摘要
道路上許多意外，常常來自駕駛人打瞌睡，造成死傷。如何降低交通事故，減少傷亡，一直是全世界各國努力的目標。本研究首先整理出各種疲勞偵測方法，包括腦波、脈搏、心跳、血壓及眨眼變化，比較其優缺點，再針對影像處理法偵測眨眼變化部份，國內外發展現況，作進一步詳述。本論文導入影像處理理論，包括影像追蹤、連續影像相減、RGB色彩系統、HSV色彩系統、灰階轉換、二質化等理論，並結合自行架設之硬體系統來進行實驗，利用C++撰寫程式所完成之軟體功能，包括臉部位置搜尋、眼睛位置計算、眼睛狀況判斷及駕駛人疲勞狀況判斷。實驗結果所進行之分析與探討，初步證實了系統的可行性。
並列摘要
Road accidents often result in injuries and deaths, many of which are caused by dozing drivers. How to reduce victims by the accidents of this kind has become a long term issue all over the world. This work firstly gives a survey on drowsiness detecting systems, inclusive of those by the brainwave, pulsation, heartbeat, and eyelid blinking frequency, followed by focusing on the detail of eyelid blinking detection by image processing and its current development trend. This thesis then introduces an image processing system, with the application of techniques of image tracking, successive images substraction, the RGB and HSV color systems conversion, and grayscale and binary transformations. The corresponding hardware is then set up with software by C++ programming language. Experimental results show that the system is capable of determining the levels of the driver’s drowsiness. The feasibility of this drowsiness detecting system is therefore verified.