| 摘要: | 隨著綠色環保意識逐漸受到重視，智慧生活的議題也被廣泛的討論。本研究探討智慧生活中如何應用數位化影像技術來偵測室內人數及位置，再結合物理學上的模擬退火演算法來實作智慧型空調系統，以便於在不同的情境及負載需求下，以最合適的方式配置系統，如此不但能達成系統耗能減少，並能同時兼顧居家生活的舒適性。本研究採用方法為系統實證法，透過實地架設系統平台來進行情境測試。所採用的辨識方式在室內幾乎可以和人體姿態模型方式達到一樣的效果。本研究利用攝影機擷取彩色畫面，經由灰階處理、二值化及低通濾波等方式把畫面轉換成黑白圖片，進而計算物體個數來達成物體偵測的目的。物體偵測結果再加上房間大小與室溫等參數，可經由模擬退火法計算成數據作為室內空調調整的依據，藉此使智慧型控制更為人性化。現在一般家用冷氣大多是需要人為控制，如設定風速大小、冷氣強弱等。我們藉由物體偵測、輸入理想溫度等，結合模擬退火法形成一個物體偵測介面來代替現今冷氣仍須手動輸入的部分，經由實驗結果證實此物體偵測技術再配合模擬退火法能夠有效代替冷氣手動部分。經由實驗我們驗證智慧屋用於家電管理確實能夠達到智慧型控制、便利與人性化的目標。
With environmental awareness becomes an important issue, the topic of intelligent living has also been widely discussed. We investigate the application of digital imaging technology to detect the number and location of interior parties. Combined with simulated annealing algorithm, we implemented an intelligent air conditioning system, which can adjust itself automatically to the most appropriate configuration in different contexts and load demands. The system will not only achieve system energy optimization, but also take the comfort of living into account. In this study, we actually set up the system platform in the real environment, and mainly focus on the implementation of indoor object detection and the temperature control. We use gray scale, binary, and low-pass filtering to achieve image recognition and thus to determine the number of people in the room. Together with the room size and the object size, we are able to calculate the location of the party, which can then be used as the basis for adjusting the air-conditioning. The identification result is used by the simulated annealing algorithm. Together with the size and the temperature of the room, one can achieve a more humane and intelligent control. |
