Lane Line Detection by Using Hough Transform

Abstract

In this study, a mixed approach is proposed to increase the detection success of lane-line especially for heavily shaded road images. These kinds of studies are the basis of lane departure warning and lane keeping assistance systems. Accurate and automatic detection is an important feature for reducing the accident risk and increasing driving safety. In most of the existing studies carried out on this subject, it has been observed that MATLAB software is widely used. In this study, the application was developed by using two different software in LabVIEW which is a graphical development platform. The text-based source codes developed in MATLAB has been integrated into the LabVIEW platform using the Mathscript Node tool for Hough Transformation (HT). By combining the superior features of both platforms, a stronger user interface in terms of visual objects compared with the MATLAB GUI with satisfactory analytical abilities was get. This mixed approach provides the software developers to choose the most appropriate syntax in a single platform. The proposed method has been tested on a limited number (4) of images, especially for heavily shaded images selected from dataset of the Carnegie Mellon University Robotics Institute Vision and Autonomous Systems Center and up to 100% success has been achieved in images exposed to high disturbance.