techniques.

vi
 CHAPTER 1

INTRODUCTION

1.1 GENERAL
Urbanization refers to the population shift from rural to urban areas. It is predominantly
the process by which towns and cities are formed and become larger as more people
begin living and working in central areas in search of better employment opportunities
generated by revolutions in information technology and other industries, leading to an
expansion of the cityscape with the development of better infrastructure and housing
(Sekertekin et al., 2016). The increase in urbanization is, in a way, aiding life in the
cities, but it has proved to be a boon as well as a bane.

Urbanization has resulted in a change in the natural landscape of the region since dense
forested areas, agricultural lands, and water bodies have been converted to
accommodate roads, buildings, and other concrete features that do not allow rainwater
to percolate. The natural balance is disturbed, leading to major changes in the physical
and surface characteristics of the region due to urbanization. The replacement of the
natural landscape by manmade features led to changes in the radioactive flux, thermal
capacity, reflectance, conductivity, and moisture of the land surface. (Shepherd, 2005).
There was an increase of 1.2% in urbanization in India during 2001-2011 when
compared to 1991-2001(Bhagat, 2011). By 2031, India’s urban population is predicted
to exceed 60 million people (Heilig, 2012). The urbanization trend in India and the
world is depicted in Figure 1.1

1.2 IMPACT OF URBANIZATION ON LULC

Rapid, spontaneous, and unsystematic urbanization has significant negative impacts on
the environment, especially on land and water. Changes in landscape due to rapid
urbanization often led to increased runoff, decreased groundwater recharge, urban
flooding, and increased surface temperatures. Variations in landscapes, changes in
regional climate, deforestation, degraded soil fertility, decreased air and water quality,
pollution, overcrowding, and an increased crime rate are a few problems associated
with unplanned rapid urbanization (Kantakumar et al., 2016; Kaya, 2007).

1
 Figure 1.1: Trend of Urbanization in India and the World
(Source: Equity Master Survey)

Analyzing the dynamics of land cover change using a conventional method of

surveying is a tedious process. The dynamics of changes on the earth can be better
understood in less time with the aid of Remote Sensing (RS) and Geographical
Information Systems (GIS) technology. Spatial and temporal mapping of urbanization
is an essential process for ensuring sustainable development. The accurate prediction
of changes in LULC resulting from urbanization is imperative for developing countries
such as India. This can be achieved through the continuous and systematic monitoring
of these changes using satellite imagery. Though high-resolution satellite imagery
provides more accuracy, it is not feasible to work on larger areas due to the

high cost. So, many researchers use medium-resolution satellite images, which are
freely available, but the accuracy depends on the number of training samples and human
interpretation, which is a major challenge. In the early days, researchers mainly used
the maximum likelihood classification technique (Tso and Mather, 2001) for the
preparation of LULC maps, but with the advent of the latest techniques, such as
machine learning and deep learning, and increased computing facilities, researchers are
using various machine learning algorithms like SVM, Random Forests (RF), K-Nearest
Neighbour (K-NN), and Classification and Regression Tree (CART) (Rahman et al.,
2017).

2
In addition to the above techniques, researchers developed spectral indices, which are
a combination of the spectral reflectance of two or more band wavelengths and
represent the relative abundance of features of interest. Normalized Difference Built-
up Index (Zha et al., 2003), Built-up Index (He et al., 2010), Modified Built-up Index
(Prasomsup et al., 2020), Enhanced Built-up and Bareness Index (As-Syakur et al.,
2012), SwiRedIndex(Capolupo et al., 2020), Index-based built-up index ( Xu, 2008),
and Built-up area extraction index (Bouzekri et al., 2015) are few indices that are used
to extract land cover of interest.

Patra et al. (2018) analyzed the LULC changes of Howrah Municipal Corporation
during 1975-2015 and concluded that the built-up area has increased by 58.16% at the
cost of other land cover classes. Nath et al. (2021) analyzed the impact of urbanization
on LULC of Guwahati City and reported that the built-up area has increased by 1-3%
during the last 30 years. Debsarma et al. (2023) analyzed the changes in LULC of
Gandhi Nagar district in Gujarat during 1991-2021 and the results revealed an increase
in built-up area of 234% at the cost of other land cover classes. Understanding the
dynamics of LULC change at a profound level due to urbanization is thus necessary for
facilitating sustainability, as increased urban areas are leading to various environmental
concerns. In view of this, the present research is aimed at examining the impact of
urbanization on changes in LULC in Visakhapatnam city of Andhra Pradesh during
2019-2024.

