Design Of Automotive Fatigue Driving Warning System Based On Improved Adaboost Algorithm

With the popularity of cars, people are no longer satisfied with the function ofa car as a mere travel tool, more and more people are interested in cars that provideunique safety features. To meet this demand, an warning system for fatigued drivershas been proposed which draws on advanced foreign design concepts.In order to satisfy people’s pursuit of automotive fatigue driving safety, withthe analysis of the structure and working principle of automotive fatigue drivingwarning system, image capture and processing system based on high-speed digitalsignal processor DSP is proposed. Because of the requirements that are non-contact,real time and all-weather of the warning system, electronic component selection,image acquisition process analysis, software algorithm design and debugging of theprogram are formed for the image acquisition module, then this automotive driverfatigue hardware circuit modular is designed. The principles of histogramequalization, median filtering and face recognition on the image preprocessing arestudied, then the image histogram equalization, median filtering and facerecognition software algorithm are designed. In the design process, the design anddebugging of the software algorithm goes in CCS.This system uses the machine vision method. It gets video sequences of adriver from the CCD camera. To achieve real-time processing of face images,cascade classifiers and Gentleboost-based strong classifiers based on MBLBPfeatures are used. And then it detects the face, eyelid. Finally, it uses the PERCLOScriterion to determine the state of drowsiness. The experimental results show thatthis system has higher accuracy and speed, and can satisfy the demands of non-contact, all illumination condition and real-time monitoring.Face recognition is the most important part of the fatigue driving warningsystem, which directly affects the real-time, accuracy and sensitivity of the system.In this paper, firstly, I design hardware compositions and software flow according toearly warning design theory and requests of this system; then I write and debugalgorithm of face recognition through CCS; finally, I optimize the designing ofsoftware algorithm and achieving face recognition according to the optimizedalgorithm.