A deep learning approach for automatic traffic surveillance under extreme climatic conditions

Venkataramani, Saravana Ganesh (2023) A deep learning approach for automatic traffic surveillance under extreme climatic conditions. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (567kB) | Preview
Preview	PDF (Configuration manual) Download (571kB) | Preview

Abstract

Machine learning for traffic surveillance has proven to be highly useful in the recent times. In this research two ML (machine learning) models have been developed with VGG16-Net and tested on a secondary dataset. One model predicts the weather conditions prevailing when vehicles travel on the road and the other model predicts accidents caused by certain weather conditions on the road. Two types of weather conditions namely rainy and snowy are being predicted. SGD optimizer and Adam optimization algorithms have been used to optimize the model The loss function used is cross entropy loss. Evaluation of performance is done through accuracy, precision, f1-score and recall parameters. The findings show that the accident prediction model rendered 90% accuracy and the weather prediction model rendered 92% accuracy.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Muslim Jameel, Syed UNSPECIFIED
Uncontrolled Keywords:	Traffic accident; weather prediction; snowy; rainy; VGG16 Net; convolutional neural networks; CNN
Subjects:	G Geography. Anthropology. Recreation > GE Environmental Sciences Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > TE Highway engineering. Roads and pavements Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	08 Jan 2025 18:41
Last Modified:	08 Jan 2025 18:41
URI:	https://norma.ncirl.ie/id/eprint/7287

Actions (login required)

View Item