驾驶员辅助系统的图像模式识别算法研究

汽车已成为现代社会一种不可缺少的交通工具，但是因车辆驾驶而出现的大量交通事故是现代社会所面临的一个严峻问题。驾驶员辅助系统可以帮助驾驶员在行车驾驶过程中提前意识到潜在的危险，进而降低事故的发生率，为保证驾驶员的生命安全起到重要的作用。介绍在该系统设计中关于驾驶员状态检测所涉及到的图像模式识别算法研究，研究的主要问题是检测驾驶员在驾驶过程中是否双手在方向盘上，以确保驾驶安全。算法的关键是实现驾驶员双手位置的准确和实时检测。算法使用阈值在HSV颜色空间对人体肤色区域进行肤色分割，然后对肤色分割区域使用AdaBoost算法进行驾驶员手部的检测。通过软件实验证明，该算法在驾驶员双手检测中快速、便捷、有效。

Research on Image Pattern Recognition Algorithms of Driver Assistance Systems

Car has become an indispensable mean of transport in modern society, but a large number of traffic accidents due to vehicles driving is a serious problem facing modern society. Driver assistance system can help the driver aware of the potential dangers in advance while driv-ing, thereby reducing the incidence of accidents, that plays an important role to ensure the safety of the driver. Introduces the study of image pattern recognition algorithm that detects the driver＆#39;s state in the system design, the main research question is to detect whether the driver＆#39; hands are on the steering wheel while driving, in order to ensure safe driving. Key to the algorithm is to achieve the accurate and real-time detection about the position of driver＆#39;s hands. Firstly, human skin color segmentation is realized with the use of thresholds in HSV color space, and then in the color segmentation region, uses AdaBoost algorithm to detect the driver＆#39;s hands. Software experiments show that the algorithm is fast, convenient, and effective to detect the driver＆#39;s hands.