Skip to Main Content
Networks of visual sensors have recently emerged as a new type of sensor-based intelligent system, with performance and complexity challenges that go far beyond those of existing wireless sensor networks (WSN). For many years now, networks of cameras have been used for surveillance and security. These networked cameras, unable to process any data locally, would send their individual streams to a centralized location and rely upon human vision and cognition to detect events and anomalies. Advances in computational technologies have allowed for much improvement in the hardware of these systems, such as higher picture quality and motorization capabilities. However, only recently have there been attempts to integrate the latest research developments in human and computer vision into current sensor technologies. The ultimate goal of visual sensor network (VSN) designers is to enable the cameras (nodes) to perform more advanced vision tasks locally, in a fully automated way. When combined, these cameras would constitute an intelligent visual system capable of detecting not only objects within their range, but also the events created by these objects.