大部份的家電都附有其專屬的遙控器，原因在於它為家電控制帶來更多的方便性。然而你可能遇過這樣的經驗，想要看電視但找不到遙控器，或是很多個使用者同時一起看電視，為了操控家電而反覆傳遞遙控器，這些狀況帶來了許多麻煩。因此，近年來智慧型家電遙控成了熱門研究的主題。本論文所設計的系統讓使用者不用佩帶任何發射器或感應器，直接利用拍手聲以及手勢達成遙控家電的目的。本系統可以辨識出 12 種靜態手勢與 6 種動態手勢作為控制的指令。 首先，要遙控家電的使用者連續拍手三次，系統抓取環境中的聲音訊號並進行音頻分析，辨識是否為拍手訊號。若是拍手訊號則以模糊控制找到使用者的位置，攝影機會旋轉到使用者前方並且擷取使用者的原始影像。經過影像前處理後，得到完整的手勢影像並加以辨識，不同的手勢影像代表不同控制家電的指令。最後，控制指令會轉換成特定頻率的紅外線訊號，操控所指定的家電。 本論文主要針對以下的主題進行研究，包含拍手訊號的辨別、聲源定位的模糊控制、手勢影像辨識的處理以及紅外線訊號發射模組的製作。 To control home appliances conveniently, most of home appliances have their own remote controllers. However, you may have the following experience such as: you want to watch TV but you can’t find a remote controller. Or many users to watch TV simultaneously, the remote controller may need to be passed to each other for controlling the TV. Due to the above inconveniences, intelligent remote control for home appliances has become much more popular in recent years. In this thesis, we design a system in which the users can control the home appliances intelligently. Based on the sound of handclap and some hand gestures, the user can achieve the above goals without taking any transmitter or sensor. The system can recognize twelve static gestures and six dynamic gestures which stand for different commands. First, the user applauds three times continuously; the sound signal is captured and recorded. Then the system identifies whether the signal is a hand clap or not. If it is a hand clap, the system can use fuzzy control to find the user’s location. Once the use’s position is detected, the camera rotates toward the user and captures the original image of the user. After image pre-processing, the gestures of the user is captured and recognized. Different gestures denote different commands for remote control of the home appliances. Finally, the recognized commands are transferred to control the home appliances by the infrared rays of specific frequency. In this thesis, the following topics are studied, such as signal identification of hand clap sound, sound source location by fuzzy control, image processing for gesture recognition and the making of the infrared rays module.
