People Counting Using Multi-Mode Multi-Target Tracking Scheme

Conventional video surveillance systems often have several shortcomings. First, target detection can't be accurate under the light variation environment. Second, multiple target tracking becomes difficult on a crowd scene. Third, it is difficult to the partition the tracked targets from a merged image blob. Finally, the tracking efficiency and precision are reduced by the inaccurate foreground detection. In this paper, the fusion of temporal and texture background model, multi-mode tracking scheme, color-based difference projection, and ground point detection are proposed to improve the abovementioned problems. In addition, we propose a people counting scheme based on the multi-mode multi-target tracking method on a crowd scene. Experimental results show that the targets on the scene may be detected robustly with the rate above 10 fps and counted with the accuracy above 90%.

1. INTRODUCTION

In the conventional tracking systems, some typical methods are applied to extract the moving objects, e.g., background subtraction [1] and temporal difference analysis [2]. Both kinds of methods are sensitive to the illumination variation and background changing. In the following, the pixel-based temporal probability model [3] is proposed to extract the moving objects. In this paper, we improve the pixel-based temporal statistical model by constructing the spatial-temporal statistical model [4]. By considering the spatial distribution around each pixel (texture distribution) the problem of slight background variation may be overcome.