Crack detector on railway tracks with animal

disturbance alert
Ashwin S,Nishaan K S,Preethika N,Sivabalan S

Ashwin S is with the Department of Instrumentation & Control Engineering, Sri Krishna College of Technology, Coimbatore, Tamil Nadu, India 641042
(Phone: +91 9345631486; e-mail: [email protected]). (Corresponding Author)

Nishaan K S is with the Department of Instrumentation & Control Engineering, Sri Krishna College of Technology, Coimbatore, Tamil Nadu, India 641042
(Phone: +91 6383057791; e-mail: [email protected]). (Corresponding Author)

Preethika N is with the Department of Instrumentation & Control Engineering, Sri Krishna College of Technology, Coimbatore, Tamil Nadu, India 641042
(Phone: +919080567620; e-mail: [email protected]). (Corresponding Author)

Sivabalan S is with the Department of Instrumentation & Control Engineering, Sri Krishna College of Technology, Coimbatore, Tamil Nadu, India 641042
(Phone: +916383098446; e-mail: [email protected]). (Corresponding Author)

Abstract—Railway transport is the most cost- effective and derailments. The seasonal variations lead to the contraction and
best mode of passenger transportation in India, as well as for expansion of railway tracks, potentially causing cracks to
long-distance and suburban traffic. Railway track crossings and develop. In response to this, the proposed system utilizes
unseen cracks in railway tracks are the leading causes of railway sensors to identify these fractures and transmits the precise
accidents. Railway track inspection and monitoring are currently locations of the detected cracks using a GPS module. Notably,
performed manually, which is time-consuming and inaccurate a significant number of accidents tend to occur during train
due to the high possibility of human error.Railway infrastructure shunting at speeds below 45 km/h.. Incidents involving trains
plays a vital role in transportation, and ensuring its safety is
and animals or humans have garnered increased attention
paramount. This abstract introduces an innovative system
designed to enhance the monitoring and safety of railway tracks
recently. Each year, numerous animals meet unfortunate ends
by integrating crack detection technology with animal on train tracks, disrupting ecosystems and causing substantial
disturbance alerts. The proposed system utilizes advanced delays in railway operations. Similarly, accidents frequently
sensors and image processing techniques to detect cracks and occur at rail gates where motor vehicles cross train lines, often
structural abnormalities on railway tracks. In addition to crack due to inadequate object detection. The reliance on human
detection, the system incorporates animal disturbance alert judgment for detection introduces the possibility of occasional
functionality. Wildlife incursions onto railway tracks pose a errors, especially in timely horn signaling, leading to accidents.
significant safety risk, not only to the animals but also to train Therefore, there is a critical need for an automated solution to
operations. The system employs infrared sensors and machine accurately detect objects on train tracks and promptly alert the
learning algorithms to identify and alert railway authorities driver, mitigating the risk of accidents. The Integrated Railway
about the presence of animals on or near the tracks, allowing for Track Crack Detection System with Animal Disturbance Alert.
timely intervention to prevent potential collisions. The In India, where the railway network is the backbone of
integration of crack detection and animal disturbance alert commercial transport, ensuring the safety of tracks is
features into a unified system provides a holistic approach to paramount. This cutting-edge system brings together
railway track safety. Real-time data from the system can be sophisticated crack detection technology and animal
transmitted to a central monitoring station, enabling prompt disturbance alerts, utilizing state-of-the-art sensors and image
response to any identified issues. This innovation aims to processing techniques. The primary goal of this system is to
minimize the risk of accidents, improve overall railway safety, proactively identify and address potential threats such as track
and contribute to the efficient and reliable operation of rail fractures and the presence of animals on or near the tracks. By
networks.
combining these crucial functionalities, the system aims to
Keywords—Railways, Cracks, Disturbance, Sensors, Image provide a comprehensive solution to enhance railway safety
Processing and minimize the risk of accidents.In a landscape where current
systems may fall short due to persistent challenges, our
I. INTRODUCTION Integrated Railway Track Crack Detection System with Animal
Disturbance Alert emerges as a promising solution. It aspires to
In India, the majority of commercial transportation relies on contribute significantly to the overall safety and reliability of
the railway network. Any issues with the railways have the the railway infrastructure, safeguarding both lives and the
potential to cause significant harm to the economy, leading to efficiency of commercial transportation.
loss of life due to accidents.But India has world’s fourth largest
railway community, but It faces significant challenges like gate
crossing issues, fires, and unnoticed track problems, leading to PROBLEM STATEMENT

