Due to the video surveillance systems on the market only support the functions of screen capture and saving, the theft cases cannot be warned in real time while these cases are occurred. Furthermore, such the cases are usually not detected because of lacking the clues. For this issue, we construct a real-time visual-based surveillance system which can notice the related personnel when the doubtful objects are detected in order to prevent the cases occurred. In this thesis, we adopt the ARM-based embedded development board and image acquisition by webcam to implement the real-time visual-based surveillance system. Additionally, we combine Short Message Service (SMS) in Global System for Mobile Communications (GSM) with our proposed system to demonstrate a video surveillance system which can transmit the warning messages to the related personnel.