3
 CHAPTER 2

LITERATURE REVIEW

Applications of Remote Sensing and Machine Learning in LULC Classification

The satellite image classification process can be used to extract LULC. The
classification process involves categorizing all satellite imagery pixels into various
LULC themes. The classified image data is used to create LULC thematic maps of the
area of interest. The LULC extraction process requires multispectral data, which
provides spectral information. The typical LULC extraction techniques are divided into
three different classification categories: supervised classification, unsupervised
classification, and hybrid classification, with each having its own merits and demerits.
MLC, SVM, and Random Forests (RF) are a few supervised classification techniques,
and K-means clustering, Iterative Self-Organizing Data Analysis (ISODATA)
clustering are examples of unsupervised classification. A Literature review of some of
the methods mentioned above is presented here.

Savitha and Talari (2023) generated cropland extent maps using RF and SVM
techniques for the Tadepalligudem watershed and concluded that SVM outperformed
RF classification with an accuracy of 94.3%.

Basheer et al., (2022) generated LULC maps of Charlottetown town in Canada using
SVM, Random Forest, and MLC techniques and concluded that SVM performed better
with 85% accuracy as compared to Random Forest and MLC techniques.

Nyamekye et al., (2021) generated LULC maps of three regions in Ghana for 1986,
2002, 2010, and 2020 using a support vector machine classifier and reported that the
overall accuracy was more than 90% in all the cases and the kappa coefficient was more
than 0.86, which indicates good classification results.

Dabija et al., (2021) did a comparison of SVM and Random Forests for Corine land
cover mapping and concluded that the SVM algorithm with the Radial Kernel Function
(SVM RBF) achieved higher overall accuracy results (6–7%) than Random Forest.

4
Prabuddh et al., (2020) analyzed the temporal changes in a watershed in the Sikkim
Himalayan region using LULC maps prepared by the Maximum Likelihood
Classification (MLC) technique. They achieved an overall classification accuracy of
85.67% with a kappa statistic (k) of 0.8117, and they emphasized the significance of
environmentally responsible land management practices in relation to policy
implications.

Hu et al., (2020) generated LULC maps for a highly urbanized Central Beijing area for
1984, 1999, 2009, and 2019 using a support vector machine and achieved an overall
accuracy of more than 90% and a kappa coefficient of more than 0.887.

Jia et al., (2019) investigated the capability of high-resolution data obtained from the
Chinese civil satellite GF-2 by extracting LULC from Shandong Province, China. To
classify GF-2 data, they used MLC and SVM classification techniques to perform pixel-
and object-based LULC extraction. The process generated LULC maps for P-MLC, P-
SVM, O-MLC, and O-SVM strategies. The corresponding confusion matrices were
determined. According to the results, the O-SVM classifier had the highest LULC
extraction accuracy.

Jamsran et al., (2019) generated LULC maps of the UVS Lake basin in Mongolia
using SVM and MLC techniques and reported that SVM achieved an accuracy of 92%
and a kappa coefficient of 0.89 as compared to MLC with an accuracy of 87% and
kappa value of 0.83.

Shubham et al., (2018) highlighted the serious consequences of urbanisation on

changes in LULC due to population growth in megacities. They used a maximum
likelihood supervised classification method to identify changes and found that the
region's basin hydrology and water security had been affected by the region's
unprecedented growth.

The importance of RS technologies in detecting LULC changes was emphasised by

Tesfa et al., (2018), who employed the MLC tool to generate classified maps and
assessed their accuracy. The findings of their study revealed significant changes,
particularly an increase in cultivable land of nearly 13.78% over an eleven-year period
due to population growth.

