基于双核DSP的列车司机驾驶疲劳检测系统研究-计算机应用技术专业论文.docxVIP

优秀毕业论文 精品参考文献资料 ABSTRACTThe ABSTRACT The safety traffic of locomotive 1S the most important and core part of railway security．Especially due to the development of heavy load train and high speed train in our country,adds much more requirement for the safety traffic of locomotive．Driver fatigue driving is one of the major causes of locomotive traffic accidents．How to effectively detect and prevent driver from fatigue driving has much real significance to reduce railway traffic accidents and personnel mortality． Firstly,the cause of driving fatigue of locomotive driver and some detection methods are analyzed in this thesis，and the machine vision based methods are selected．According to the characteristics and correlation of the fatigue，the PERCLOS(Percent eye closure)is adopted as a principle to determine the degree of fatigue for real—time detection． Thereby,the main scheme of locomotive driver fatigue detecting system is proposed． Secondly,a set of integrated locomotive driver fatigue detecting algorithms are designed．In these algorithms，the gauss distribution model is used to detect human face，while a weighted hybrid projection function is proposed for the location of eyes，which determines the weight of projection functions based on certain conditions and makes the location more accurate．Accordingly,the deformable templates are used to detect and track the contour of eyes．In order to improve the convergence speed of the energy function of deformable templates and the match accuracy,a method is advanced，in which the SUSAN operator is used to extract eye comers firstly,and then the deformable templates are adopted to match eyelids，lastly the locomotive driver fatigue is judged by calculating the value of PERCLOS continuously． Then，the dual·core DSP based locomotive driver fatigue detecting system is build up，and the technologies of realizing and optimizing such system are given．In the system，CCD camera is used to collect the face images of locomotive driver which are