Real-Time Tracking and Lane Line Detection Technique for an Autonomous Ground Vehicle System

Abstract

Real-time detection and tracking of a moving car in sequence of image frames acquired by a static camera and lane line detection using an onboard camera have been elucidated in this paper. Vehicle tracking has been performed by exploiting modelling behaviour of Kalman filter, and lane line detection has been done using modified Hough transform. In case of car tracking, image acquisition, detection of vehicle and locating its centroid are done using adaptive background subtraction technique. Parameter estimation of car’s position is further performed using Kalman filter. In case of lane detection, pre-processing of image, edge detection and masking results’ detection of broken lane lines are done. Hough transform, for detecting curvatures by exploiting the duality between points on a curve and curve parameters, is applied to join all broken parts and to get solid lane lines. The entire system has been implemented on a ‘VEEROBOT’ robotic car in laboratory environment. Real-time image has been captured using a ‘Logitech’ web camera. All relevant processing has been carried out by ‘Raspberry pi 3’. All the analyses of experimental results are elaborately presented in this paper.