© 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.

ORG

UNDERSTANDING THE CHANGE IN LAND

PATTERN USING SUPPORT VECTOR
MACHINE
R Aditya1 Nikhil J P1 Hariharan C1 Rakshitha S1 Sachin R Kulkarni2
1Student 2Assistant Professor
1,2
Department of Civil Engineering
1,2
RNS Institute of Technology, Bengaluru, Karnataka, India

ABSTRACT: This study investigates land use changes in Bangalore, India, from 2013 to 2028 using Landsat imagery and Support
Vector Machine(SVM) classification. Significant shifts are observed in both rural and urban areas. Rural regions experience a drastic
reduction in forest cover, while urban areas witness a decline in water bodies and a rise in built-up areas. These findings highlight the
pressing need for sustainable land management practices to address environmental degradation and urbanization's rapid pace.
Graphical representations visually underscore these changes, emphasizing the urgency of conservation efforts and policy
interventions for a balanced and resilient future. This research underscores the importance of deep learning techniques in
understanding land use dynamics and informing evidence-based decision-making for environmental management and sustainable
development.
Keywords: Landuse Landcover, SVM, Land Management,

1. INTRODUCTION
Nowadays, Remote Sensing (RS) plays an important role in providing rich data for various applications. It has now become an
important part of achieving lofty goals and many goals, such as long-term climate studies, analysis of population evolution, and some
even sit first, with damage prevention [1]. Emery and Camps [2] reported that our ability to view the Earth from low Earth orbit and
geostationary satellites continues to improve. This development requires significant changes in the way we use and manage remote
sensing. One of the main tasks of remote sensing is location classification. Cheng et al. [3] defined classification as the division of
the far-field image into distinct land cover and land use group elements. Scene classification is an important part of remote sensing,
which is important for many remote sensing applications such as urban planning [ 4 ], land management [ 5 ], and fire characteristics
[ 6, 7 ]. Land use and land cover (LULC) classification has emerged as a major challenge in the field of remote sensing to better
manage the world's land for various purposes. To overcome this challenge. This advanced technology can provide insight for decision-
making and land planning by accurately and efficiently classifying land features. Essentially, LULC classification involves classifying
and identifying various aspects of the Earth's surface from satellite data. By understanding these characteristics, we can make
informed decisions about how to use and protect our land resources. Zhou et al. [8] proposed a location-CNN for location recognition;
recognition accuracy is best for the design model. Appears to provide new ideas for solving scene recognition problems. But source-
CNN needs more data and more time for training. Donahue [9] proposed a DeCAF network that is different from traditional deep
networks. It has a wide range of tools and can be used for many visual experiences without the need for extensive training. As the
amount of information increases, remote-sensing images introduce two major problems in location recognition: visual inconsistencies
and unclear labels.
2. LAND COVER ANALYSIS USING GIS
Land analysis using geographic information systems (GIS) involves analyzing and classifying the earth's surface according to physical
and land use. GIS uses spatial data to map, describe, and interpret different land cover types and provides information on
environmental management, urban planning, and resource conservation. Classification of land cover is important because it provides
information that can be used for modelling, especially models related to the environment such as climate change and development
treatment law. Therefore, mixed land use and land cover provide an integrated approach to understanding the interplay between the
biophysical world, social practices, and interactions [12]. It is used with geographic information system (GIS) technology. Distance
measurement forms the basis of many thematic data important to GIS analysis, including land use data and terrain features. Aerial
and Landsat images are also frequently used to assess land cover distribution and update existing geographic features. With the
introduction of remote sensing systems and image processing software, the importance of remote sensing in geographic information
systems (GIS) has increased greatly [13]. The accuracy of the terrain model depends on the interpolation technology. Land is scarce
in coastal cities like Mumbai as they are surrounded by the sea. As urbanization accelerates, land use and land use are rapidly changing
into hydrological processes. When heavy rains and major floods occur simultaneously, existing rivers will lose their capacity and

IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e533

c533
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG
cause floods [16]. A proper soil analysis forms the basis for choosing the best type of use for farmers. It is designed to evaluate the
physical potential of land use in climate, soil type, physical geography, and geological conditions.

