Android smart phone based participatory sensing

• Main Author: Goh, Yee Ting
• Other Authors: Li Mo
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/62803
• Institution: Nanyang Technological University

In recent years, there is an increasing market demand for mobile navigation application. However, current navigation application only provides a driving path which drivers may not know which lane to stay in to make a turn. One of the improvements is to provide lane information to drivers through real-time lane detection with the use of an Android mobile phone. This project proposes a robust and real-time lane detection that is able to detect all the lanes and to identify the current car position. Firstly, the algorithm of lane detection is based on transforming a front facing image to a bird-eye view image in order to make lane markers to appear vertical with almost equal width. Secondly, Canny edge detector is used to obtain edges of an image and lines are detected by Hough line transform. Next, grouping of lines is used to obtain a robust line from a group of nearby lines where these lines are used to identify the lane markers on the road and to identify the current car position. The experimental result shows that all the lanes are detected and is able to keep track of current car position when there is a lane change. In addition, the proposed algorithm is able to identify the missing lane markers and dynamically insert a lane marker at the correct position.