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

 In the existing system, techniques such as Image parameters. The researchers also suggest automating
processing, video transmission, Image segmentation,Machine the extraction of points forming the sleeper's
learning algorithms ,Deep learning and Magnetic field methods curvature, which wasn't fully achieved in this study
can identify the cracks on the railway tracks and animal due to the assistance needed by the method used
disturbance alert. One of the oldest ways to check the railway (snakes). Additionally, they recommend comparing
tracks involves people manually scanning all the necessary different defect detection techniques, emphasizing the
parts. Even though this method is commonly used in India, it potential for better results with a more extensive
often leads to poor results. A simpler approach is using a image database.
camera to keep a constant eye on the tracks and stream what it
3. Tanish Tallam et al(2023) proposed crack detection
sees. However, this method may miss small cracks and
in railway tracks using image processing and deep
disturbance and it can be quite expensive.
neural networks.The study introduces a technique
Objectives of the study based on image processing to detect fractures in
railway tracks. Instead of manually inspecting track
 To address the limitations of the current methods used to
segments, the method automates the process, allowing
find problematic railway tracks and obstacle detection
for the installation of video cameras at various points
 Integrating Railway Track Crack Detection System with along the track to capture images. These images are
Animal Disturbance Alert will gives better results then analyzed by the proposed system to identify
fractures, reducing the risk of accidents by promptly
 To lower the chance of people getting hurt or killed. detecting cracks as they appear. The autonomous
 To ensure safe journeys. nature of the system minimizes the need for operator
involvement, ensuring optimal efficiency. The project
demonstrates the Railway Crack Detection System,
employing a machine learning approach in the
II. LITERATURE REVIEW
experiment to accurately identify fractures on the
1. Y. Lavanya et al(2023) proposed Iot smart engine track surface. The inclusion of a GPS module
for detecting railway crack in this the new system enhances precision by providing the exact position of
combines the usual way of finding cracks with a the fracture. This method offers a quick and effective
creative approach using live video streaming and IoT. way to detect track fractures. By incorporating image
This entire setup is placed on a four-wheeler that
processing and machine learning models, the system
moves along the tracks. Compared to the current
can wirelessly transmit information about detected
system using IR transmitter and receiver, this new
fractures in the railway track to the authorities. This
method is a smart way that makes things easier for
advanced approach improves accuracy compared to
authorities and improves how accurately we detect
existing methods, contributing to enhanced railway
cracks. The checking process happens regularly to
safety.
find cracks and prevent accidents. The main goal is to
make sure we find faulty rails on time to stop train 3 T. Dedeepya et al(2023) proposed Iot based railway
accidents, saving lives and property. track crack detection and alert system. In this
project, they employ infrared sensors to detect
2. Taynara Martins Gonc¸alves et al(2023) proposed
fractures in railway tracks. Looking ahead, there's a
Digital Image Processing in the Diagnosis of
potential to utilize visual videos captured from the
Cracks in Steel Sheet.The study introduces a new
tracks through systems equipped with IP-based
method using pattern recognition and digital image
cameras. Enhancements like linked solar panels and
processing to automatically investigate defects in
Notification chatbot support could be incorporated for
railways, focusing on sleepers and rail maintenance.
improved visibility and interaction with the robot.
The main goal is to enhance the management of rail
This device is designed to identify faulty rail tracks
maintenance by analyzing and classifying sleepers.
without requiring human intervention. When
The innovative process involves extracting
compared to existing detection methods, the proposed
information from the curvature of the sleeper surface
system offers several advantages, including reduced
to optimize railway maintenance, ultimately
costs, lower power consumption, and quicker analysis
preventing accidents and reducing costs. The
times. The purpose of a railway track crack-detecting
classification system, based on the curvature of the
vehicle is to identify cracks or irregularities in the
sleeper surface, performed well in analyzing the upper
track that, if promptly repaired, can reduce the risk of
curvature, correctly identifying 94.24% of validation
train accidents. With this suggested approach,
data and detecting all defective sleepers. However,
damaged tracks can be easily located and repaired
when analyzing points below the sleeper, the average
promptly, potentially saving numerous lives.
V. Vivek et al(2023) proposed Towards the
accuracy for damaged sleepers was 81.82%. The
system is designed with a user-friendly interface, but 4
the study acknowledges the need for a larger development of obstacle detection in railway tracks
validation dataset to improve accuracy. Future work is using thermal imaging .This study introduces a