5
Huang et al., (2018) proposed a method for automatic extraction of impervious
surfaces from high-resolution remote sensing images based on deep learning techniques.
(AEISHIDL), A Convolutional Neural Network (CNN) is used to calculate CNN
feature averages of all pixel neighbourhoods in each object. Finally, the Fuzzy C-Means
(FCM)clustering algorithm is employed to extract the impervious features. The
experimental results showed that AEISHIDL has higher accuracy compared to other
techniques.

Sophia and Ndambuki (2017) generated LULC maps using the MLC technique in
their study area in Limpopo Province. The classification efficiency of the maps
produced was quantitatively assessed using the kappa statistic. Their study yielded an
overall classification accuracy of 81.7% and a kappa statistic (k) of 0.722, both of which
were within acceptable limits.

Shravya and Sridhar (2017) highlighted the significance of geospatial technologies

for efficient land use planning in the context of future research. Using GIS techniques,
they examined the changes in land cover in their proposed study area over the past
twenty-five years and produced comprehensive LULC maps highlighting the temporal
changes that occurred over time.

Deilmai et al., (2014) used SVM and MLC classification techniques to find the best
and most accurate method for land use classification and found that SVM is more
accurate than MLC. They stated that MLC cannot handle complex images, so many
pixels cannot be correctly classified, but SVM can overcome the image's mixed-pixel
problem.

Taati et al., (2014) created a land use map using SVM and MLC classifiers from
Landsat 5 TM images in Qazvin, Iran. The overall accuracy and kappa coefficient were
used to compare the accuracy of the two algorithms. The evaluation results confirmed
that the SVM algorithm outperforms the MLC algorithm in land use mapping, with an
overall accuracy of 86.67% and a kappa coefficient of 0.82. As a result of its higher
accuracy and consistency, this algorithm has been proposed as an optimal classifier for
the extraction of land use maps.

The convergence of remote sensing and machine learning techniques has transformed
the field of land use and land cover (LULC) classification, providing unprecedented
insights into landscape dynamics and environment monitoring. The combination of
6
remote sensing and machine learning not only improves the efficiency and accuracy of
LULC classification, but also enables the detection of subtle changes over time,
supports large-scale environmental assessments, and facilitates more effective
decision-making for sustainable land management and conservation practices. This
integration leverages the strengths of high-resolution satellite imagery and advanced
algorithms, making it possible to analyze vast amounts of data with greater precision
and speed than traditional methods.

7
 CHAPTER 3

STUDY AREA

Visakhapatnam, often referred to as Vizag, is a prominent coastal city in the Indian state
of Andhra Pradesh is taken as the study area. It is situated on the east coast of India,
along the Bay of Bengal. Geographical coordinates of this study area have 17.6868° N
latitude and 83.2185° E longitude. Key geographical features include the Simhachalam
Hills and the Eastern Ghats. Visakhapatnam experiences a tropical savanna climate.
Summers are hot and humid, with temperatures often exceeding 35°C (95°F). Winters
are mild, with temperatures ranging between 15°C (59°F) and 25°C (77°F). The
monsoon season occurs between June and September. Rapid industrialization and
urbanization have led to environmental issues such as air and water pollution,
deforestation, and habitat loss. The coastline is vulnerable to erosion and extreme
weather events like cyclones. The city has well-developed infrastructure, including an
international airport, extensive road and rail networks, and modern healthcare facilities.
While the city has witnessed substantial growth and development especially between
the years 2019 and 2024, it also faces challenges related to environmental sustainability
and urban management. Strategic planning and conservation efforts are crucial to
maintaining the delicate balance between development and ecological preservation in
Visakhapatnam.

8
 CHAPTER 4

METHODOLOGY

4.1 OVERVIEW

To study the change detection of Visakhapatnam's LULC, Sentinel-2A satellite images

have been used. Images from 2019 and 2024 have been acquired with 10-meter spatial
resolution from Google Earth Engine for the areas of Gajuwaka, Bheemunipatnam,
Visakhapatnam Rural and Visakhapatnam Urban. Later the satellite images are
classified into Five classes: Built-up, Waterbody, Vegetation, Agriculture, and Barren
land using Support Vector Machine Algorithm in Google Earth Engine platform. Firstly,
geometry is defined and imported under feature collection and with ‘Class’ as property.
75 Training samples of each feature Class are marked on the loaded Sentiel-2A satellite
images of each year and are fed as input to the code for training. 80 % of the samples
are taken for training the model and 20% of the samples are taken for validation of the
model. Later the Confusion Matrix, Producer Accuracy, User Accuracy, Overall
Accuracy and Kappa Coefficient are calculated.

