Abstract: Aiming at the problems of difficulty for road congestion state detection and high computational complexity for traffic basic parameters,a holographic road traffic state detection method based on machine vision in lightweight was presented in this paper．In order to achieve automatic detection of the road traffic congestion state and a variety of traffic basic parameters simultaneously using machine vision in embedded system,firstly,according to the design idea of "points stand for a plane",road areas is customized,and the uniform distribution of sampling points in the areas is generated automatically,which contributes to reduce the computing and storage resources．Secondly, by combining the background subtraction algorithm with the frame differential algorithm,some import data are obtained,such as the non-existence sampling points,the existence sampling points,the motion existence sampling points and the motionless exis-tence sampling points reflecting the road congestion states．Then,road scene background is updated by the gray value of non-existence sampling points,and some important traffic basic parameters are obtained by calculating the space distribution regularity of the existence sampling points,and traffic congestion state is achieved by analyzing the space arrangement of the motionless existence sampling points．The experiment results show that the proposed algorithms have the advantage of high computational efficiency,less resource consumption,large detection range and strong robustness,etc.,and can quickly and accurately detect various traffic basic parameters and road congestion states.

Key words: Machine vision,Sampling points,Background update,Road congestion detection,Traffic basic parameters detection