Research on Real-time Monitoring System of Abnormal Driving Behavior Based on Multi-angle Fusion

Based on multi angle feature, this paper focuses on the study of the abnormal driving behavior monitoring technology. On the one hand, in order to conduct a judgment of abnormal driving, the program uses wireless sensor technology to monitor the trajectory of motor vehicles on the road;on the other hand, face detection and eye and mouth location are realized by face key point detection technology; drivers’ fatigue state are determined by using EAR algorithm to calculate the aspect ratio of eyes; in the process of driving, MAR algorithm is used to calculate the mouth opening to determine whether the driver is sleepy or not while driving. Once the above situation occurs, the embedded system platform will issue a voice alarm to prompt the driver to stop and have a rest, thus the traffic accident rate will be reduced further.