3. DEEP LEARNING/ MACHINE LEARNING

Machine learning is a popular intelligence science of the 21st century. It is a method, method, and research that uses data, observations,
and information to model and describe computer-based learning processes. Machine learning is generally divided into three
categories: supervised learning, unsupervised learning, and incremental learning [17]. Supervised learning focuses on classifying or
predicting features based on a set of unknown data. Unsupervised learning focuses on grouping objects based on their similarities.
Learning support focuses on encouraging change for performance and improvement by learning from mistakes. Distributed tracking
is an open-source research that uses deep ML models on RS images [18]. Machine learning uses general-purpose algorithms to extract
hidden information from data. Data is the engine, fuel, and process of machine learning algorithms. Therefore, machine learning can
be considered as knowledge management [19]. Feature extractors and classifiers build deep-learning models for classification. To
achieve a better classification, researchers often use the state model directly or improve the results by adding prior information [21].
Multiscale feature extractors, rotation invariance extractors, and hand-generated features have been proposed to obtain data-driven
features [22]. However, this method faces many problems [23]. Unfortunately, low signal-to-noise ratio (SNR) and the distance of
different types of objects (such as the same objects in different directions or different objects with good image quality) worsen the
problem. Some researchers believe the problem stems from insufficient data.

4. SVM IN LAND COVER DETECTION

Support Vector Machines (SVMs) are instrumental in the realm of land cover detection, providing a robust and effective means of
classifying and mapping diverse land cover types within satellite or aerial imagery. The journey begins with data acquisition,
obtaining high-resolution imagery of the target area alongside ground truth data that offers labelled information about various land
cover classes. SVM has emerged as one of the most popular and useful techniques for data classification [32]. The objective of SVM
is to produce a model that predicts the target value of data instances in the testing set in which only attributes are given. The
classification goal in SVM is to separate the two classes by means of a function devised from available data and thereby to produce
a classifier that will work well on further unseen data [33]. Vapnik proposed a statistical learning theory-based machine learning
method which is known as a Support vector machine (SVM) [24]. SVM is considered one of the most important and simple methods
for solving problems related to classification [25] and learning and prediction [26]. Support vectors are the data points closest to the
decision point [27]. It performs the classification of data vectors according to hyperplanes in a wide space [28]. The simplest
classification problem is [29]. The maximum margin classifier is used to find hyperplanes with maximum margins in real-world
problems [30]. The main advantage of SVM is its ability to solve a variety of classification problems, including large-scale and
nonlinear problems. One of the disadvantages of SVM is that it requires correct tuning of many parameters to obtain good
classification results [31].
5. STUDY AREA:
Bengaluru is located in the southeastern part of the state of Karnataka, India. It lies at an average elevation of 920 meters (3,020 feet)
above sea level on the Deccan Plateau. The city is characterized by undulating terrain with several small hills and valleys. Bengaluru
is surrounded by the Eastern Ghats to the west and the flat plains of the Bay of Bengal to the east. Bengaluru experiences a tropical
savanna climate characterized by mild temperatures throughout the year. The city enjoys a moderate climate due to its high elevation.
Summers, from March, are warm with temperatures ranging from 25°C to 35°C (77°F to 95°F). Monsoon season begins in June and
lasts until September, bringing heavy rainfall. Winters, from December to February, are cool and pleasant, with temperatures
averaging around 15°C to 20°C(59°F to68°F). Studying Land Use and Land Cover (LULC) in Bengaluru is crucial for managing
rapid urbanization, assessing environmental impacts, optimizing resource allocation, and enhancing climate resilience. It provides
essential insights for sustainable development, informed decision-making, and safeguarding the city's ecological balance amidst
dynamic socio-economic changes.

IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e534

c534
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG
Figure 01: (A) India, (B) State Of Karnataka, (C) Bengaluru Rural, (D) Bengaluru Urban
6. METHODOLOGY
Here's a condensed point-by-point analysis of the methodology:
1) Define research goals and boundaries. Consider geographical scope, data accessibility, and representativeness. Ensure the
selected area mirrors broader regional characteristics.
2) Use Landsat 8 and 9 satellite imagery for LULC classification in Bengaluru.
3) Preprocess data to rectify distortions and enhance visual clarity.
4) Customize algorithms for tasks like image classification or change detection. Select algorithms based on research objectives
and data characteristics.
5) Utilize Support Vector Machines (SVM) for LULC classification, separating classes based on spectral information. Enable
accurate mapping of Bengaluru's landscape for urban planning and environmental monitoring.
6) Merge spatial datasets like satellite imagery, GIS layers, and socio-economic data.
7) Scrutinize land cover data to discern patterns and attributes.
8) Compare land cover datasets and methodologies to identify discrepancies and improvements.
7. RESULTS AND DISCUSSION
This research investigates the correlation between urban and rural changes in Bangalore, India, from 2013 to 2023. Utilizing Landsat
satellite imagery, Land Use and Land Cover (LULC) classification was conducted for the years 2013, 2018, and 2023, employing the
support vector machine(SVM)classification algorithm. Land use analysis: Land use analysis was carried out using remote sensing
data for 2013,2018,2023 and 2028. The results are given in the table. Here major portions we have considered are Barren Land, Water
Bodies, built-up areas, Forest Areas, and Agricultural Areas.

Figure 02 Figure 03

Figure 04
Figure 05
IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e535
c535
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG
Figure (02,03,04,05): Land Use Dynamics Of Bengaluru Rural

The Land use and land cover in the Bengaluru rural catchment after analysing the results reveals that the area under Forest has reduced
from 66.25% (2013) to 3.95% (2028).

Table 01: Land Use Dynamics of Bengaluru Rural

LULC 2013 2018 2023 2028

classes Area (sq. %Area Area (sq. km) %Area Area %Area Area (sq.Km) %Area
km) (sq.Km)
Barren Land 332.08 5.63 370.06 6.27 847.85 14.37 389.77 6.61
Water 412.14 6.99 220.04 3.73 116.25 1.97 274.25 4.65
Bodies
Built-up 388.04 6.58 518.22 8.78 2306.02 39.09 4196.30 71.13
Area
Forest Area 3908.77 66.25 3501.87 59.36 2123.20 35.99 233.00 3.95
Agricultural 858.75 14.56 1289.60 21.86 506.47 8.58 806.47 13.67
Area

Figure 06 Figure 07

Figure 08 Figure 09

IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e536

c536
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG

Figure 10
Figure (06,07,08,09,10): Graphical Representation Dynamics Of Bengaluru Rural

Figure 08 Figure 09

Figure 10 Figure 11
Figure (08,09,10,11): Land Use Dynamics Of Bengaluru Urban
The Land use and land cover in the Bengaluru Urban catchment after analysing the results reveals that the area under Water
IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e537
c537
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG
bodies has reduced from 5.80% (2013) to 3.66%(2028) and Built-uparearaised from 57.11% (2013) to 64.52% (2028)
Table 02: Land Use Dynamics Of Bengaluru Urban

Figure 12 Figure 13

Figure 14 Figure 15

2013 2018 2023 2028

LULC classes Area (sq. km) %Area Area %Area Area (sq. %Area Area %Area
(sq.km) km) (sq.km)
Barren Land 205.52 5.80 485.22 13.70 287.82 8.13 129.72 3.66
Water
Bodies 231.29 6.53 420.85 11.88 784.81 22.16 284.81 8.04
Built-up Area
2023.18 57.11 1524.90 43.05 1570.52 44.34 2285.52 64.52
Forest Area 304.50 8.60 488.19 13.78 455.63 12.86 198.63 5.61
Agricultural Area
777.82 21.96 623.14 17.59 443.53 12.52 643.53 18.17

Figure 16
Figure (12,13,14,15,16): Land Use Dynamics Of Bengaluru Urban

8. CONCLUSION
Understanding changes in land patterns through deep learning techniques is pivotal for advancing our understanding of land use
dynamics, supporting sustainable development goals, and facilitating evidence-based decision-making in environmental management.

IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e538

c538
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG
The ongoing research and technological innovations in this domain are critical for addressing emerging challenges and harnessing the
full potential of deep learning in land pattern analysis.
An analysis of Land Use and Land Cover (LULC) dynamics in Bengaluru, encompassing both rural and urban areas, reveals
significant transformations over the studied period. In rural regions, there's a stark reduction in forest cover, plummeting from 66.25%
in 2013 to a mere 3.95% in 2028, indicating alarming environmental degradation and habitat loss. Conversely, the urban landscape
shows a noticeable decline in water bodies, dropping from 5.80% to 3.66%, accompanied by a substantial increase in built-up areas,
rising from 57.11% to 64.52%. These shifts underscore the rapid pace of urbanization and infrastructure expansion, highlighting the
urgent need for sustainable land management practices.
Graphical representations vividly depict these changes, emphasizing the criticality of proactive measures to conserve natural
resources, mitigate environmental degradation, and promote sustainable urban development. Strategic planning and policy
interventions are imperative to balance developmental needs with ecological preservation, ensuring a resilient and livable future for
Bengaluru's residents.

