A vision based intelligent transportation system for smart cities : A

Abstract

Keeping track of congested roads to overcome the problem of slow traffic has always been a perennial and principal task for road authorities. In order to pave way for tackling this outstanding concern of road bottlenecks, traffic analysis systems implemented ought to be highly efficient and immensely accurate. With the acquirement of highly accurate traffic data, road officials can then have the capability to administer the required steps to maximise and greatly improve road traffic conditions. In this project, we will be sampling videos taken from an iPhone XS and use Python 3 scripts to read these videos and produce outputs for analysis. The scripts involve vehicle detection and tracking, speed estimation as well as counting the number of vehicles passing a region of interest (ROI). The Python scripts would run in a Linux environment and for this project, it would be Ubuntu. In addition, frames are extracted from the sample videos which can be used for image processing in the Python graphic user interface (GUI). With this research, we can figure out more opportunities of using OpenCV-Python in traffic applications. Comparative analysis of various implementation of image segmentation techniques and image enhancement processes was tested to discover which created the most precise outputs.