HANWHA VISION ANNOUNCES WISEDETECTOR, ENABLING USERS TO ‘TRAIN’ DEVICES TO DETECT NEW OBJECT TYPES

Hanwha Vision has launched WiseDetector, a machine-learning-based feature that can train video cameras to detect new object types according to a model defined by the user. This gives users greater flexibility in adapting object detection to their needs. WiseDetector is supported in P Series AI cameras and is licence-free.

The tool works effectively with detecting distinct objects. Some examples of objects it can be trained to detect are boxes posing a hazard on a shop floor, cones obstructing roads, or signs blocking corridors.

The flexibility of machine-learning-based detection also creates opportunities for safety and security teams to reinforce other measures, such as access control. For example, WiseDetector can learn to recognise open or closed doors and provide a visual alert that improves access control.

Training the WiseDetector model is achieved through the Device Manager and users do not need deep technical knowledge to get started. This makes it relatively easy to set up and further expands the possibilities of video analytics. Users simply collect images of a target object and the machine-learning model learns to recognise it. When the trained WiseDetector model is applied, it can then use existing video analytics applications on the camera to trigger live events.