ROAD BOUNDARY DETECTION BASED ON THE RASTER SCANNING AND THE RANDOMIZED HOUGH TRANSFORM

Abstract

Detecting the road boundary is an important aspect for the applications of Intelligent Vehicle (IV). Automatic road lane detection is one of the crucial parts of the vision based driver assistance system. The driver assisting system can reduce the accidents, and thereby control the flow of traffic. The existing system uses the Laser Interferometry Detection and Ranging (LIDAR) by generating the Region of Interest (ROI) as prior information to detect the object or the path. The problem associated with the existing technique is that the deficiencies in the software realm and in need to select the proper ground filtering algorithm. The proposed system is used to improve the road lane detection system by using Hybrid Median filter. The system also produces a warning or an alert message to the user for proper driving in the road. It also uses the Dynamic Programming (DP), a powerful algorithm for optimal path finding on a cost field. Randomized Hough transform is applied on the filtered edge image to detect practically useful road boundaries with straight line segments. Finally, we aim to evaluate our system with the use of various filters for estimating the processing speed, time consumption for each filter method thereby to achieve high efficiency over the existing methods.