Three Brand New Features to Deep Learning Video Analytics

Videcon are delighted to announce a sweep of updates to Deep Learning Video Analytics on Professional recording units, improving the accuracy of our two popular methods of detection and incorporating three new innovative features designed to further reduce false alarms.

Double Knock Alarm Triggers

A double knock can be configured for DLVA in conjunction with an alarm triggering devices such as a PIR sensor wired into the recorder. The Double knock feature adds an additional layer of false alarm filtering; when activated an event notification will only be triggered if both the DLVA and the PIR sensor trigger an event simultaneously.

Confidence Threshold

The Concept Pro Deep Learning Video Analytics technology can detect human beings and vehicles and provides a confidence percentage for each. For example, a person who is far away in the image with their back turned and crouching down may return a percentage of 50% whereas a person facing the camera close to the image, may return a percentage of 100%. To further reduce the number of false alarms, the system can be configured to ignore any detections below a certain confidence threshold. For example if a 60% confidence threshold is set on human deteciton then an alarm will only be activated when the system detects a human with a confidence of 60% or greater. The confidence percentage will also be displayed in the event notifications and when searching through past events.

Static Object Filter

The static object filter in the DLVA set up menu tells the system to ignore detected objects that are not moving. For example, a vehicle drives into a restricted space and stops, an alarm is triggered once, rather than repeatedly for the entire time the vehicle stays in the area, this reduces the unnecessary number of trigger events logged and makes reviewing of the alarm points much easier. The recommended setting for static object filter is 'high' however this can be configured to 'medium' or 'low' to suit the installation environment.

 

Frequently Asked Questions

What is the Detection Range for DLVA Detection?

Detection Type Camera Lens Size Distance (Approx)
Human Detection 2.8mm 10 Meters
Human Detection 10mm 20 Meters
Vehicle Detection 2.8mm 25 Meters
Vehicle Detection 10mmm 60 Meters

 

How do I get the latest features if I already have a Concept Pro Professional Recording Device?

To update your recording device to the latest firmware, please contact Technical Support on +44(0)1924) 528006


Activating a New Licence

Deep Learning Video Analytics requires a license to be activated on the recorder. The license can be purchased at the point of ordering a new Concept Pro recorder and will ensure the recorder comes with the Deep Learning Video Analytics license activated. Alternatively, the license can be activated after the recorder has been purchased, simply call Videcon on +44(0)1924 528000 to order the license; you will need to provide the MAC Address and the first 5 digits of the firmware version of the recorder. To add a license post-sale, the recorder must have been purchased after the launch of Deep Learning Video Analytics. 

Supporting Documentation

Deep Learning Video Analytics User Guide - Download 

 How to set up Deep Learning Video Analytics - Watch Now 

 

 

Videcon Ltd, Unit 1 Concept Bus. Park, Smithies Lane, Heckmondwike, West Yorkshire, WF16 0PN
Sales : +44(0)1924 528000 | Technical: +44(0)1924 528004 | Customer Service: +44(0)1924 528006

Videcon South, Unit 5, Bridge Gate Centre, Martinfield, Welwyn, Garden City, AL7 1JG
Office: (01707) 932 555 | Sales : +44(0)1924 528000 | Customer Service: +44(0)1924 528006