Automatic detection of security risks in our streets
Friday, June 8th, 2012
Author: Zeev Smilansky – Emza Visual Sense, Israel.
Security is a growing concern around the world. There are no quick and easy solutions. One concept that is becoming accepted around the world is that of surveillance cameras. In many places, people have become accustomed to the concept that a surveillance camera is a solution for reducing crime and alerting for dangerous situations. Around the world, cities are installing surveillance cameras to improve safety and security (200,000 in Moscow, millions in London, 100,000,000 sold annually) around the world. However, almost always, these cameras are – as their name implies – recording devices and the video they captured is stored and almost never looked at.
SQS is participating in a project that proposes a new approach, instead of sending and storing raw video, the system will generate real-time alerts in order to prevent dangerous and criminal events; will require low infrastructure in terms of power supply, communication bandwidth and storage; be simple to install and operate; and help city officials improve the safety and security of their city.
To develop the system, we needed benchmark movies. Two approaches were considered. The first approach is recording real urban scenes. This can be very difficult and time consuming if you count on finding “natural” events, or expensive and complicated if you plan to “direct” events with actors and attendants. The second approach is creating animated movies. For simple cases, this is certainly possible. For more complicated scenes, the effort required can quickly become prohibitive. In the end, we started with simple animations (created by us (SQS)) and move gradually to real-life captured movies. A frame from one of these simulated movies is shown on the left.
The project is now at the middle of its duration. Next year we will see if the performance matches the expectations.
Link to HUSIMS project: http://projects.celtic-initiative.org/husims/index.html


Can a large organization adopt agile approaches to software development when the organization holds the notion that not all projects should be agile? In other words, can there be a mix of waterfall-type projects and agile projects in the same organization? The short answer is yes, however there is a cost that must be paid for this coexistence.

