Research

 

Intelligent Vehicle Experimental Test-beds

 

Human-Centered Driver Assistance Systems

Vision Systems for Occupant Posture Analysis

 

Vision Systems for Driver Analysis and Interface

 

Related Research

 

Surround Vehicle Detection Research

Motion-Based Surround Detection

An estimate of the planar motion of the ground can be found by calculating the ego-motion of the vehicle based on the fusion of CAN bus readings with motion information from image features.  This in turn can be compared to the motion from the video sensors to detect objects that are not part of the ground.  This technique can be applied to all types of cameras and lenses.  Below are some results from this research.

Model-Based Surround Detection

For detecting vehicles to the side of the car, a model-based wheel detector can be used.  After de-warping the omnidirectional image to generate a side view, a Gaussian Mixture Model based probabilistic detection system is used to extract possible wheel locations.  These candidates are tracked and wheel position estimates are further refined based on this tracking.  Below are some results from this research.

Videos of these results can be viewed in the Videos portion of this web site

More details on these research topics can be viewed in the Publications portion of this web site.