Fatigue driving state detection method based on face video

The invention discloses a face video-based fatigue driving state detection method. The method comprises the following steps of S1, acquiring required data in a multi-thread mode; s2, acquiring face key points; s3, pulse waves are obtained through a TC-CAN method; s4, processing the face key points through a ResNet18 network to obtain external behavior characteristics; s5, extracting required heart rate features from the pulse wave data obtained in the step S3; and S6, fusing the external behavior features and the heart rate features for classification. According to the method, the face behavior state and the heart rate state of the driver are obtained through the video analysis technology, the external behavior state and the heart rate characteristics related to fatigue driving can be extracted and fused through the model, end-to-end judgment from the video to the fatigue driving state is achieved, and the driving safety is improved. Therefore, the accuracy and reliability of fatigue driving state detection a