Robust 3D people tracking and positioning system in a semi-overlapped multi-camera environment

Abstract

People positioning and tracking in 3D indoor environments are challenging tasks due to background clutter and occlusions. Current works are focused on solving people occlusions in low-cluttered backgrounds, but fail in high-cluttered scenarios, specially when foreground objects occlude people. In this paper, a novel 3D people positioning and tracking system is presented, which shows itself robust to both possible occlusion sources: static scene objects and other people. The system holds on a set oí múltiple cameras with partially overlapped fields of view. Moving regions are segmented independently in each camera stream by means oí a new background modeling strategy based on Gabor filters. People detection is carried out on these segmentations through a template-based correlation strategy. Detected people are tracked independently in each camera view by means oí a graph-based matching strategy, which estimates the best correspondences between consecutive people segmentations. Finally, 3D tracking and positioning of people is achieved by geometrical consistency analysis over the tracked 2D candidates, using head position (instead of object centroids) to increase robustness to foreground occlusions.

Estimation of 3D head region using gait motion for surveillance video

“…The first is a 3D assisted approach which uses view geometry and a 3D human shape model. Mohedano et al [2] built a multi-camera geometry-based 3D tracker which uses multi-dimensional background subtraction and human template correlation. This method detected people even when they were occluded by static foreground objects.…”

Resource-aware distributed particle filtering for cluster-based object tracking in wireless camera networks

“…If multiple cameras, each with different viewing area and direction, collaborate on object tracking, one can expect to have a larger tracking coverage with more robustness against occlusion [1][2][3][4][5]. Fig.…”

“…Let us say e is the energy required to transmit a packet. Then, the total transmission energy E T X is expressed as 5) where N T X is the number of transmitted packets. Similarly, the amount of wasted energy E N R is computed as 6) where N N R is the number of lost packets.…”

“…They are mostly categorized into direct information fusion and probabilistic information fusion. In the direct information approach [3][4][5], 2D trackers (single view based trackers) are executed individually and the final 3D track is estimated by combining the 2D tracks with geometric or appearance constraints. Let us sayx i,t is the 2D track result of the i th camera and …”

“…The main bottleneck of developing object tracking algorithms is that a camera may have insufficient target information caused by object occlusion and by the indistinct nature of the object in front of cluttered backgrounds. Recent studies of object tracking methods using multiple cameras [1][2][3][4][5] rely on the concept of collaborative tracking of a moving target to overcome these difficulties.…”