Visual Surveillance

Visual data captures more information than data captured by other type of sensors. The visual surveillance and analytics solution can be configured to automatically detect activities in an area under observation of a camera. The camera records this information in a database and utilizing this recorded information, the system allows users to search for activities of interest as well as provide analytical reports to understand the inherent patterns and trends.

Business Challenges

With the rise in criminal and anti-social activities, there is an increased pressure on organizations and government agencies to incorporate a higher level of security within their systems. Even in day-to-day business, the need to understand the trend of the market and requirement of the customers has become quite essential. So, the need of the hour is to have a system that can automatically detect and track movements of objects (vehicles and humans) in real-time.

Overview 

The major components of the visual surveillance solution are:

• Visual Surveillance: Image processing that involves detection, tracking, and location of objects of interest.

 Analytics: Interpretation of detected motion, i.e. modelling of unusual behaviour, detection of deviant activities, determination of behavioural patterns, and so on.
The system surveillance runs on Linux. Currently the system supports single camera surveillance and work is in progress for multi-camera tracking.

Benefits

The computer vision based automated visual surveillance and analytics system can detect movements within the field of view of a camera and stores all relevant information in a database. It allows the users to query on the database, hence pre-empts hours of manual effort in monitoring hours of video footage. The system can provide real-time alerts and also detect suspicious or unusual activities.

Such a system not only increases the accuracy in visual surveillance but also reduces the cost of manual inspection in the long run. The analytics presents graphical charts to the users based on the search queries, which equips them to understand trends and patterns of activities occurring over a period of time and draw conclusions and take pro-active measures

Use Case Scenarios

Visual Surveillance can find several possible industrial applications such as:

• Security and monitoring

     _ in corporate offices, public places, secluded and prohibited places

• Law obedience

      – wrong parking of vehicles, jumping lanes, taking wrong turns, wrong lane driving

      –  multiple users in ATM counters

• User behavior study - exploring statistical patterns/trends

      –   in retail stores, shop-floors

      –   can be used for traffic planning

• Stopping pilferage

      –   number of cars crossing a toll-point vs revenue

      –   number of units crossing a point on manufacturing assembly line

• Suspicious activity detection

      –  someone dropping or picking up an object (association-dissociation amongst objects)

      –  someone following unusual entry/exit pattern

      –  car appearing/exiting from wrong place