Speed Bump System Based on Vehicle Speed Using Adaptive Background Subtraction with Haar Cascade Classifier

Abstract

Driving at high speed and recklessly is the main cause of traffic accidents. In several places speed bumps are installed as a medium to warn drivers to slow down the speed of the vehicle, but the installation of speed bumps in several places has become a problem in itself with inconvenience for drivers traveling at low speeds, so it is necessary to develop an intelligent system to warn drivers when speeding. vehicles break safety boundaries, making drivers safer and more comfortable. At the vehicle identification stage, a combination of the Adaptive Background Subtraction method with the Haar Cascade Classifier is proposed, and vehicle speed estimation is carried out by calculating the time difference in the detection area or Region of Interest (ROI). Testing was carried out using traffic videos with three conditions, namely day, evening and night, with each condition using the same object data, namely 55 images of car objects. The proposed method produces car detection accuracy with an average of 85% and MSE 0.5 in vehicle speed measurements