9. REFERENCES
1. A. Bharath, M. Manjunatha, T. V. Reshma, Ranjitha B. Tangadagi, and Sifatullah Bahij, “Evaluation of Land Use/Land
Cover Changes due to Urban Sprawl in Bengaluru Rural, Karnataka, India”. “Discrete Dynamics in Nature and Society”, “Vol.2023,
“15 Pages”.
2. A Venukumar, Dr. B Laxmaiah, “Role of GIS and Remote Sensing in Land Use and Land Cover Studies”, “January 2022”.
3. Abhishek M. J & Renukadevi.S, “Use and Land Cover (LULC) of Bangalore City - A Geospatial Approach”, “December
2021”.
4. Aishwarya S Hanjagi, P Surendra, Shahnin Irfan, Ashok D Hanjagi, “Urban growth and development in Mandya taluk using
GIS and remote Sensing”. “Geo Eye”, “Vol.8, “Issue 2”, “Page no.6-11”.
5. Aishwarya S Hanjagi, Shahnin Irfan, P Surendra, Ashok D Hanjagi, “Land use / Land cover changes in Mandya taluk using
GIS and Remote Sensing”. “Geo Eye”, “Vol.8, “Issue 1”, “Page no.1-4”.
6. Amrutha Mary Varkey, “An Analysis of LULC Change Detection using GIS; A Case Study of Peri-Urban Bengaluru”.
“International Journal of Science and Research”, “Vol.8, “Issue 1”, “January 2019”.
7. Asmita & Anirudh, “Re-Visiting Urban Water Bodies: A Case Study for Mir Alam Tank, Hyderabad”, “August 2021”.
8. Azra Jawaid, Devi Rahul K, “Revisiting Urban Commons: A Land Use Land Cover Analysis of Bengaluru City”. “August
2021”.
9. B. Aneesha Satya & M Shashi & Pratap Deva, “Future Land Use Land Cover Scenario Simulation Using Open-Source GIS
for the City of Warangal, Telangana, India”. “January 2020”.
10. Bharath Setturu, K.S. Rajan, T.V. Ramachandra, “Land Surface Temperature Responses to Land Use Land Cover
Dynamics”, “A SCITECHNOL Journal”, “Vol 1, “Issue 4”, “January 2019”.
11. C.B. Jagadeesh, H.S.Shivakumar Naiklal. Nagaraj Sitaram, “Dynamics of Rapid Urbanization of Bangalore and Its Impact
on Land Use/Land-Cover – A case study of the Vrishabhavathi sub-watershed”. “International Research Journal of Engineering and
Technology”, “Vol.02, “June 2015”.
12. Giorgos Mountrakis, Jungho Im, Caesar Ogole, “Support vector machines in remote sensing: A review”, “November 2010”.
13. Hashem Dadashpoor, Parviz Azizi, Mahdis Moghadasi, “Land Use Change, Urbanization, and Change in Landscape Pattern
in a Metropolitan Area”. “Journal of Environmental Sciences”, “Vol.655, “Page no.707-719”.
14. Holly Dehingia, Rima Rani Das, S. Abdul Rahaman, P. Surendra, Ashok D. Hanjagi “Decadal Transformation of Land Use
- Land Cover and Future Spatial Expansion in Banglore Metropolitan Region”, “The International Archives of the Photogrammetry,
Remote Sensing and Spatial Information Science”, “Vol.43-B3-2022,”, “Page no.6- 11”.
15. Krity Gautami, Nila Nair, “Use Land-Cover Change Analysis of Delhi”, “August 2021”. Understanding the Change in Land
Pattern for Bengaluru City. 50
16. Mohamed A.E. Abdul Rahman, A. Natarajan, Rajendra Hegde “Assessment of land suitability and capability by integrating
remote sensing and GIS for agriculture in Chamarajanagar district, Karnataka, India”. “The Egyptian Journal of Remote Sensing and
Space Sciences”, “Vol.19”, “Issue.1”, “Pages 125-141”.
17. P.E. Zope, T.I. Eldho, V. Jothiprakash, “Impacts of Land Use–Land Cover Change and Urbanization on Flooding: A Case
Study of Oshiwara River Basin in Mumbai, India”. “CATENA”, “Vol.145”, “Pages 142-154”.
