The frequent occurrence of automobile traffic accidents seriously threatens the safety of human life and property. Therefore, fatigue driving detection has important social value and research significance. In consideration of the market demand of intelligent assistant driving system, we design a real-time driver fatigue detection system based on deep learning and ARM platform, which uses Samsung 6818A53 series ARM as the driver fatigue real-time detection platform. In order to reduce the interference caused by the change of light and the occlusion of sunglasses in the actual driving environment, the driver's face image is captured by USB infrared camera. Firstly, face detection and alignment are carried out by multi-task cascaded convolutional neural network; Then the eye region is obtained according to geometric relationship between the feature points; Moreover, the driver's eye state is identified by Convolutional Neural Network (CNN); Finally, fatigue judgment is made based on PERCLOS criterion. The system has been tested in the experimental simulation environment and the actual driving environment. The experimental results show that detection speed of the system can reach more than 20 frames per second, which meets the requirement of real-time detection.
Real-time Fatigue Driving Recognition System Based on Deep Learning and Embedded Platform
2019-03-06
American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS); Vol 53 No 1 (2019); 164-175 ; 2313-4402 ; 2313-4410
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
DDC: | 629 |
Real-Time Distraction Detection from Driving Data Based Personal Driving Model Using Deep Learning
Springer Verlag | 2022
|Real-time driving risk assessment using deep learning with XGBoost
Elsevier | 2022
|