proposed with a larger database and varied training novel and efficient method for identifying railway
 tracks and obstacles. The approach involves breaking Intelligent Detection System (TIDS). The future
down thermal images into multiple information- direction of this technology aims for high reliability
carrying elements using 2D SSA (Singular Spectrum (RAMS). When the Fully Autonomous Operation
Analysis). The FR-CNN algorithm then utilizes each (FAO) line operates without a driver, the highly
element to recognize obstacles on railway lines. reliable SIL4 TIDS acts as the "eyes" of the train,
Additionally, the study focuses on using SSA for track detecting foreign objects entering the boundary in real
identification. By combining FR-CNN with 2D SSA, time and reporting to the signal system. Current
the aim is to create an advanced warning system for methods primarily concentrate on onboard obstacle
preventing rail accidents and enhancing rail safety. detection, but the train's braking distance is often too
The experimental results support the effectiveness of long. Future efforts should focus on developing
this proposed approach. Surface architectures are collaborative perception between vehicles and roads
highlighted for providing a balance between low to enhance obstacle detection distance. The study uses
computational load and excellent reliability, offering a convolutional neural network, which introduces
performance benefits. In future research, emphasis uncertainty, especially in dynamic environments.
should be placed on incorporating brightness images Future research aims to estimate and improve the
and stereo information from the dataset along with interpretability of the model by addressing this
heat information to further enhance performance uncertainty.
measures.
8 Jinyan Qu et al(2023) proposed Research on Railway
5 Alfored George Charles Prem et al(2023) proposed Obstacle Detection Method Based on Developed
Survey On Smart Siren to Alert Animals About Euclidean Clustering Safety is crucial in railroad
Train Tracks in Image Processing This study operations, and obstacles on the tracks pose a
highlights the severity of animal-train accidents, significant threat to traffic safety. To enhance the
which can impact safety, finances, and the accuracy and efficiency of detecting obstacles on
environment. Various technical solutions have been railways, we propose an effective method in this
proposed to prevent these collisions, but they have paper. Our approach utilizes the RANSAC algorithm
limitations and may not be universally effective. The to isolate the Region of Interest (ROI) and an
use of image processing, specifically object detection improved Euclidean clustering method to address
and picture segmentation, is explored for identifying issues like over-segmentation and under-
animals and providing early warning signals for segmentation. Experimental results indicate that this
approaching trains. One proposed solution involves a algorithm accurately segments the ROI, providing
smart siren system that generates tailored sound improved detection outcomes.Their method involves
signals to alert and deter animals from railway tracks. using the RANSAC algorithm to determine the rail
While both smart siren systems and image processing position after dividing the roadbed plane, followed by
technologies show promise in reducing animal-train segmenting the ROI based on the rail position. The
accidents, further research is necessary to enhance experimental procedure demonstrates the
their accuracy, reduce false alarms, and optimize their effectiveness of this segmentation process.They
effectiveness in different scenarios To successfully introduce an enhanced Euclidean clustering algorithm
implement these solutions and minimize the impact of for obstacle segmentation, which stands out for its
animal-train collisions, collaboration between wildlife adaptability with a variable distance threshold,
conservationists, transportation authorities, and allowing it to segment diverse point cloud data.
railway operators is essential. Comparative testing with other clustering algorithms
shows that, despite having a longer operation time, the
6 Qiang Zhang et al(2023) proposed Automatic improved Euclidean clustering algorithm achieves an
Obstacle Detection Method for the Train Based on 11% increase in correct segmentation rates.
Deep Learning This paper introduces a smart system Considering both operation time and accuracy, it
for detecting trains in urban rail settings. By using proves to be the superior choice. Testing the entire
various focal cameras, the system expands its
system with raw point cloud data reveals that the
coverage. Additionally, LiDAR is incorporated to
detection results are better when obstacles are closer
provide distance information and identify small
to the LiDAR, covering a larger area and having
obstacles not visible in images. The system utilizes
greater height. Moreover, our algorithm performs well
deep learning-based algorithms for object detection
in actual railroad scenes for obstacle detection. Future
and semantic segmentation to identify relevant areas
plans involve further development, particularly in
based on subway scene characteristics. The spatial
controls, to enhance accuracy and timeliness in
judgment algorithm ensures effective perception in
obstacle detection. Additionally, efforts will be made
rail scenarios, successfully detecting obstacles on the
to classify and identify detected obstacles, expanding
