A Novel Machine Learning Framework for Real-Time Railway Crack and Squat Detection

Abstract

Preventing accidents and maximizing maintenance efforts require secure railway infrastructure. This study introduces a machine learning-based method for detecting cracks in railway tracks that makes use of computational intelligence and image processing techniques. The system analyses and tracks photos, extracts fault features, and accurately classifies abnormalities using MATLAB. The suggested system efficiently detects fractures, squats, and structural flaws by combining deep learning algorithms, edge detection methods, and pattern recognition models. Furthermore, a hardware module based on Arduino analyses identified defects and sends real-time notifications to a central monitoring system, allowing for prompt remedial action. By enabling proactive resource allocation and ongoing track monitoring, this connection improves predictive maintenance, which in turn raises railway safety and operating effectiveness.

Introduction

I. Introduction

Railway tracks are vital for transporting passengers and goods, but structural flaws like cracks and squatters pose safety risks, potentially leading to derailments. Traditional manual inspections are:

• Time-consuming

• Labor-intensive

• Error-prone

As trains operate continuously, real-time and automated crack detection systems are necessary to improve safety, reduce maintenance costs, and ensure efficiency.

II. Project Overview

This study proposes a MATLAB-based automated crack and squat detection system using:

• Image processing

• Machine learning (ML)

• Fixed camera setup

• Integration with Arduino for alerts

The system supports predictive maintenance and ensures long-term infrastructure reliability by monitoring tracks continuously.

III. Literature Survey

1. Sensor-Based Detection with GPS Alerts:

   • Uses IR sensors, GPS, and microcontrollers to detect cracks.

   • Challenges: Affected by environmental conditions, poor detection of small/deep cracks.

2. Improved Monitoring Systems:

   • Emphasizes the need for automated systems to reduce rail accidents and prevent animal injuries.

3. Collision Prevention & Crack Detection:

   • Combines IR, ultrasonic, RFID, GPS, and Bluetooth.

   • Future focus on AI and ML to enhance detection precision and system efficiency.

IV. Methodology

A. Data Collection

• Fixed camera captures images at regular intervals.

• Images processed using noise reduction, contrast enhancement, and edge detection.

• Detected cracks trigger alerts via LEDs, buzzers, SMS, or GSM modules.

• IoT integration allows remote monitoring and cloud-based data sharing.

B. System Architecture Modules

1. Image Acquisition Module

   • High-resolution images taken in real-time or at intervals.

   • Good lighting and stable positioning enhance accuracy.

2. Preprocessing Module

   • Enhances image quality for better crack visibility.

   • Involves denoising, contrast boosting, and edge detection.

3. Feature Extraction Module

   • Extracts crack-specific features (e.g., width, length, orientation).

   • Uses traditional filters (Sobel, Canny) and CNNs for deeper analysis.

4. Crack Detection & Classification Module

   • Detects cracks via image segmentation or ML algorithms (SVM, Random Forest).

   • CNNs provide more accurate and robust detection with fewer false positives.

5. Output & Decision-Making Module

   • Displays results and triggers actions:

     • Visual highlights

     • Text alerts

     • Auto-generated reports

     • Real-time notifications (SMS, alarms)

6. Arduino Integration

   • MATLAB communicates with Arduino via serial port.

   • Arduino triggers alarms or safety signals based on crack severity.

V. Key Advantages

• Real-time and remote monitoring

• Reduces human error

• Accurate detection of even small or hidden cracks

• Supports predictive maintenance

• Cost-effective and scalable

Conclusion

This study demonstrates how well an automated system based on machine learning can identify cracks in railway rails. Through the use of image processing algorithms, the system guarantees early crack diagnosis, reducing the likelihood of accidents and enhancing track maintenance. Railway safety and efficiency are further improved by combining automatic response systems, such as Arduino-based controls, with real-time monitoring. By drastically lowering the need for manual inspections, this method improves the scalability and dependability of railway monitoring. The accuracy of detection can be further increased with developments in machine learning, deep learning, and sensor technologies. Future developments in automated inspection systems and image processing will help create a safer and more effective railway network.

