Hanwha Techwin America, a global supplier of IP and analogue video surveillance solutions, announced a line of 4K AI cameras. Part of the Wisenet P series, the new AI-based cameras capture pristine images at up to 4K resolution while including powerful, in-camera deep learning algorithms for advanced object detection, classification and error-free analytics. Utilising object recognition versus motion detection all but eliminates false alarms while also providing valuable business and operations insight.
“Our new AI cameras have solid performance in both analytics and deep learning applications,” said Ray Cooke, Vice President – Products, Solutions, and Integration, Hanwha Techwin America. “The included, licence-free analytics detect and classify a range of objects including people, vehicles, licence plates, and faces. This technology will provide more reliable edge based intelligence, and open new opportunities in security as well as business and operations intelligence.”
Unique attributes of the objects are also stored as metadata alongside the video information including Colours of people’s clothes for both top and bottom; People’s clothing length for both top and bottom; People wearing of glasses or not; People wearing or carrying a bag; Age group; Gender; Vehicle types and Vehicle colours.
This metadata can be read by a VMS server and used in post-event forensic search to significantly reduce time spent investigating specific events. Wisenet P series AI cameras offer integration with Wisenet WAVE and other popular VMSs from Milestone and Genetec using Hanwha’s plugin.
As businesses reopen with new policies regarding social distancing and occupancy limits, it’s never been more important to accurately count customers and employees as they enter and exit the premises. Unlike traditional, imprecise methods of people counting with dedicated overhead cameras, the Wisenet P series AI cameras deliver accuracy with people counting based on the camera’s AI object detection algorithm.
Wisenet P series AI cameras are equipped with a ‘BestShot’ feature which ensures that only the most suitable image of classified objects is sent to a backend server. The full video image can be up to 4K resolution to see details of objects, along with tagged attributes metadata. Users have the option to set the image resolution to less than 4K for full video streaming or use BestShot to minimise bandwidth and storage requirements for server-based analytics.
