Health Monitoring System Of Driver Fatigue And Implementation

Along with the rapid development of transportation in our country, the amount of malignant traffic accidents are also going upward. The accidents which were caused by drivers'fatigue account for 20% of all the traffic accidents and 40% of the serious traffic accidents. Accordingly, the research of a fatigue detecting and controlling system for drivers, especially a car-carried, real-time system, is more and more significant.According to PERCLOS physiological theory, the system in this dissertation judges driver's fatigue by means of supervising the frequency of driver's eye-blinking. The front part receives the image data through a digital camera that is infrared radiation functional, which can provide the round-the-clock infrared light for image obtaining. The part of data processing is based on the ADI Blackfin-533 DSP. The end part adopts the modules of vocal warning and taillight alert. If the system detects that the driver is in fatigue, the DSP will send out a control signal to turn on the vocal warning circuit and taillight alert circuit. The key point to design the system is to locate the eyebrow and eye area and to distinguish the state of eyes. Considering the relative uniformity of driving environment and the features of human face, we adopted Improved Adaboost, a high accuracy and high speed algorithm, to determine the position of the eyebrow and eye area. The combined projection algorithms as well as the geometry character are used in identifying the state of eyes. It is verified that this real time system can obtain the information of eyes image accurately, process and identify it in time. Accordingly, the system achieves the goal to monitor the driver's fatigue effectively.At the beginning of this article, it briefly introduced the system and PERCLOS criterion of fatigue. Then, it brought some descriptions of the design and realization of the software and hardware, along with the comparison of different algorithms. Among that, we put much more emphasize on the Adaboost algorithm, which is used to locate the region of eyes, and mixed projection algorithm, which is used to test the state of eyes. The experiment results were given at the end of the article.