Driver Fatigue Estimation Using Image Processing Technique

A Fatigue Detection system has been\uddeveloped using non-intrusive vision based approach. The\udsystem uses a Logitech USB camera which points towards\uddriver‟s face and monitors face and eyes to detect driver\udfatigue. The system is developed on Linux operating\udsystem and used DM3730 processor as hardware. The\udalgorithm is developed to estimate whether eyes are open\udor closed and fatigue is estimated using PERCLOS\udmethod. Normal human blinks eyes 12 times in a minute.\udIf the eyes are closed for 15 consecutive frames in a\udminute or if PERCLOS > 80% than system issues warning\udto stop the vehicle. The algorithm is tested on 45 different\udpersons i.e., 15 women, 15 men and 15 persons wearing\udspectacles and the detection rate is 99.2%. The system\udtakes <5ms of time to detect whether eyes are open or\udclosed and the hardware used is small in size and easily\udimplementable