適合影像處理的串流模組設計

本論文是以TSSA ( Trimedia Streaming Software Architecture ) 軟體架構所建立的模組結構以及其模組化的串流方式為探討的重點。模組化的概念是將不同性質的影像處理程式區分開來，讓各模組具有單一化的執行功能，並且具備獨立的資料結構以及執行函式。此外模組會以單純的介面函式來與應用程式以及系統連結，提供給外界簡易的操作方式，而模組本身的影像處理內容，則是透過模組內部的函式來呼叫執行，不讓外部的應用程式或者其它模組直接來操作，避免模組本身的資料內容遭到修改，給予模組多一層的保護。TSSA模組的串流方式是透過傳輸模組來運作。在兩兩影像處理模組之間會連接一個傳輸模組，而傳輸模組上的傳輸條件以及傳送單位，則會根據前後連接的影像處理模組特性來設定。當模組之間要傳送影像資訊時，會透過系統函式來取得傳輸模組上的封包，提供給影像處理模組來讀取或寫入資料。這部分的運作機制會透過對應用程式執行上實際的測試來了解，此外會再藉由於影像處理模組中增添新的函式功能，對模組的運作方式做進一步的了解與應用。

關鍵字

模組化；串流

並列摘要

The primary issue of this thesis is the component architecture and streaming of TSSA ( Trimedia Streaming Software Architecture ). The concept of modular is separating different types of image processing programs into several parts each with its own data structures and functions. In addition, each component would be linked to the application program and system by some simple interfaces, which provide an easy way for users to operate. Meanwhile, components’ image processing procedure would be executed by the internal functions to avoid modifying from application programs or other components. The streaming of TSSA components is operating on the in-out descriptor, which is generalized as a connection component. The in-out descriptor is connected between two image processing components, and the transmitting conditions of in-out descriptor are based on the properties of components it’s connected. While the image processing components need to transmit data, the packets of in-out descriptor are provided via system functions. The mechanism of communication between components would be verified through the experiment of some programs and an additional function would be added into the component to further clarify the operations.