References

[1] DevelopmentofRailwayTrackCrackDetectionSystem using Arduino Prof. Najma Siddiqui1, Prof. Akil Ahmed1, Tejas Thakre2, Sujyot Jibhe2, Yogesh Karnake2, Samir Sheikh2, Yash Sangole2, SohelrazaSheikh2,InternationalJournalofScientificResearchin Science and Technology [2] Railway Track Crack Detection Arun Kumar R,Vanishree K, Shweta K,Nandini C, Shweta G Department of Computer Science and Engineering Tontadarya College of Engineering Gadag, India, International Journal of Engineering Research & Technology (IJERT)S [3] Real-Time Obstacle Detection Over Railway Track using Deep Neural Networks Fahim Ur Rahmana, Md. TanvirAhmeda,Md.MehediHasana,NusratJahanb*a, bDepartment of CSE, Daffodil International University,Dhaka,Bangladesh,ScienceDirect [4] IoT Based Railway Track Faults Detection and Localization using Acoustic Analysis and Signalling Likhith Chandra1 and Harsha V2 Department of Information Science and Engineering1,2 Global AcademyofTechnology,Bengaluru,India,International Journal of Advanced Research in Science, Communication and Technology (IJARSCT) [5] Implementation of Railway Track Crack Detection and Protection N.Karthick1, R.Nagarajan2, S.Suresh3, R.Prabhu4, International Journal Of Engineering And Computer Science ISSN:2319-7242 [6] AUTOMATIC RAILWAY TRACK CRACK DETECTION SYSTEM USING IOT R. Dharanidaran1, S. Jagan2, B. Anandhakeerthi3, S. Kaveyen4, U. Gowtham5, International Research Journal of Engineering and Technology (IRJET) [7] RAILWAY TRACK CRACK DETECTION USING DEEP LEARNING MODELS Mrs.G Sri Lakshmi, Assistant professor, Department of IT,sree.gpk@gmail.comG.Aihika,BTech, Department of IT,gundaaihika@gmail.comK.Sasidhar Reddy, BTech, Department of IT, sasidhar6898@gmail.com S.Reshma, BTech, Department of IT,shaik.reshu67@gmail.comK.Soniya, BTech, Department of IT,kurapatisoniya082@gmail.com, Journal of Engineering and Science [8] A Deep Learning Approach Towards Railway Safety Risk Assessment, HAMADALAWAD 1,(Graduate Student Member, IEEE), SAKDIRAT KAEWUNRUEN 1,ANDMINAN2,IEEE Access. [9] SCueU-Net: Efficient Damage Detection Method for Railway Rail JUN LU JI LIU 1,BOLIANG 1,2, QUJIANG LEI 2,JIE XU 2,XIUHAO LI 2,JUNHAOLIU 2, 2,ANDWEIJUN WANG 2, IEEE Access,IDPI [10] Robust Automatic Railway Track Crack Detection using Unsupervised Multi-Scale CNN S. Nandhini1 , Mohammed Saif2 , Kailash Venthan P3 , Sridhar E4 Assistant Professor, Department of Computer Science and Engineering1 B. Tech Students, Department of Computer Science and Engineering2,3,4 International Journal of Advanced Research in Science, Communication and Technology (IJARSCT) [11] Automated rail surface crack analytics using deep data-driven models and transfer learning Zhong Zheng, Haoyang Qi, Li Zhuang*, Zijun Zhang,Science Direct. [12] Classification of Rail Track Crack using CNN with Pre-Trained VGG16 Model Shreetha Bhat School of Computer Science and Applications REVA University Bengaluru, Karnataka, India shreetha.bhat@reva.edu.in Vidya S School of Computer Science and Applications REVA University Bengaluru, Karnataka, India vidyadhare13@gmail.com Asha Gowda Karegowda Department of MCA Siddaganga Institute of Technology Tumkur, Karnataka, India asha_gowda@sit.ac.in Devika G Dept of Computer Science and Engineering, Government Engineering College, KR Pet. Karnataka, India. sgdevika@gmail.com. 2022 IEEE International Conference on Distributed Computing and Electrical Circuits and Electronics (ICDCECE) [13] Designing of Improved Monitoring System for Crack Detection on Railway Tracks, Nilisha Patil1, Dipakkumar Shahare1, Shreya Hanwate1, Pranali Bagde1, Karuna Kamble1, Prof. Manoj Titre2 Proceedings of the Third International Conference on Intelligent Communication Technologies and Virtual Mobile Networks (ICICV 2021). IEEE Xplore Part Number: CFP21ONG-ART; 978-0-7381-1183-4 [14] RAILWAY TRACK CRACK DETECTION & OBSTACLE DETECTION SYSTEM Sagar Kotkondawar*1, Samyak Meshram*2,AbhishekRangari*3, Dr. J.P.Sathe*4, Dipali Nankate*5, Eshakha Sarve*6International Research Journal of Modernization in Engineering Technology andScience. [15] An Analysis Of Convolutional Neural Networks For Image Classification Neha Sharma,Vibhor Jain, Anju Mishra,An Analysis Of Convolutional Neural Networks

Copyright

Copyright © 2025 Sowmya S, Surendhar R, Karthikeyan , Manoj R, Priyanka S. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.