VIDEO SURVEILLANCE & VIDEO ANALYTICS

Video surveillance, access control, and alarms have much in common, and they often work together as an integrated system to keep intruders out of secure areas, limit access to security-sensitive areas, and remotely monitor critical areas to reduce the risk of crime and security incidents.

Today, lots of video images are being used around us as cities and technologies have been developed.  Various mobile equipment and black boxes are being widely used,  and we are receiving many different information from them.  Also, CCTV cameras are in every corner of the city to cover areas that are constantly being monitored and managed by different institutions and organizations  But this also means that finding out the very information you need from mass video data and properly responding to situations are getting more and more difficult.  To handle that Sureworks analyzes video and offers customized information to clients. 

Now, let’s see how  Video Analytics works in real-life.  Here, a child has gone missing in a large and crowded shopping mall.  How can we find this child from many different monitors filled with people?  First, we use the Complex Event Search function here, we enter information on the missing child such as the time she went missing,  last known location, and her height and clothing colors.  Also, for a more accurate search, we can add a searching criterion by drawing her expected movement direction using Visual Query function.  Visual Query is a unique function that enables visual criteria-based searches rather than text-based ones.  Now, we’ve narrowed down lots of videos into a few,  containing an object resembling the child’s looks and wearing.  And after a moment, we finally get the child’s facial information.  It’s now time to use the White-List face recognition function to track the girl’s route.  White-List tracking allows you to detect facial data that has not been previously saved in the database and track the route of your target entity, all in real-time.  Now you confirmed the girl’s route and current location.  Now, all the information is transferred to an onsite agent’s mobile device for him to confirm the missing girl’s face, movement route,   and current location, and quickly find the girl. 

Also, we can analyze the Grouping event and Path-passing event so that it would give alarms when more than a certain number of people are gathering in the square or an object is moving towards the square.  However, if it gives an alarm every time either situation happens,   there would be over-detection of alarms which are not related to the illegal demonstration.  In such situations, the Complex Event Processing function will help you predict the situation you want to find more effective.  By binding the 2 situations together with the search query operator ‘and’ using GUI Tool,  Video Analytics activates alarms only when the 2 events occur simultaneously.  Then tie all 3 cameras again with the operator  ‘and’ so that it would give alarms only when the 2 events happen at all 3 entrances.  By doing this, the false alarm rate goes dramatically down,  and predicting the very situation you find gets much efficient.  As you saw, Samsung SDS Video Analytics does more than just collect video and image data with the object, face, and license plate recognition functions,  it also processes and analyzes data from them and provide valuable information to clients.  And this can be utilized for more specialized purposes across various sectors such as urban surveillance, facility, retail, manufacturing, healthcare, finance, et  With its Video Analytics solution, Samsung  SDS has proved its extraordinary competency through Moscow Safe City Project in Russia and Hi-1 resort casino in Korea.  Furthermore, Samsung SDS is following its roadmap to increase the accuracy of analysis and quickly process a mass amount of data by interfacing various sensors and devices to videos.