Distributed Sensing and Processing for Multi-Camera Networks

TitleDistributed Sensing and Processing for Multi-Camera Networks
Publication TypeBook Chapters
Year of Publication2011
AuthorsSankaranarayanan AC, Chellappa R, Baraniuk RG
EditorBhanu B, Ravishankar CV, Roy-Chowdhury AK, Aghajan H, Terzopoulos D
Book TitleDistributed Video Sensor NetworksDistributed Video Sensor Networks
Pagination85 - 101
PublisherSpringer London
ISBN Number978-0-85729-127-1

Sensor networks with large numbers of cameras are becoming increasingly prevalent in a wide range of applications, including video conferencing, motion capture, surveillance, and clinical diagnostics. In this chapter, we identify some of the fundamental challenges in designing such systems: robust statistical inference, computationally efficiency, and opportunistic and parsimonious sensing. We show that the geometric constraints induced by the imaging process are extremely useful for identifying and designing optimal estimators for object detection and tracking tasks. We also derive pipelined and parallelized implementations of popular tools used for statistical inference in non-linear systems, of which multi-camera systems are examples. Finally, we highlight the use of the emerging theory of compressive sensing in reducing the amount of data sensed and communicated by a camera network.