4.2 DATA COLLECTION

In this present study, sentinel satellite images of the Visakhapatnam region were
acquired from Google Earth Engine for the years 2019 and 2024 with a cloudy pixel
percentage of 20 and scale: 10. Here bands ‘B2’, ‘B3’, ‘B4’, and ‘B8’ in sentinel (Blue,
Green, Red, NIR) of Copernicus data are used for loading the satellite images.

Table 4.1: Data collection

Satellite Acquisition Year Spatial Source

Data Resolution

Sentinel-2A 2019-01-01 to 2019-12-31 10m Google Earth Engine

Sentinel-2A 2024-01-01 to 2024-03-31 10m Google Earth Engine

9
4.3 DATA PREPROCESSING

4.3.1 Filter Data and Cloud Masking: Apply spatial and temporal filters to focus on
your area and period of interest.

Figure 4.1: Filter data

4.3.2 Set Visualization Parameter:

The are a few ee.Image methods that produce RGB visual representations of image data,
for example: visualize(), getThumbURL(), getMap(), getMapId() and, Map.addLayer().
By default, these methods assign the first three bands to red, green and blue,
respectively. The default stretch is based on the type of data in the bands. To achieve
desired visualization effects, you can provide visualization parameters such as bands,
min, max, gain, bias, gamma, format etc.

Figure 4.2: Visualization Parameters

4.4 SUPERVISED IMAGE CLASSIFICATION

Training data is instrumental to supervised image classification. The training dataset is

a labeled set of data that is used to inform or “train” a classifier. The trained classifier
can then be applied to new data to create a classification.

The process for creating, collecting and the training data in Google Earth Engine is
detailed in the steps below.

4.4.1 Create Training Data

Firstly, a new feature class is created for each required land cover and the training data
is stored there. Here, the 5 feature classes which are imported are Built-up, Waterbody,
Vegetation, Agriculture and Barren land. These are labeled as Class 1 to Class 5
respectively. Based on these labels, the classifier can predict the most likely land cover

10
class for each pixel in an image. From the loaded Sentinel image, the training data
points are manually marked and defined as different feature classes using point marker.

The next step is to merge each land cover feature class into one final feature class with
all land covers aggregated.

Figure 4.3: Creation of Training Data

4.4.2 Training the Classifier

The Classifier package handles supervised classification by traditional ML algorithms

running in Earth Engine. These classifiers include CART, Random Forest, Naive Bayes
and SVM. The general workflow for classification is:

1. Collect training data.

2. Instantiate a classifier. Set its parameters if necessary.

3. Train the classifier using the training data.

4. Classify an image or feature collection.

5. Estimate classification error with independent validation data.

Figure 4.4: Training the Classifier

11
4.4.3 Applying the classifier

Figure 4.5: Applying the Classifier

4.5 POST-PROCESSING

Post processing involves Accuracy Assessment of the classification model, exporting

data to google drive for LULC change analysis and lastly, change detection monitoring
using Q-GIS.

4.5.1 Accuracy Assessment

Accuracy assessment is a crucial step in evaluating the performance of your land cover
classification model in Google Earth Engine (GEE). This involves comparing the
classified image against ground truth data to compute metrics such as overall accuracy,
producer’s accuracy, user’s accuracy, and the kappa coefficient.

Figure 4.6: Accuracy Assessment

12
4.5.2 Exporting Data

The results and classified image dataset are exported and saved to Google Drive for
further analysis if any. In this case the dataset is used to analyze spatio- temporal
dynamics of LULC of Visakhapatnam with the help of Q- GIS software to produce
the final LULC maps.

Figure 4.7: Exporting Data

Figure 4.8: Google Earth Engine interface with 2024 Sentinel image

Figure 4.8 is the snapshot of the Google Earth Engine (GEE) interface with our study’s
region of interest on which we have worked on to load sentinel images for supervised
image classification. Leveraging Google's cloud infrastructure, GEE can handle large
datasets and complex computations efficiently, making it possible to perform
supervised classification over extensive areas and long time periods.

