Context-Aware Smart Video-Surveillance for Assistance of Vehicle's Public Safety
Smart Public Safety, Distributed Systems, Ubiquitous Computing, Context Awareness, Object Recognition
Smart video surveillance plays a fundamental role on technological assistance of public safety scenarios, by its potential in allowing detecting events and objects in real-time. This technology attracts special attention in public safety, particularly, in vehicular environments since it offers in mission critical aspects, real-time operations to achieve low time response in detections (assaults, kidnaps, violence, and etc.) with high accuracy and correctness. With efficient smart video surveillance in public safety vehicular scenarios, the expectation is enable the tradeoff between reactive and pro-active authorities actions kept in alert by customized systems and more efficient planning, thus providing the society more quality of life. In vehicular scenarios empowered with broadband wired network infrastructure (regularly deployed), video/audio streaming fits well due to capacity in provisioning high quality video streaming transport towards remote applications. However, this situation is totally different in vehicle environments. By adopting mobile wireless network technologies, there are a lot of challenges (packet loss, increased latency and delays, etc.) to guarantee real-time video streaming in motion. Thus, it is almost impossible to support remote monitoring based smart surveillance applications in terms of ensuring minimum video quality to meet an accurate video processing. In this work, we propose a solution with local video streaming processing for the coupled approach deployed among remote Smart Surveillance applications and multimedia sensors embodying vehicles.