tracks in experiments.
the research scope to autonomous driving systems.
7 As technology for active obstacle detection gains
9 Istiak Mahmud et al(2023) proposed Advancing
wider acceptance, it's crucial to focus on the key
Wildlife Protection: Mask R-CNN for Rail Track
technologies and implementation methods of the Train
Identification and Unwanted Object Detection To
 prevent wildlife deaths and avoid train accidents on a cost-effective and efficient safety system designed
railway tracks, it is crucial to implement a smart to elevate railway transportation safety and service
automated system capable of detecting objects on the quality to the next level.
tracks and promptly notifying authorities. Utilizing
12 Fahim Ur Rahman et al(2022) proposed Real-Time
machine learning techniques for object detection on
Obstacle Detection Over Railway Track using
train lines can significantly enhance safety and
Deep Neural Networks They suggest a method to
operational efficiency. In this context, this research
detect obstacles on train tracks using a Convolutional
proposes a system based on mask-RCNN to identify
Neural Network called "MobileNetV2." We apply
the train line and detect unwanted objects on it.
transfer learning, using pre-trained weights of
Compared to previous approaches, our system
MobileNetV2, to reduce training time and resources.
demonstrates superior performance, with a specific
This model is efficient in detecting obstacles,
focus on wildlife safety. To achieve this, our dataset
providing a quick response to control centers. By
includes a higher representation of animals as objects.
The proposed system can serve as a smart and determining if a track crossing is obstacle-free, our
automated assistant for train operations, playing a system alerts trains beforehand, enhancing safety.
crucial role in minimizing accidents on railway tracks. Beyond signaling, our model can facilitate advanced
steps to remove obstacles. While collecting manual
10 M. Mustafa SARITAS et al(2023) proposed Railway data from rail tracks is feasible with handheld devices,
Track Fault Detection with ResNet Deep Learning using drone-shot images could improve accuracy
Models while In assessing the accuracy of various further. Their model, based on handheld images,
ResNet deep learning models, it is evident that all achieved 97.00% accuracy, which can improve with
models generally exhibit high accuracy. Notably, diverse perspectives during training. However, their
ResNet-101 stands out with the lowest accuracy, dataset only includes images from bright sunny days,
while ResNet-18 demonstrates the best performance. lacking bad weather conditions. Including such
The success of ResNet-18 is attributed to its superior images is crucial, as bad weather affects image clarity.
learning capability and effective transfer of precise Detecting obstacles accurately and sending timely
low-level features. In the evaluation of ResNet-18, signals benefits resource conservation and overall
ResNet-50, and ResNet-101 models on the Railway infrastructure safety. Future efforts will address these
Track Fault Detection dataset, performance metrics issues to enhance their system's performance
were employed. Overall, ResNet-18 displayed the accuracy.
highest sensitivity and accuracy on the dataset.
However, the ResNet-50 model outperformed in terms 13 Aaditya Aurobindo et al(2023) proposed
of specificity and precision metrics. These findings Development of a Robotic System for Non-
provide insights into the performance variations Destructive Crack Detection on Railway Tracks
among ResNet models of different depths, aiding in They have created a Robot to find cracks in railway
informed model selection. tracks, which is crucial for preventing serious
accidents. This robot uses machine learning and
11 Mr.Nissanka R.R.N.V et al(2022) proposed Train ultrasonic sensors to spot defects as tiny as 0.5mm and
Accident Prevention & Breakdown Detection using 0.3cm. In a test on a rail with similar dimensions to a
Machine Learning based Safety System This railway track, the robot performed well, showing
research focuses on improving railway safety through potential for similar experiments on actual tracks with
a Machine Learning-based Safety System designed to the right data. However, the study acknowledges that
prevent accidents and detect issues like elephant outside elements like dust and feathers can sometimes
collisions, drowsiness, cracks, and breakdowns. The cause the robot to incorrectly identify defects. The
system, utilizing ESP32 and Neo 6m modules for data angle of the camera is also critical for accuracy, with
processing, employs machine learning algorithms for the best results at a 180-degree angle, and accuracy
classification and detection. The study's experiments dropping with significant changes in angle. In
demonstrated that the safety system is highly accurate summary, their machine vision system looks
in identifying potential accidents and providing real- promising for detecting track defects, and ultrasonic
time alerts to train operators and control rooms. It sensors give accurate measurements. Still, we need to
performed well in detecting elephants, drowsiness, consider external factors and camera angles to
and cracks, while also spotting abnormalities in maintain accuracy. Further research is essential to
critical components like the engine, brakes, and enhance the system's performance and explore its
wheels. The system's accuracy indicates its reliability practical use in the real world.