13
 CHAPTER 5

RESULT AND ANALYSIS

Land use change from 2019 to 2024:

Table 5.1 shows the tremendous impact on land use composition and configuration.
Urbanization and comprehensive development have significantly contributed to the
rapid land use change in this region. It is apparent that large built-up areas have replaced
Water bodies within the study period. Similarly, the other urban land uses such as
commercial, industry, and transportation increased with different scales. Overall, all the
land uses have different trends of changes. Built-up areas increased whereas others
decreased, removing the other land uses. The accuracy of the LULC classification is
done using overall accuracy, and kappa coefficient. The overall accuracy obtained was
88% and 91% during 2019 and 2024 respectively, and the kappa coefficient was 0.86,
and 0.89 respectively.

Table 5.1: Land use changes from 2019 to 2024

Spatiotemporal Change Analysis:

The LULC maps, area statistics, and change in area are shown in Figure 5.1 and 5.2
and Table 5.1. During the study period, we observed an uneven shift in land use due
to rapid urban expansion. There was an increase in vegetation from 175.8 km2 to
223.9 km2.In contrast, Barren land decreased from 86.166 km2 to 18.2565 km2, which
is 78.8% decrease of barren land. Similarly, water faced a decrease from 55.6159 km2
to 29.3559 km2.We can see that Water Body has decreased by almost 50 % (47.21 %).
Another major difference can be seen in the Built-up area, and it has decreased by

14
46.85% in 5 years. Agricultural lands have been cut down to 51.3756 km2 which was
previously having an area of 60.878 km2 during 2019–2024. LULC of the years 2019
and 2024 are shown in Figures 5.1 and 5.2.

Figure 5.1: LULC of the year 2019

Figure 5.2: LULC of the year 2024

15
 CHAPTER 6

CONCLUSION

In this study, we analyzed spatiotemporal LULC changes based on temporal Sentinel

data. Barren land and Water bodies decreased because of fast urban growth. From the
analysis we can deduce that barren land has been converted to built-up area and
vegetation, and even waterbodies are replaced by built-up areas. Similarly Agricultural
area has urbanized with various built- up projects. Therefore, most of the other features
in 2019 are replaced by built- up areas in 2024, indicating fast urbanization and growing
cities. In some areas we noticed that barren lands have been transformed to vegetation
and in some places vegetation itself is turned into built- up. And in the case of
Waterbodies, heavy encroachment can be observed, which leads to urban sprawl. These
factors exacerbate the difficulties associated with sustaining regional development and
environmental conservation. Additionally, agriculture and development policies can be
linked to promote sustainable urbanization. It is advised that future studies use more
factors and data to investigate their effects on landscape patterns. Changes in LULC,
particularly the increase in impervious surfaces (Built- up), have contributed to altered
hydrological patterns, including reduced groundwater recharge and increased surface
runoff, posing challenges for water resource management. The expansion of urban areas
has spurred economic growth and development, providing improved infrastructure,
services, and employment opportunities. However, it has also resulted in the
displacement of rural communities, loss of agricultural lands, and increased pressure
on existing urban infrastructure and services. There is a critical need for implementing
sustainable land management practices to balance development needs with
environmental conservation. Strategies should include promoting reforestation,
protecting natural habitats, and encouraging sustainable agriculture. In conclusion, the
spatiotemporal analysis of LULC changes has highlighted the profound and far-
reaching impacts of human activities on the landscape. It underscores the importance
of adopting a holistic and integrated approach to land management that considers the
ecological, social, and economic dimensions of land use. By implementing sustainable
practices and policies, it is possible to mitigate negative impacts, enhance resilience to
16
climate change, and promote a more sustainable and harmonious relationship between
human development and the natural environment.

