You are on page 1of 2

NAME: GNANA SELVI AGUSTA. A GUIDED BY: Dr. K.

ARUL DEEPA

REG NO: 2020178018 - MCA (R)

A Deep Learning based Railway Track Fault Detection

INTRODUCTION:

Around the world rail systems are among the most preferred public transportation methods
and are becoming busier requiring them to operate with increasing levels of availability and reliability.
In addition, the speed and loads of trains have also been increasing greatly in recent years, and all
these factors inevitably raise the risk of producing rail defects. Railway operators worldwide have been
highly concerned about such defects as if treated late, they may lead to major revenue loss and have
bigger implications like loss of life due to accidents.

ABSTRACT:
Rail track is a critical component of rail systems. Accidents or interruptions caused by rail track
anomalies usually possess severe outcomes. Reliable and economical inspection of rail tracks is
paramount to ensure the safe and timely operation of the railway network. Traditionally, a trained
person visually inspects the rail for defects which makes the whole process slow, subjective and
dangerous. Later different non-destructive testing (NDT) techniques have been used such as Eddy
current testing, magnetic flux leakage testing, ultrasonic testing, phased array detection, guided wave
detection, and so on. Automated vision based track inspection utilizing computer vision and pattern
recognition techniques have been regarded recently as the most attractive technique for track surface
defect detection due to its low-cost, high-speed, and appealing performance. However, the different
modes of failures along with the immense range of image variations that can potentially trigger false
alarms makes the vision based track inspection a very challenging task. In this project, I am going to
use a multiphase deep learning based technique to detect railway track fault which is much better
than visual inspection. In this proposed system, initially the raw images are transferred to the image
processing subsystem where pre-processing is performed and then the deep learning model is trained
with the training data and validated by the testing data.

MODULES:

The proposed system consists of three modules:

1. Pre-processing subsystem
• Image Standardization
• Colour Analysis
• Edge Analysis using Sobel
• HSV(Hue Saturation Value) Transformation
• Image Augmentation
2. Model generation subsystem
• Train the dataset using CNN(Convolutional Neural Network) and save
the generated module
3. Testing the input image
• Use the generated model to test the input image of the railway track
to classify the it as defective or non defective

TECHNOLOGIES USED:

1. Tools: Anaconda
2. Language: Python
3. Libraries: Tensorflow, Keras, Cv2

GUIDE SIGNATURE

You might also like