Citation

Gopichand G, Sampath R, Ramesh LK. ARPN J. Eng. Appl. Sci. 2023; 18(20): 2304-2310.

Copyright

(Copyright © 2023, Asian Research Publishing Network)

DOI

10.59018/1023282

PMID

unavailable

Abstract

A vast quantity of information about vehicular traffic is logged into the monitoring system that monitors traffic each second. It takes a lot of work to manually monitor this data, and it also necessitates hiring staff who are just responsible for monitoring. Deep learning, also known as convolutional neural networks, is a methodology that has the potential to be used for the purpose of controlling and monitoring traffic. After going through some preliminary processing, the data from the various traffic monitoring systems are included in the training Traffic-Net dataset. Therefore, this work implemented the deep learning convolutional neural network (DLCNN) for analysing road images, which can detect the accident, a fire occurred, normal, and dense traffic classes. Initially, the TrafficNet dataset is divided into eighty percent for training and twenty percent for testing. After that, a dataset preparation procedure is carried out to standardise the complete dataset. During the procedure of image pre-processing, all the pictures are scaled down to the same dimensions. In addition, DLCNN is used for the prediction of traffic status. The simulation results showed that the proposed DLCNN resulted in superior traffic analysis performance measurement as compared to other methods.

Language: en