DESIGN AND IMPLEMENTATION OF A SPEED DETECTION SYSTEM USING RADAR AND IMAGE PROCESSING

Abstract The issue of vehicle accidents occurring on the roads on a daily basis, which only seems to be increased by the minute is what this paper researches and seeks to tackle. With the simple technology that has been made available globally, a cost-effective system can be developed to fight overspeeding, reduce road accidents and save lives. As a result, in this project, an Automatic Number Plate Recognition program and a web service for speed detection were implemented. This was done with the aim of monitoring frequently used roads, and those areaswhere the accidents normally take place in order to sanction drivers guilty of driving above set speeding limits. This made use of machine learning-based Optical Character Recognition program couples with a web application to issue speeding fines dynamically to offenders via email. In the course of this project a validation accuracy of 97% was achieved for the training done with Nigerian License plate characters. The mailing was fully functionally at the end of the project. Though, a few hitches were experienced on the license plate detection phase as some license plate shapes vary slightly. This proved the need to improve the license plate detection to detect modded license plates in addition to the regular ones.