17
 REFERENCES

1. Chirasmayee Savitha, Reshma Talari.(2023). Mapping cropland extent using

sentinel-2 datasets and machine learning algorithms for an agriculture
watershed, Smart Agricultural Technology, Volume 4,100193,ISSN 2772-3755.
https://doi.org/10.1016/j.atech.2023.100193
2. Basheer, Sana & Wang, Xiuquan (Xander) & Farooque, Aitazaz & Nawaz,
Rana & Liu, Kai & Adekanmbi, Toyin & Liu, Suqi. (2022). Comparison of
Land Use Land Cover Classifiers Using Different Satellite Imagery and
Machine Learning Techniques. Remote Sensing, 14.4978.10.3390/rs14194978
3. Clement Nyamekye, Samuel Kwofie, Emmanuel Agyapong, Samuel Anim
Ofosu, Richard Arthur, Linda Boamah Appiah. (2021). Integrating support
vector machine and cellular automata for modelling land cover change in the
tropical rainforest under equatorial climate in Ghana. Current Research in
Environmental Sustainability, Volume 3,100052,ISSN 2666-0490,
https://doi.org/10.1016/j.crsust.2021.100052
4. Dabija, A., Kluczek, M., Zagajewski, B., Raczko, E., Kycko, M., Al-Sulttani,
A.H., Tardà, A., Pineda, L., Corbera, J. (2021). Comparison of Support Vector
Machines and Random Forests for Corine Land Cover Mapping. Remote
Sens, 13, 777. https://doi.org/10.3390/rs13040777
5. Prabuddh Kumar Mishra, Aman Rai, and Suresh Chand Rai. (2020). Land use
and land cover change detection using geospatial techniques in the Sikkim
Himalaya, India. Egypt. J. Remote Sensing Space Sci, 23, 133–143.
https://doi.org/10.1016/j.ejrs.2019.02.001
6. Hu, S., Fan, Y., & Zhang, T. (2020). Assessing the effect of land use change on
surface runoff in a rapidly urbanized city: A case study of the central area of
Beijing. Land, 9(1), 17.
7. Buyan-Erdene Jamsran, Chinsu Lin, Ishgaldan Byambakhuu, Jamsran Raash,
Khaulenbek Akhmadi. (2019). Applying a support vector model to assess land
cover changes in the UVS Lake Basin ecoregion in Mongolia, Information
Processing in Agriculture, Volume 6, Issue 1,158-169, ISSN 2214-3173.
https://doi.org/10.1016/j.inpa.2018.07.007

18
8. Subham Mukherjee, Wiebke Bebermeier, and Brigitta Schütt. (2018). An
Overview of the Impacts of Land Use Land Cover Changes (1980–2014) on
Urban Water Security of Kolkata. Land, 7(91), 1-25.
https://doi.org/10.3390/land7030091
9. Tesfa Gebrie Andualem, Gizew Belay, and Adebabay Guadie. (2018). Land Use
Change Detection Using Remote Sensing Technology. Journal of Earth Science
& Climatic Change, 9(10), 1-6. DOI: 10.4172/2157-7617.1000496
10. F. Huang, Y. Yu, T. Feng. (2018). Automatic Extraction of Impervious Surfaces
from High Resolution Remote Sensing Images Based on Deep Learning. J. Vis.
Commun. Image R. https://doi.org/ 10.1016/j.jvcir.2018.11.041
11. Sophia S. Rwanga, and J. M. Ndambuki. (2017). Accuracy Assessment of Land
Use/Land Cover Classification Using Remote Sensing and GIS. International
Journal of Geosciences, 8, 611-622. DOI: 10.4236/ijg.2017.84033
12. Shravya. S, and Sridhar. P. (2017).Land Use and Land Cover Change Detection
for Delhi Region through Remote Sensing Approach. International Journal for
Scientific Research & Development (IJSRD), 4(11), 54 - 57.
13. Deilmai, B. R., Ahmad, B. B., & Zabihi, H. (2014). Comparison of two
classification methods (MLC and SVM) to extract land use and land cover in
Johor Malaysia. IOP Conference Series: Earth and Environmental Science.
https:// doi. org/ 10. 1088/ 1755- 1315/ 20/1/ 012052
14. Taati, A., Sarmadian, F., Mousavi, A., Pour, C. T. H., & Shahir, A. H. E. (2014).
Land Use Classification using Support Vector Machine and Maximum
Likelihood Algorithms by Landsat 5 TM Images. Walailak Journal of Science
and Technology (WJST), 12(8), 681–687.