本論文研究開發一套分散式影像處理平台為目標，應用於電腦操作介面 及ARM 系統兩端以分散運算量及降低資料通訊量，考慮過去在影像處理系 統的優缺點，發展一套可移植ARM 的處理及影像辨識演算法函式庫提供開 發者使用。 該架構改良傳輸原始影像的設計，使得影像擷取控制端截取影像後，直 接進行影像處理及影像辨識運算，傳輸影像辨識後參數，當影像擷取裝置 數量變多能夠有效的降低與電腦操作端的運算量及通訊量。利用所開發之 影像平台可供開發演算法驗證結果。 在影像辨識方面，改善弦波對應找線演算法的可移植化及辨識資訊。而 電腦操作介面可嵌入至VMIO 系統，協助未來在智慧型攝像系統影像前置處 理及辨識能夠獲得更多的效益。;In this paper, the research and development of a distributed image processing platform for the goal. Computer Operator Interface and ARM systems both ends to distributed computing capacity and reduce data traffic. Consider the advantages and disadvantages over the image processing system. Development of a portable ARM, processing and image recognition algorithms to provide library developers to use. The architecture design of improved transmission of the original image. Makes image capture control terminal in the interception of the image. Then direct image processing and image recognition arithmetic, transfer images parameters. When the number of image capture device can effectively reduce the increase in the amount of computers and computing and communications volume. Using our developed image processing platform for developers algorithms verify the results. In terms of image recognition, improved sine corresponding linear detection algorithm, the portability and identification information. The computer′s user interface can be embedded to VMIO system to help in the future in the Smart perturbation of the system to do pre-processing and image recognition can get more benefits.
