本論文最主要利用線性形狀記憶合金做為致動器，搭配光學成像系統，架設出對焦平台，使用ARM微處理器加上達靈頓放大電路來控制電流，再配合不同的控制法則來達到自動對焦的目的。 本實驗設計最主要分為硬體架構和軟體開發兩大部分，硬體部分利用形狀記憶合金線在低溫態時預先施加外力使其變形伸長，ARM微處理器加上達靈頓放大電路輸出PWM訊號來控制電流，合金線通電加熱至高溫態後麻田散體相逆轉變成沃斯田體相產生記憶效應，恢復原來形狀來帶動成像鏡頭組，達到對焦致動器的往復來回運動。 軟體的部分包括找尋對焦值作為被動式對焦演算法中的性能指標，來判斷影像的清晰程度，和使用Borland C++ Builder (BCB) 做為人機介面，比較對焦前與對焦後的影像清晰程度差異。本研究的對焦法則使用基因演算法、模糊控制、固定及調變步伐控制。 最後，分別介紹並呈現三種對焦控制法的對焦實驗結果，並比較三種對焦法則的優缺點。 In this research, the wire-type shape memory alloy (SMA) is used for the autofocusing actuator. The main experimental design is divided into two parts. One is the hardware setup another is the software development. At the part of the hardware, we use an actuator and an optical imaging system to set up the focus platform. The temperature is controlled by ARM microprocessor through a Darlington amplifier circuit. At the part of the software, we are looking for focus values as passive autofocus algorithm’s performance indicators to determining the clarity of the image. And we write the Borland C++ Builder (BCB) as a user interface, comparing the image clarity difference between before focus and after focus. Furthermore, we try genetic algorithms, fuzzy control, steady and unsteady steps to be the rules of focusing to achieve the purpose of autofocus. Finally, we introduce and present the experimental results, and compare the advantages and disadvantages of three focus rules.
