Motion and intrusion detection or similar have been a standard feature for some time now on even some of the most basic cameras and recorders, however for the purpose of reliable alerts to an intruder these functions have been woefully unreliable due to this all being down to basic pixel analysis (how many change within the set area) – so it is easily fooled.
Cats, blowing rubbish/leaves, rain or simply sudden cloud cover changes are often enough to set off repeated alerts – so this unfortunately is not an ideal solution unless in a controlled environment, while decent quality infra-red detectors have been the go to choice for reliable alerts otherwise (or even radar now or thermal) these do add to the complexity of a system and cost and each have their own limitations.
However, we now have a range of ‘deep learning’ technologies available on certain new Cameras and recorders that are able to decide whether an object is a human or vehicle with a reasonably high accuracy - Hikvision call theirs ‘Deep in Mind’ or ‘Acusense’ depending on the range and level of technology.
The Acusense series of DVRs, NVRs and IP Cameras are now able to filter out up to 80% of false alarms with the latest generation. Cameras also feature the ‘Powered by DarkFighter’ imaging to provide-decent low light picture quality.
The latest and greatest ‘deep learning’ DeepinMind (NVR) and DeepinView (Camera) series offer a superb 90% reduction in false alarms with it either being done back at the NVR (with high quality cameras providing the images) or within the camera itself which can be more cost effective for smaller numbers of cameras, again classifying human and vehicle targets. For best results the NVRs should be paired with the Ultra 5 series cameras (Darkfighter) – which is what a DeepinView camera is based upon.
Using this technology, a system can be deployed to provide full Facial Recognition or Number Plate Recognition so an individual or vehicle can actually be identified and an alert sent to an operator or a barrier/gate opened – for example.