in preventing accidents and enhancing overall railway
safety. However, the research has some limitations, 14 Qi Li et al(2023) proposed Feature Selection and
such as a lack of data for rural area railways due to Damage Identification for Urban Railway Track
challenges in communicating with relevant Using Bayesian Globally Sparse Principal
individuals quickly. Fetching accurate data via Google Component Analysis This paper introduces a
Map API is also difficult in certain areas, affecting the Bayesian method for effectively identifying and
precision of test results. Overall, the study introduces selecting features in urban railway track monitoring
 data to detect damage. They focus on train-induced friendly support for mass transportation, improving
dynamic responses, aiming to find shared patterns of safety and asset digitalization. However, challenges
sparseness in the data. Unlike traditional methods that include improved navigation in GPS-denied areas,
focus on local sparseness for individual components, local data storage, safety measures, cost optimization
our approach globally selects relevant features for all in maintenance design, and adherence to railway
components, aiding in automatic unsupervised feature standards. Most existing mobile manipulators aren't
selection. To achieve this, They use a variational tailored for railway use, necessitating future research
expectation-maximization algorithm to find the most for railway-compatible designs. Addressing issues like
probable uncertain parameters, maximizing the poor data transfer signals at railway sites and
likelihood function. They validate our Bayesian exploring teleoperation and mixed reality-based
globally sparse PCA method by comparing it with the mobile manipulation are potential research areas.
current state-of-the-art method, using actual train- Additionally, considering the variety of tools needed
induced vibration data from an urban railway track for railway track maintenance, exploring self-
their method successfully handles unsupervised reconfigurable end-of-arm tools is another research
feature selection for high-dimensional structural prospect in the field of mobile manipulators.
health monitoring (SHM) data, accurately identifying
track bed plate crack damage through inferred global
sparse patterns. They introduce three indexes to III. PROPOSED METHODOLOGY
quantify the damage identification performance.
Given its high accuracy in identifying actual damage,
our method represents a promising step towards
practical damage identification for urban railway  Railway networks often intersect with
tracks. The timely detection of damage is crucial for natural habitats, posing significant risks to wildlife,
ensuring the safety of railway tracks and preventing particularly elephants, and the safety of railway
discomfort for passengers due to vibrations. In future operations. Collisions between trains and elephants
studies, they aim to explore physics-informed not only endanger lives but also threaten the
supervised deep learning methods to improve track conservation of these majestic animals. In response to
damage localization and classification. Additionally, this critical challenge, our project introduces a
addressing limitations, such as the absence of physical comprehensive solution that integrates cutting-edge
model information and the inability to estimate technologies—image processing, IoT (Internet of
damage severity and type, will be a focus. Exploring Things), and crack detection systems—to mitigate
the application of the method to other types of tracks, these risks and protect both wildlife and railway
including high-speed railway tracks, is also infrastructure.
considered for future research.  First, to provide timely alerts to railway operators
14 Miftahur Rahman et al(2023) A Review on the upon the detection of elephants in proximity to
Prospects of Mobile Manipulators for Smart railway tracks, enabling swift preventive measures to
Maintenance of Railway Track This study states avoid accidents. Second, leveraging IoT sensors, the
that Railways are an eco-friendly and reliable public project aims to identify and report track faults,
transport system, with track maintenance being crucial including cracks or damages, ensuring the safety and
for safety and comfort. Despite being older than reliability of railway infrastructure.
automotive and aerospace industries, railways often  By harnessing image processing capabilities and IoT-
lack modern robotic technologies. As train demand based sensors, our innovative system not only aims to
grows, there's a need to upgrade maintenance using prevent collisions between trains and elephants but
robotic and autonomous solutions. Existing railway also collects valuable data on elephant behaviors,
robots mainly focus on inspection, lacking repair movement patterns, and habitat usage. This data not
capability. Research highlights the potential of mobile
only facilitates immediate preventive actions but also
manipulators for maintenance tasks across various
contributes essential insights for research and
industries. They're categorized by robot type and
application area, including ground, aerial, and conservation strategies.The integration of crack
underwater robots. While widely used for high- detection technology further fortifies the railway
voltage power lines, mobile manipulators find infrastructure, ensuring its durability and safety.
applications in nuclear reactors, aerospace,  This holistic approach underscores our commitment
