W. Zheng, Y. Nie

Pages: 147-157

Abstract
In order to improve the early warning effect of traditional traffic situation monitoring methods, this paper proposes a road environment traffic safety situation early warning and monitoring method based on machine vision. Firstly, the quadratic polynomial curve model of road environment is constructed according to machine vision. Secondly, the linear filter is used for road image smoothing and noise reduction to obtain real-time traffic information. Then, the equilibrium state equation of road safety is constructed to obtain the maximum lateral velocity of the vehicle. Finally, traffic safety situation early warning and monitoring is realized according to the situation early warning and monitoring equation. The experimental results show that this method can obtain the vehicle safety situation within 7 minutes, and the highest accuracy of safety situation early warning and monitoring can be 99.7%, indicating that the traffic safety situation early warning and monitoring of this method is obviously better.
Keywords: security situation early warning; machine vision; eliminate redundancy; weighted mean filter mask; smooth noise reduction; grayscale processing