Lane and Curve Detection

Lane and Curve Detection

Shikhar Gupta¹, Ayush Kumar², Tarun Sharma³, Devyansh Sharma⁴, Sidhant Singh⁵

Department of Computer Science, Chandigarh University

Abstract - Developing new features is one of the most important step in distribution arising time for forthcoming machine sector which ensures the auto-mobilist’s safety and dropped the number of roads accidents.

Indeed, the road and lane discovery is truly complex and challenging task. It includes localisation of the road and the determination of the relative position between vehicles and Road digital cameras on the vehicle will capture the picture of road and lanes using various programming algorithms. The lanes are pulled using Hough transfigure through a brace of hyperbolas which are fitted to the edges of the lanes. The proposed lane discovery system can be applied on both painted and unpainted roads as well as crooked and straight road in different downfall conditions. The proposed system does not bear any spare information analogous as lane range, time to lane crossing and neutralize between the centre of the lanes. In addition, camera estimation and coordinate transformation are also not demanded. The system was excavated under various situations of changing illumination, and murk Goods in various road types without speed limit.

Keywords:

Hough Transformation, Canny Edge Detection, Bilateral filter, Greyscale, Region of Interest (ROI), Convolutional Neural networks (CNN), Recurrent Neural Network (RNN).