agriculture, and construction. By incorporating not only to railway safety but also to the preservation
design, instrumentation, algorithms, and workflow of wildlife habitats and species, particularly the
frameworks, mobile manipulators can enhance endangered elephant populations.Through the
railway track maintenance. Previous studies show amalgamation of technology and conservation efforts,
economic and safety benefits in different railway this project strives to revolutionize railway safety
applications. Using intelligent mobile manipulators protocols while actively contributing to the protection
can offer robust, reliable, and environmentally of wildlife and their ecosystems.
 PROBLEM STATEMENT 2. Animal Obstacle Detection:
In the context of, for example, autonomous
vehicles or wildlife monitoring, image processing is
 The intersection of railway tracks with used to detect obstacles such as animals on the road.
elephant habitats has emerged as a critical concern Cameras capture images or video frames, and image
due to the heightened risk of collisions between trains processing algorithms are applied to identify and
and elephants. Current manual monitoring methods classify objects in these images. Machine learning
exhibit significant shortcomings in accuracy, real- techniques can be integrated to train the system to
time responsiveness, and consistency, amplifying the recognize different types of animals, distinguishing
threat of accidents and exerting adverse effects on them from other objects.
wildlife conservation endeavors.
 The Challenge: To address this pressing issue, the In both cases, the process typically involves the
challenge is to devise an automated system capable of following steps:
proactively averting accidents involving trains and
elephants on railway tracks. This system necessitates Image Acquisition: Capture images using cameras or
the integration of state-of-the-art technologies, sensors.
including image processing, IoT (Internet of Things),
alongside crack detection mechanisms.. Pre-processing: Enhance or clean up the images to
improve the quality and make it easier for algorithms
to analyze.
INTEGRATION OF TECHNOLOGIES
Feature Extraction: Identify relevant features in the
Image Processing: Utilize cameras along railway images, such as edges, textures, or patterns associated
tracks to capture images of the surroundings(elephant with cracks or animals.
and crack detection). Image processing is a field of
study and technology that involves manipulating and Detection: Apply algorithms to detect and locate the
analyzing visual information, typically in the form of presence of cracks or obstacles based on the extracted
images or videos. The goal of image processing is to features.
improve the visibility or interpretability of the content
in an image or to extract useful information from it. Decision Making: Based on the analysis, the system
image processing works like a set of tools that can makes decisions, such as alerting humans to inspect a
enhance, modify, or extract features from images. It structure for potential issues or warning a vehicle's
involves techniques such as filtering, edge detection, control system to take evasive action.
image compression, and more. Image processing is
widely used in various applications, including Overall, image processing enables the automation of
medical imaging, surveillance, facial recognition, tasks that involve interpreting visual information,
computer vision, and even in everyday applications making it valuable in applications ranging from
like photo editing apps on smartphones. Image infrastructure inspection to autonomous systems.
processing is a field of study that involves analyzing
and manipulating images to extract useful
information or enhance certain features. When it IoT (Internet of Things): Connect cameras and
comes to detecting cracks or obstacles like animals in sensors to a centralized system for real-time data
images, image processing plays a crucial role in transmission. The Internet of Things (IoT) is
automated systems and computer vision applications. increasingly being applied in the railway industry to
Here's a simple explanation: enhance safety, efficiency, and maintenance
processes. In the context of detecting cracks and
1. Crack Detection: animal obstacles, IoT technologies can be employed
In infrastructure inspection, like checking for in the following ways:
cracks in buildings or bridges, image processing
techniques can be employed. The process involves 1. Sensor Networks: IoT-enabled sensors can be
capturing images of the structure and then using installed along railway tracks to continuously monitor
algorithms to analyze these images. The algorithms the condition of the tracks. These sensors can detect
can identify patterns and characteristics associated anomalies, such as cracks or abnormalities in the
with cracks, helping to automatically detect and track structure. Accelerometers, strain gauges, and
highlight areas where there might be structural issues. other types of sensors can be utilized for this purpose.
 Data Collection: Collects valuable data on elephant
2. Vibration and Acoustic Sensors: These sensors can behavior and track usage patterns.
be embedded along the railway tracks to detect
unusual vibrations or sounds that may indicate a
potential issue, such as a crack in the rail. Machine BLOCK DIAGRAM
learning algorithms can analyze the data collected
from these sensors to identify patterns associated with
track defects.

