Deep Learning based Artificial Intelligent Systems in Road Traffic Density Estimation and Congestion Classification

Abstract

Objectives:The main objective of this paper is to employ the subset of artificial intelligence, namely, deep learning to estimate road traffic density and thus mitigate the undesirable effects caused by traffic congestion and improve the quality of life of people. Methods: This work presents a method of classification of road traffic conditions based on video surveillance data obtained from CCTV cameras mounted on highways. A simple, basic architecture of deep convolutional neural network (DCNN) based method is introduced that learns traffic density from pre-labeled images in order to estimate the traffic flow density in highways. Findings: The standard publicly available UCSD dataset of real videos is used for experimental verification. The experimental results obtained shows that the proposed model outperformed all the existing conventional methods in the literature by reaching the highest accuracy and classifies the test video in less computational time. Novelty: The proposed methodology employs Matlab deep learning network designer with hyper parameter tuning, cross validation and activation maps to classify the road traffic density into three different states namely light, medium and heavy.