Occupant posture analysis using reflectance and stereo image for "smart" airbag deployment

Robust detection of vehicle occupant posture is necessary for intelligent airbag deployment. This paper presents a vision-based method of estimating the size, posture and pose of the occupant. Utilizing raw reflectance and stereo disparity images, this algorithm presents a mixed-mode approach to finding occupant features. Extensive experiments show the viability of this mixed-mode method in identifying the occupant's head location and suggest the feasibility of extending the analysis system to include detection of features such as occupant arms and foreign objects.