3. Drones and Cameras: Drones equipped with

cameras and other sensors can be employed for aerial
surveillance of railway tracks. They can capture high-
resolution images and videos, which can be analyzed
for signs of track damage or the presence of
obstacles, such as fallen trees or animals.

4. Animal Detection Systems: In areas where there is

a risk of animal obstacles on the tracks, infrared
sensors, cameras, or other motion detection devices
can be installed to identify the presence of animals.
This information can then be relayed to the control HARDWARE COMPONENTS
center or train operators to take appropriate action,
such as slowing down trains or activating warning 1. NOD MCU
systems. 2. LCD
3. PC
5. Data Analytics and Predictive Maintenance: 4. GPS
Collected data from various sensors can be analyzed 5. IOT
using data analytics and machine learning algorithms
to predict potential issues before they become critical. SOFTWARE TOOLS
This proactive approach allows for timely
maintenance and reduces the risk of accidents or 1. EMBEDDED C
disruptions. 2. ARDUINO IDE
3. PYTHON
By integrating IoT technologies into railway 4. OPEN CV
infrastructure, operators can create a smarter and
more responsive system that enhances safety, reduces
downtime, and improves overall
operational efficiency.

ADVANTAGES

Enhanced Accuracy: Image processing provides

accurate elephant detection, reducing false alarms.

Real-time Alerts: IoT enable instant communication,

reducing response time.

Automated System: Reduces dependency on manual

patrols, ensuring continuous monitoring.

