- 學位論文
使用空間關係與關鍵畫面查詢的多攝影機視訊監控系統
A Multi-Camera Video Surveillance System Use the Spatial Relationship and Content-based Retrieval
摘要
第三代保全監控系統成為全球監控業發展趨勢之後，多攝影機的使用已成為最主要的應用，然而大部分的多攝影機視訊監控系統，仍然採用傳統沒效率的人工方式來監督是否發生異常事件，並且也沒有去注意彼此多攝影之間的協同關係。因此，如何能架構一個多攝影機視訊監控系統，有效增進彼此間關係及檢索效能，成為了一個重要的課題。 本論文提出一個一般環境下對攝影機端(Camera端)以及中央控制端(Central Control端)的架構流程，去描述兩者之間相互溝通以及架構的過程。介紹攝影機端所需處理的事，並當有異常狀態發生時，如何告知中央控制端。以及中央控制端如何與攝影機端相互溝通，並能於多攝影機視訊監控環境下，建立相對空間關係下的空間拓樸，使監控系統具備彈性、適應性以及強韌性；並且能結合關鍵異常畫面的擷取，將異常影片去依照六種特徵，找出最能代表異常狀態影片下的關鍵畫面。如此的攝影機端以及中央控制端的架構，不僅詳知其建構過程，多攝影機的監控系統也能有彼此的協同關係，所擁有的視訊資料庫亦可提供查詢者利用時間關係、空間關係、以及關鍵異常畫面去加以檢索，以建立出一個完整的多攝影機視訊監控系統。
並列摘要
By using a Multi-Camera architecture has become the primary trend of the third generation surveillance system. However, the third generation surveillance system still uses the traditional manner to observe the video events by human. And, the Multi-Camera cooperative relationship is not to be cared. So, how to improve the Multi-Camera construction is an important subject. The thesis is to propose a process to construct the Camera side and the Central Control side under the usual environment. We describe how the Camera side notifies the Central Control side when the exceptional condition happened, and how the Central Control side communicates with the Camera side. Building a spatial topology is to use spatial relationship under the Multi-Camera surveillance system. And, the system is flexible, adaptability, and toughness. We also find the most representative key frame by using six characteristics in the exceptional video. Finally, we can use different items to find the most interesting video in the surveillance video database. In the thesis, we not only know the construction of Multi-Camera two sides, but also build the Multi-Camera cooperative of surveillance system.