This video analytics technology, with licence plate recognition with high speed deblurring, sits on a surveillance system of CCTV cameras and Digital Video Recorders (DVRs), giving it 'intelligence' by monitoring actively within the entire camera view. It transforms an ordinary passive system into an active one.
The core engine has the ability to detect, track, trace and alert an operator by means of visual and audible modes, on any individual or group present within a surveillance zone. Any unattended objects such as bags and people who stop moving along its path will be detected and an alert sent to the operator. Additional features are available to minimize false positives. External interferences such as reflections of raindrops, changes in brightness and background in outdoor environments, are being filtered out in the detection and tracking processes. One example is the tracking and sending an immediate alarm when human-like objects are detected walking among moving vehicles in a busy multi-lane motorway.
Apart from the core engine, the Licence Plate Recognition (LPR) coupled with high speed deblurring solution is designed to detect and identify fast moving vehicles entering to a designated area within the camera view, capturing vehicle's license plate number for specific traffic enforcement/control use. Images of vehicles are stored using customized image compression functions. High speed deblurring, image enhancement and checksum functions were developed to enhance and boost recognition rate.
Video Analytics key features include:
LPR key features include:
The solution enables on duty round the clock surveillance without any day off or showing signs of fatigue, distraction or loss of concentration, helping to avert an unwanted incident.
The technology can be potentially implemented in many different situations:
The LPR potential applications in:
Our technology has dual layers of operations to give it the unsurpassed level of accuracy and reliability:
(i) Detection Layer (Core Engine) and
(ii) Application Layer (Filtering Engine and Plug-ins).
Its Core Engine detects objects based on parameter sets while moving and stationary objects are tracked and distinguished by different markers. Conventional solutions also lack the Application Layer which has the ability to filter unwanted blobs due to fast moving camera, shapes classification, color & position, timing & distance, path tracing, vehicle headlights etc. Additional ‘plugins’ are available for specific applications.
Real time alerts are also available in the forms of SMS, MMS and the central monitoring system (CMS). The CMS is capable of monitoring 32 cameras and reporting positive events with queue buffers. The technology also controls Pelco-D, IP and Sony PanTiltZoom (PTZ) cameras to capture close-up view of the intruder(s).