With the ever growing requirement for safety, more and more cameras have been applied in surveillance applications. Most previous work on visual surveillance has focused on single camera processing including object detection, tracking, classification and activity recognition. Tasks under each cameras field of view are independent from each other. Instead of having a high resolution camera with a limited field of view, multiple cameras provide a solution to wide area surveillance by extending the field of view of a single camera. One problem associated with a multi-camera system is to automatically analyze and fuse information gathered from multiple cameras so that human intervention is reduced to a maximum extent. In this thesis, we concentrate on the problem of continuously tracking objects across multiple non-overlapping cameras. The key issues include: (1) Object representation and matching, (2) Topology estimation and (3) Data association across cameras. Specifically, the following topics are addressed in the thesis: 1.We propose a new solution to the problem of appearance matching across multiple non-overlapping cameras. Objects of interest, pedestrians are represented by a set of region signatures centered at points sampled from edges. The problem of frame-to-frame appearance matching is formulated as finding corresponding points in two images as minimization of a cost function over the space of correspondence. The correspondence problem is solved under integer optimization framework where the cost function is determined by similarity of region signatures as well as geometric constraints between points. Experimental results demonstrate the effectiveness of the proposed method. 2.Since moving objects do not move around the scene randomly, by observing the targets behavior in a short period of time, conclusions can be drawn about the semantic model of the scene. The semantic model of multiple cameras is represented by a topology graph. The nodes in the topology graph are defined as entry(exit) zones in each camera. The connections in the topology graph are used to indicate the connectivity between nodes in a network of cameras. We propose an unsupervised method for recovering the topology of multiple cameras with non-overlapping fields of view. The topology graph of multiple cameras helps to predict the reappearance of moving objects. If one object disappears from the node of one cameras field of view, we only need to search the reappearance of t...
修改评论