| In recent years, the global traffic accident death toll continued to rise. When analyzing the cause of traffic accident, most of overseas and domestic experts think that artificial factor is the most important factor in the three factors of human, vehicle and road. To reduce such incidents, the tutor of writer promotes the application of scientific and technological achievements of the project "The Intelligence Device for monitoring the Driver State"in Shanxi Province in 2006. Carry on to real-time supervise the driver state (Including whether the driver was still in, whether fatigue driving, drink, etc.) on the basis of 5 patents which already gaining, and implement the control of different class according to the result of supervision. The controls include the hint, alarm and brake.This paper is the son of the subject, its main task is to judge the driver fatigue state, and give single chip the outcome of judgments, the single chip implement relevant control according to the outcome (other members from the project team complete the part of control). Therefore, the task is the key element in monitoring the driver state. Based on my national invention patents "the device and method of fatigue driving state monitoring" (ZL: 200610012623.6), use the driver eyes state as the fundamental basis of identifying the fatigue, design the system for supervising the driver fatigue with the DSP TMS320DM642 as the core processor in the text. Carried on following several aspect works:1. Study and improve the method for identifying eye state on the base of checking a large amount of literature.The method for identifying eye state that has been which not only compute largely, but also being affect by the environment factor(for example illumination) easily. Therefore, put forward a kind of new method for identifying eye state in the paper. Firstly, use movement information and skin color characteristics to detect the human face on the real-time images. Secondly, use the Adaboost algorithm examine human eye region on the upper half face region which was examined. After the examinations divided this region makes the normal opened eye template with the eyebrow and the eye. Finally, carried on the match with the upper half human face region which was not examined the eye. When this region biggest similarity value is smaller than some hypothesis threshold value, we thought the eye is close; otherwise thought the eye is open. And calculate the mean time which eye closed and the PERCLOS value in unit-time according to the eye state. Then, carries on the comparison with the hypothesis threshold value to judge the driver is whether or not fatigue.2. Carried on hardware design and the correlation software of the system for supervising driver fatigue.Design the hardware platform with TMS320DM642 as the core processor, aim at the need of the algorithm of identifying fatigue. The hardware platform includes video capture circuit, external memory, the hardware interface between single chip with DSP, power supply circuit, and also on. Design the software based on DSP/BIOS on the hardware platform which include three parts: setting BIOS, initializing the system and implementing the BIOS task.3. Carry out large amount of simulation experimentCarry on large numbers of simulative experiments based on above works. The result of experiment indicate, the judgment method of the driver's fatigue state is higher accurate rate, can suffice the request of system which supervise the driver's state of fatigue. |
PDF Full Text Request