本論文是以人臉辨識為目的,進行嵌入式系統的開發。我們選擇具ARM架構處理器的pcDuino V2為硬體平台,使用Linux內核的Ubuntu為作業系統,並於硬體周邊設置一個攝像鏡頭以截取人臉圖示,以便讓辨識程式能利用人臉的多重生物特徵,執行辨識動作。系統中負責執行人臉辨識功能的程式是基於Haar-like特徵和AdaBoost算法所開發的,Haar-like特徵是作為辨識的依據,AdaBoost是機器學習的演算法,用於訓練分類器。由實驗結果得知,在不同情境下,所開發的嵌入式系統能以小於一秒鐘的時間,迅速辨識出受測者的身分。因此,能保證此嵌入式系統的即時性。
For the purposes of facial recognition, in this thesis, we design and implement an embedded system whose platform consists of pcDuono-V2 board with ARM-processor inside and a Linux-kernel-based operating system, Ubuntu. A camera is set up on the platform to take human face pictures, and a program is used to recognize these pictures via biological multi-features. The corresponding facial recognition program running on the embedded system is based on Haar-like features and AdaBoost algorithm. Haar-like features are the foundation of recognition, while AdaBoost is a machine learning algorithm for training classifiers. According to the experimental results, the resultant embedded system can recognize the experimental subjects during one second in every our considered situations, which ensures the real-time performance.