Research And Design Of Fatigue Driving Detection System For Vehicle Embedded Platform

The popularity of passenger car brought great convenience to people's travel, people's work and life style also changed dramatically, but also to the road traffic are of great potential safety hazard, traffic accidents frequently happen to many families bring heavy disaster, serious traffic problems to people's life has brought great trouble. Fatigue as a serious one of the main causes of traffic accidents, how to in the process of driving for rapid detection and early warning is always an important research direction of researchers in recent years, various detection methods have been put forward and repeated experiments and improvements. By comparing the findings: detection method based on image processing technology for a non-invasive, thus has broad market prospect.Condition of infrared and visible light during the day nig ht, the author of this paper respectively under the condition of fatigue test method was designed, the main research content is as follows:(1) in the process of driving at night, through the infrared camera to collect images, and designs the infrared nigh t under the condition of face detection, eye location method, and compares the different fatigue characteristic parameters, designs and realizes the PERCLOS fatigue parameters, the driver's head displacement and the pilot blink frequency extraction method, the system integrated the three characteristic parameters comprehensive determine whether the driver is in a state of fatigue;(2) under the condition of visible light during the day, through the facial image data transmission to Face Core cloud computing open platform, analysis of the current driver identity, by monitoring the engine state to calculate the vehicle running time, and finally integrated the two information determine driving fatigue state. For daytime algorithm of digital image processing with the color images of complex, and the limited resources of embedded system defects, so the on-line detection method, the complex operation shifted to cloud detection platform, reduce the burden of the processing of the embedded system. The experimental results show that the system has high detection accuracy, real-time detection is strong, can achieve half real-time test results;(3) on the basis of the former two key technology research, through the analysis of system requirements and hardware selection, b uild up the embedded system development platform, the detection algorithm is transplanted into the embedded system, embedded platform was designed and implemented a set of fatigue detection system detection accuracy, and the performance of the algorithm in embedded systems.Based on utility as the goal, for the ARMv7 architecture as well as tuning ARM Cotext A8 embedded processor was used to construct the embedded system development platform, and studies the fatigue driving technique based on image and fatigue detection based on identification technology, designs and realizes an embedded platform fatigue detection system. Early stage of system design in combination with the practical situation of the on-board system, give full consideration to the actual application scenarios, and according to the modular design thought, and system easy to transplant, can be convenient and integrate existing vehicle system.