Distributive and Governing System for Descriptive Error Identification of High Speed Railway Illustrations and Images using Convolutional Neural Networks
Lipsa Das, Smita Sharma, Astik Naval, Akanksha Singh, Pooja Anand
Abstract
During the operations of high-speed railway the deformities and errors during the course of its actions is determined with the help of 2 or more images of a particular point on the train which is also use for maintaining various kinds of parameters and used for maintenance. These types of images are often termed as bitemporal images however it is of great responsibility dad the conclusions derived from these images must be of great accuracy and must possess higher percentage of correctness. Through this paper we show a process in which 2 neural networks (convolutional) are set up to perform the desired tasks. This process has no link to the conventional method of defect analysis however this system has the potential of discriminately classifying the 2 bitemporal images and evaluating the differences between them. This type of technology can be used in areas of validation of decisions which are based upon a single instance (image). This type of technology also uses various deep learning algorithms which would not be possible with conventional methods of error detection and thus is the most potential safety protocol technology used in high-speed railway.