ALGORITHM OF PROPOSED SYSTEM

 1. Cameras and sensors installed along railway and deep neural networks” International Research
tracks detect elephant presence and crack Journal of Modernization in Engineering Technology
detection (demo). and Science -May 2023
2. Image processing algorithms analyze
images to identify elephants accurately. [4] T. Dedeepya, A. Kalyani, D. Niranjan Reddy, Ms.
3. Data is transmitted in real-time to a central S. Sushma “Iot based railway track crack detection
control system using IoT and alert system” Dogo Rangsang Research Journal
4. Control system assesses the risk and sends Vol-13, Issue-4, No. 6, April 2023
alerts to train operators and relevant
authorities. [5] V.Vivek, J. Hemalatha , T.P. Latchoumi , S.
Mohan, “Towards the development of obstacle
detection in railway tracks using thermal imaging”
June 1, 2023

[6] Alfored George Charles Prem, Rajakumar

Singarayan,Subraja Rajarethnam, Nalini Mohan
“Survey On Smart Siren to Alert Animals About
Train Tracks in Image Processing” 2023

[7] Qiang Zhang, Fei. Yan, Weina Song,Rui

IV. CONCLUSION Wang,Gen Li Automatic Obstacle Detection Method
for the Train Based on Deep Learning Sustainability
2023
Our project successfully integrates image processing,
IoT, and LoRa technology to detect elephants near [8] Jinyan Qu,Shaobin Li,Yanman Li and Liu Liu
railway tracks, reducing accident risks. Real-time Research on Railway Obstacle Detection Method
notifications to train operators prevent collisions, Based on Developed Euclidean Clustering 2023
showcasing the potential of rapid response systems.
Data collected aids wildlife protection, contributing [9] Istiak Mahmud, Md.Mohsin Kabir, Jungphil Shin,
to broader conservation efforts and coexistence. Our Chayan Mistry, Yoichi Tomioka, M.F.Mridha
project demonstrates how innovation can foster safer “Advancing Wildlife Protection: Mask R-CNN for
environments and bridge human progress with nature. Rail Track Identification and Unwanted Object
By fostering harmony between technology and Detection” IEEE-2023
nature, we envision a future where biodiversity is
[10] Mucahid Mustafa Saritas, Ilkay Cinar, Yavuz
Selim Taspinar, Murat Koklu “Railway Track Fault
preserved through such technological solutions. Detection with ResNet Deep Learning Models
Integration with train control systems to enable Research Gate” April 2023
automatic train stoppage on detection. Machine
learning enhancements for better identification of [11] Rosebell Paul, Neeraj M, Neeraj Sagar, Vaibhav
elephant movement patterns. Expansion of the system Nair, Yadukrishnan Ps “A Swarm-based AI Aided
to cover other wildlife protection initiatives. Wheel Bot System to Detect Cracks in Railway
Tracks” IEEE- 2022
REFERENCES
[12] Fahim Ur Rahman , Md. Tanvir Ahmed , Md.
[1] Y. Lavanya ,G. Vamsi Krishna, T. Sowmya, K. Mehedi Hasan , Nusrat Jahan “Real-Time Obstacle
K. Trinadh , L. Akash Raghu,(2023) “Iot smart Detection Over Railway Track using Deep Neural
engine for detecting railway crack” Dogo Rangsang Networks” Procedia Computer Science Volume 215,
Research Journal,March 2023 2022

[2] Taynara Martins Gonc¸alves , Andrea G. Campos [13] Qi Li, Yong Huang, Jiachui Chen, Xiahui Liu,
Bianchi1 and Glauco F. Gazel Yared “Digital Image Xianghao Meng, Chao Lin “Feature Selection and
Processing in the Diagnosis of Cracks in Steel Sheet” Damage Identification for Urban Railway Track
Brazil-2023 Using Bayesian Globally Sparse Principal
Component Analysis” Sustainability 2023, 15(6),
[3] Tanish Tallam, Thejas J, Rashmi T V, “Crack 5391
detection in railway tracks using image processing
[14] Aaditya Aurobindo,Pratham Mehta,Harshal
Soman,Rajashree Daryapurkar,Ninad Mehendale
“Development of a Robotic System for Non-
Destructive Crack Detection on Railway Tracks” 31
Oct 2023

[15] Mr. Miftahur Rahman, Haochen Liu, Dr. Isidro

durazo-cardenas, Andrew Starr, Amanda Hall, Robert
Anderson “A Review on the Prospects of Mobile
Manipulators for Smart Maintenance of Railway
Track” 2023