“Real-Time Lane Detection for Driving System Using Image Processing”

In the past decade a lot of research has been done on vision based real-time advanced driver safety system (ADAS). Many OEM car manufacturing companies have stressed their research on ADAS so as to increase the vehicle and driver safety. This work is aimed at developing an algorithm for on road lane departure warning (LDW) system. LDW can be successfully used for dropping the on road accidents which happened due to driver drowsiness or ignorance. In this report a survey has been done for the study and comparison of available algorithm for LDWs. An image processing based algorithm has been developed which uses a single camera mounted on-board. The robustness of algorithm has been tested under various lightning conditions and in the presence of heavy shadowing cast by bridges, vehicles etc. The algorithm is implemented in two steps, first it finds the vanishing point on the road and then a careful selection of line segments have been done by using the information of vanishing point. A smart and robust heuristic feature based filter has been used to validate the detected lanes. Kalman tracker is then implemented in order to track the detected lane and handle the critical sections where the detection algorithm fails. The major challenge for any vision based system has always been to achieve high accuracy and precision along with low computational cost (required for real-time applications).