STTR: A system for tracking all vehicles all the time at the edge of the network
STTR: A system for tracking all vehicles all the time at the edge of the network Xu et al., DEBS’18
With apologies for only bringing you two paper write-ups this week: we moved house, which turns out to be not at all conducive to quiet study of research papers!
Today’s smart camera surveillance systems are largely alert based, which gives two main modes of operation: either you know in advance the vehicles of interest so that you can detect them in real time, or you have to trawl through lots of camera footage post-facto (expensive and time-consuming). STTR is a system designed to track all of the vehicles all of the time, and store their trajectories for ever. I certainly have mixed feelings about the kinds of state surveillance and privacy invasions that enables (it’s trivial to link back to individuals given trajectories over time), but here we’ll just focus on the technology. Since the system is design with pluggable detection and matching algorithms, then given some calculations around volume it ought to be possible to use it to track objects other than vehicles. People for example?
Assuming the availability of suitable detection and matching (figuring out if Continue reading