分散式工業影像監測系統之研製

Abstract

本論文的目的，在於建立一個以網路為基礎的即時影像系統，並利用此系 統來執行影像識別與機械手之連線。用網路連結系統的各個影像模組，有 下列幾項優點：(1)系統使用數個影像模組同時處理多架攝影機的影像資 料，或是一架攝影機的畫面分由數個影像模組處理，達成多視窗同時處理 的目標；(2)系統以網路結合數個影像模組，並行處理影像資料；(3)系統 中各個影像模組的獨立性提高，因而使得模組在設計上更為自由；(4)模 組間的接線少，因此容易拆裝；(5)使用者可自由選擇所需的模組組織系 統；(6)不論在硬體或軟體上，維護與擴充都很容易。由於半導體工業的 進步，我們藉由硬體上的設計來提升影像系統處理動態影像的能力，除了 將網路應用於影像系統外，各個影像模組中相關的數位電路，都是以PGA( Programmable Gate Array)完成，不但縮減硬體之體積、提高穩定性，同 時也使電路容易修改，增加設計的彈性。 The goal of this thesis is to develop a network-based real- time image system , and we use this system to combine pattern recognition and robot arm . Connecting the different image modules under the network has the following advantages : (1) It is possible for multiple camera application or multi-window tracking scheme used multiple modules with the same camera ; (2) The system combine multiple modules with the network and process image data in parallel ; (3) The modules in the system can be designed individually and independtly ; (4) The number of interconnection lines between modules is decreased so that it is easy to assemble or dismantle the modules ; (5) The user can freely choose what modules he needs to form a spicific system ; (6) Whether hardware or software , it is easy to maintain and extend . Because of the progress of the semi- conducted industry , we uplift the ability of dynamic image processing by means of the design for hardware . In addition to applying network on the image system , all the digital circuit is done by PGA (Programmable Gate Array) . On the one hand , it retrenchs the dimension and increase the stability . On the other hand , it is easy and flexible to rework .