Date of Award

12-14-2023

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Information Science

First Advisor

Serhan Dagtas

Abstract

Catastrophic natural disasters have an impact on millions of individuals each year, whether directly or indirectly. A successful rescue operation can save a great deal of lives in the post-disaster phase, but researchers are still facing difficulties in carrying this out. Even while they are in peril, today's people frequently post updates about their whereabouts on well-known social media platforms, sometimes even asking for help. A prompt and appropriate response to these catastrophic events depends on developing an efficient and automated approach capable of retrieving real-time data from impacted locations and extracting essential elements. The research focuses on predicting flood water levels and speed for recommendations on road pass ability utilizing real-time data from traffic cameras and social media platforms like Twitter and YouTube. There are numerous challenges. The precise size of the reference object may not be known, the submerged object may only be partially visible, the height of the flood water that appears in different areas of the image scene may vary, and it may be difficult to continuously monitor a vehicle at low-quality video. To address these issues, the suggested model makes use of a person, a traffic signal, and a vehicle as reference objects. It has also been trained using datasets of images of flooded roads to estimate flood water levels, estimate speed using pixel mapping, and show the viability of the methods. The road's pass ability is then assessed by the model using the floodwater level and speed as inputs.

Download

Share

COinS