Road tracking using particle filters with partition sampling and auxiliary variables

In this paper we present a novel algorithm for road tracking using computer vision techniques. We have extended the sampling algorithm used in the standard particle filter and integrated our new algorithm into a multiple tracker framework for road tracking. The standard particle filters use Sampling Importance Resampling (SIR) to draw particles from an importance proposal density function.Experimental results show that our method is accurate and is suitable for real road applications.

► Challenging outdoor computer vision application.
► Improved particle filter tracker using partition sampling and auxiliary variables.
► Accurate road tracking results suitable for real world applications.