A Fast ModelFree Object Tracking Algorithm（一个快速模范自由对象跟踪算法）.pdf

A Fast Model-Free Morphology- Based Object Tracking Algorithm a b a Jonathan Owens , Andrew Hunter Eric Fletcher aSchool of Computing Technology University of Sunderland, UK jonathan.owens@sunderland.ac.uk bDepartment of Computer Science Durham University, UK andrew1.hunter@durham.ac.uk Abstract This paper describes the multiple object tracking component of an automated CCTV surveillance system. The system tracks objects, and alerts the operator if unusual trajectories are discovered. Objects are detected by background differencing. Low contrast levels can present problems, leading to poor object segmentation and fragmentation, particularly on older analogue surveillance networks. The model-free tracking algorithm described in this paper addresses object fragmentation, and the object merging that occurs when proximate objects segment to the same connected component. 1 Introduction Automated visual surveillance aims to provide an attention-focussing filter to enable an operator to make an optimum decision whenever an unusual event occurs [1]. This is achieved by directing the operator’s attention only to those events classified as unusual. The backbone of such systems typically comprises something like the processing pipeline shown in figure 1. Image Object Object Object Behaviour Scene Acquisition Segmentation Tracking Classification Classification Description Figure 1: Typical image processing pipeline for automated video surveillanc