Search Results

Robust lane marking detection remains a challenge, particularly in temperate climates where markings degrade rapidly due to winter conditions and snow removal efforts. In previous work, dynamic Bayesian networks with heuristic features were used with the feature distributions trained using semi-supervised expectation maximization, which greatly reduced sensitivity to initialization. This work has been extended in three important respects. First, the tracking formulation used in previous work has been corrected to prevent false positives in situations where only poor RANSAC hypotheses were generated. Second, the null hypothesis is reformulated to guarantee that detected hypotheses satisfy a minimum likelihood. Third, the computational requirements have been greatly reduced by computing an upper bound on the marginal likelihood of all part hypotheses upon generation and rejecting parts with an upper bound less likely than the null hypothesis.

In this paper, a method of improving the automated vehicle’s perception using a multi-pose camera system (MPCS) is presented. The proposed MPCS is composed of two identical colored and high frame-rate cameras: one installed in the driver side and the other in the passenger side. Perspective of MPCS varies depending on the width of vehicle type in which MPCS is installed. To increase perspective, we use the maximum width of the host vehicle as camera to camera distance for the MPCS. In addition, angular positions of the two cameras in MPCS are controlled by two separate electric motor-based actuators. Steering wheel angle, which is available from the vehicle Controller Area Network (CAN) messages, is used to supply information to the actuators to synchronize MPCS camera positions with the host vehicle steering wheel.