Fault Detection from Images of Railroad Lines Using the Deep Learning Model Built with the Tensorflow Library

Abstract

A means of transportation is the way in which an object, person, or service is transported from one place to another. Rail transportation occupies an important place in terms of cost and reliability. Most train accidents are caused by faults in railroad tracks. Detecting faults in railroad tracks is a difficult and time-consuming process compared to conventional methods. In this study, an artificial intelligence based model is proposed that can detect faults in railroad tracks. The dataset used in the study consists of defective and non-defective railroad images. The proposed model consists of foldable neural networks developed using the Tensorflow library. Softmax method was used as a classifier. An overall accuracy of 92.21% was achieved in the experiment.

Remote Sensing and Machine Learning for Safer Railways: A Review

“…Notably, the training time for Conv2D exceeded the duration required by the other models. Sener et al (2022) disclosed that one disadvantage of their approach was that it requires preprocessing to normalize the images before training the model [22]. found that blurred images can severely decrease detection accuracy [74].…”

“…Saha et al (2022) [30],ŞENER et al (2022) [22],Sysyn et al (2019) [32],Wu et al (2022) [62],Chen et al. (2021) [40], Niu et al (2020) [60], Gibert et al (2016) [23], Zhang et al (2021) [33], Jamshidi et al (2017) [24], Guo et al (2021) [51], Gan et al (2017) [58], Bojarczak et al (2021) [57], Wang et al (2019) [53], Hashmi et al (2022) [20], Wei et al (2020) [42], Rizzo et al (2010) [38], Passos et al (2022) [26], Zhang et al (2022) [68], Wu et al (2020) [59], Jin (2021) [63], Zhang et al (2021) [18], Zhang et al (2018) [36], Lu et al (2020) [69], Chandran et al (2022) [70] Fastener Inspection Cao et al (2023) [43], Wei et al (2019) [52], Qi et al (2020) [44], Liu et al (2019) [34], Franca et al (2020) [71], Chandran et al (2021) [21], Chandran et al (2021) [35], Cui et al (2019) [61], Liu et al (2021) [55], Liu et al (2021) [25], Aytekin et al (2015) [72].…”