18. Priyadarshini J. Shetty, Shashikala Gowda, Gururaja K.V. and Sudhira H. S, “Effect of Landscape Metrics on Varied Spatial
Extents of Bangalore, India”, “Asian Journal of Geoinformatics”, “Vol.12”, Issue.1”.
19. Ramachandra TV, Bharath H. Aithal, Durgappa D. Sanna, “Insights to urban dynamics through landscape spatial pattern
analysis”. “International Journal of Applied Earth Observation and Geoinformation”, “Vol.18, “Page no.329-343”.
20. Shobha N V, “i”, “International Journal of Scientific Research in Civil Engineering”, “Vol.7”, “Issue.1”, “January-February
2021”.
21. Shreyas Goswami, “A Study on The Urbanization of Bengaluru, India Between 2015 and 2020 Based on Sentinel-1/2 Data”,
“August 2021”.
22. Shruti Kanga, Gowhar Meraj, Brian Alan Johnson, Suraj Kumar Singh, Muhammed Naseef PV, Majid Farooq, Pankaj
Kumar, Asif Marazi, Netrananda Sahu, “Understanding the Linkage between Urban Growth and Land Surface Temperature - A Case
Study of Bangalore City, India”. “Journal of Remote Sensing”, “Vol.14, “Issue17”, “August 2022”.
23. Shweta Hiremath, D.K. Prabhu raj, B.P. Lakshmikantha, Surya Deb Chakraborty, “Land Use/Land Cover Change Analysis
of Bangalore Urban District and Its Impact on Land Surface Temperature”. “Remote Sensing and GIS for Environment with Special
Emphasis on Marine and Coastal Dynamics”, “December 2013”.
24. T. V Ramachandra, Bharath.H. Aithal, and S. Vinay, “Land Use Land Cover Dynamics in Rapidly Urbanising Landscape”.
“Journal of SCIT”, “Vol.13, “August 2013”.
25. U. Icdan Basha, U. Suresh, G. Sudarsana Raju, M. Rajasekhar, G. Veeraswamy, and E. Balaji, “Land Use and Landcover
Analysis Using Remote Sensing and GIS: A Case Study in Somavathi River, Anantapur District, Andhra Pradesh, India”. “An
International Quarterly Scientific Journal”, “Vol.17”, “Issue.3”, “January 2018”.
26. Vinay Prasad Mandal, Sufia Rehman, Raihan Ahmed, Md. Masroor, Pavan Kumar, Haroon Sajjad, “Land suitability
IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e539
c539
 © 2024 IJNRD | Volume 9, Issue 5 May 2024| ISSN: 2456-4184 | IJNRD.ORG
assessment for optimal cropping sequences in Katihar district of Bihar, India using GIS and AHP”, “Korean Spatial Information
Society2020”.
27. Abhirup Dikshit, Biswajeet Pradhan, and Abdullah M. Alamri, “Temporal HydrologicalDrought Index Forecasting for New
South Wales, Australia Using Machine Learning Approaches”, June 2020
28. Ava Vali, Sara Comai and Matteo Matteucci, “Deep Learning for Land Use and LandCoverClassification Based on
Hyperspectral and Multispectral Earth Observation data: AReview”, August 2020
29. Bushra Praveen, SwapanTalukdar, Shahfahad, Susanta Mahato, Jayanta Mondal, PriteeSharma, Abu Reza Md. Towfqul
Islam & Atiqur Rahman, “Analyzing trend and forecasting of rainfall changes in India using nonparametric and machine learning
approach”, September 2020
30. Eleonora Jonasova Parelius, “A Review of Deep-Learning Methods for Change Detection in Multispectral Remote Sensing
Images”, April 2023
31. Iftikhar Ali, Felix Greifeneder, Jelena Stamenkovic, Maxim Neumann andClaudiaNotarnicola, “Review of Machine
Learning Approaches for Biomass and Soil MoistureRetrievals from Remote Sensing Data”, December 2015.

IJNRD2405456 International Journal of Novel Research and Development (www.ijnrd.org) e540

c540