Y. Wang, Y. Zou, M.J. Zhang, P. Ding

Pages: 43-54

Abstract
Aiming at the problems of low detection rate and high false alarm rate of traditional driver dangerous driving behavior detection methods, this paper proposes a driver dangerous driving behavior detection method in night driving environment based on least square ellipse fitting algorithm. Based on four dangerous driving behaviors: sharp turn, continuous lane change, emergency obstacle avoidance and emergency braking, the characteristics of drivers' dangerous driving behavior are extracted. Based on the feature extraction results, the least square ellipse fitting algorithm is used to detect the driver's dangerous driving behavior at night. The experimental results show that the detection rate of this method is high and the false alarm rate is low, which can effectively ensure the driving safety in the night environment.
Keywords: night environment; dangerous driving behavior; least squares ellipse fitting; feature extraction