Lane Departure Warning Systems (LDWS) have recently become an integral part of drive assistance systems.
Abstract
Lane Departure Warning Systems (LDWS) have recently become an integral part of drive assistance systems. However, such systems are expensive and require professional installation. On the other hand, a significant increase was made in smartphone use. In order to spread this important accessory to the mainstream market, efforts are made to develop LDWS as a smartphone application. This requires an algorithm that is both robust and can overcome various problems such as camera quality and limited processing power.
The aim of the project which is described in this document is to develop such an algorithm, while testing different options for implementation. The chosen algorithm combines image processing methods from several articles, with a Kalman-filter based tracking system. The parameters were chosen empirically.
Experimental results show that the algorithm performs reliably and robustly under different road and lighting conditions. In addition, the option of integrating data from device’s sensors was tested, in order to increase algorithm robustness.
Flowchart
Result
