A depth-based polar coordinate system for people segmentation and tracking with multiple RGB-D sensors
Emilio J. Almazán, Graeme A. Jones · Research Repository (Kingston University London) · 2014
The work presented in this paper provides a framework for monitoring wide area indoor spaces built from multiple RGB-D sensors. A polar coordinate space representation of the common ground plane is proposed to effectively aggregate the data from all sensors. Additionally, it provides with capabilities that mitigate the main issues related to RGB-D sensors i.e. limited range, degradation of resolution with distance and increasing noise. For multi-target tracking, the use of a discriminative appearance model is a chief aspect in order to properly distinguish people from one another, especially where the density of people is high. A multi-part appearance model is presented in this work the chromogram – which combines colour with the height dimension offering high discriminative capabilities. To effectively evaluate this framework a new challenging dataset is presented that contains multiple occlusions and people interactions aiming to cover the relevant aspects of multi-target tracking systems.