一般設計自動對焦的方法會包含兩部分,一部分是關於所取得之對焦值的計算方法,一部分是對焦最佳點的搜尋法;由於典型的對焦點搜尋方式往往是使用二分搜尋法,此方法之缺點在於不同的對焦環境下,所花費的平均搜尋時間過多,相對地會造成對焦時間過長。 本論文提出將灰色預測理論應用於系統控制方面,並實現在具有自動對焦功能的USB網路攝影機嵌入式系統。此系統中的硬體架構主要包括USB通信、圖像處理的視訊單晶片處理器、系統圖像感測器(Sensor)和嵌入系統軟體組成來驅動音圈馬達(VCM)。系統的軟體部分,在影像訊號處理方面,在所選擇的視訊單晶片處理器中,使用處理器所內建提供的一階微分的計算方法處理USB網照相機中視頻信號的邊緣(Edge)數據,根據得到的數據判斷是否在焦點距離內。使用灰色預測(Grey Prediction)的目的在於改進自動聚焦精確度的控制方法。最後將所提出的演算法與業界中常使用的改良式爬山搜尋法(Modify Climbing Search,MCS)來做比較,可以由實驗結果看出本論文所提的演算法在藉由控制VCM鏡頭模組的操作,完成自動對焦功能的同時,相對地可以減少MCS法對焦所花費的時間近1秒左右。實驗結果證明所提方法之可行性與有效性,它將使USB網路照相機在對焦的使用上變得更快速。
Generally, the design method of Auto-Focusing contains two parts, one is focusing on the value obtained by the method of calculating, the other is the search method of the best focus point. The search method of typical focus point uses binary search, it needs more searching and focusing time in different environments of focusing. This thesis presents the application of gray predication in system control, and the implementation of USB webcam with auto focus embedded system. The main hardware of this system includes USB communications, the single chip processor of image process, image process sensor and embedded system software to drive the VCM. In system software, the video signal processor of image signal processing built-in first-order differential calculation method to deal with the edge data of USB network video cameras. It judges by the edge data whether it’s in the focus distance or not. The goal to use the gray predication is to improve the accuracy of auto focus control method. Finally, we compared the grey predication with the Modify Climbing Search, uses in camera module industry. We can see from the results of the experiment, it reduces about 1 second than MCS while controlling VCM module operation to achieve the auto focus. It proved the feasibility and the effectively of this experiment, it can be more quickly when use USB camera on focusing.