Benchmark Innovation Awards 2018 Video Surveillance (Hardware) Winner: Hikvision – DeepinView Cameras
Hikvision’s DeepinView IP Camera Series incorporates the use of deep learning algorithms to enhance the intelligence of edge devices, thus increasing the value of surveillance system performance across a broad range of security and management applications.
The deep learning algorithms support multi-level programming which, when compared with conventional intelligent algorithms, enables faster and more in-depth processing. The algorithms are used to implement feature-learning, providing accurate and consistent video content analytics (VCA) performance. The algorithms make use of high-speed GPU processors which allow computational analysis of large amount of complex data, such as streaming video.
The advanced processing allows advanced video analytics to be implemented at the edge in video surveillance systems, adding flexibility to solutions as the cameras can use multiple sensors. This approach can, for example, allow an overall view to be captured with any activity of interest being detected in the scene. A second PTZ sensor can then focus on the activity, delivering a higher resolution stream containing the important video footage.
This enhanced versatility can be deployed in a wide range of applications, providing tracking, alarm activation, pre- and post-incident alerts, as well as improving the gathering of business intelligence or automating site management systems.
DeepinView cameras can detect people while filtering out other objects and movements within a scene. This is of particularly interest in perimeter protection applications.
Facial recognition can be easily added to a variety of applications, enabling alerts and notifications to be sent to system operators whenever persons of interest are detected. The facial modelling and processing is carried out at the smart camera.
Return on investment can be enhanced as the DeepinView cameras offer not only security, but also business intelligence. For example, the collection and analysis of accurate visitor information can assist businesses with regard to increasing profitability, ensuring compliance to health and safety policies and automating site management tasks.
People Counting analytics can log the number of people who enter or exit a defined area. The cameras can monitor foot traffic through a defined space, or count the number of people inside a venue at any given time to ensure occupancy limits are adhered to.
Traffic monitoring and automatic number plate recognition (ANPR) can be deployed to monitor vehicular traffic movements.
The DeepinView range has been designed to upgrade standard video systems to intelligent automated detection systems, delivering efficient surveillance management.
DeepinView cameras can be supplemented by DeepinMind NVRs and analytics servers to allow the creation of flexible solutions.