Guest Editorial Computational and smart cameras

Abstract

Recent advances in camera hardware, processing capabilities, storage capacities, and communication technologies suggest an evolution of traditional multi-camera systems into networks of highly capable computational and smart cameras. These networks can be seen as visual sensor nodes that combine sensing with onboard processing and storage, and that are able to communicate with other similar sensors in the vicinity. As a future development step, these visual sensor nodes will have the necessary control and coordination machinery to set up ad-hoc sensing networks, capable of capturing, analyzing, and storing large amounts of image and video data over extended spaces to support a variety of applications. Opportunities for application development abound in domains as diverse as smart environments, wearable sensing, security and surveillance, environmental monitoring and disaster response, traffic management and urban sensing, elderly care and wellness monitoring, geo-scale monitoring, etc. These networks must also exhibit a great deal of autonomy as the sheer scale of the networks and the tremendous amount of data present in them render human-in-the-loop operation for the most part infeasible. Much work is still needed to realize this vision of ad-hoc networks of computational and smart cameras capable of providing perceptive coverage of wide areas with little or no human supervision. The goal of this special issue is to bring together papers dealing with various aspects of this vision.