False Alarms – False Alarms – and more false alarms!
We’ve all had the frustration – waking up abruptly- in the middle of night, at the sound of a screeching siren. Hurriedly returning home from the school function at the call of the armed response company – only to find , you guessed it – a false alarm.
The Problem Of False Alarms
Follow this article to the end, and let’s put a stop to this – I’ll show you how at the end of the article.
The problem is so enduring, that endless different types and designs of alarm equipment is being designed. CCTV, Different Alarm Panels, New Types of Beams, Artificial Intelligence, Electric Fences – all in the quest for less false alarms, and accurate alerts.
Whats the use of an alarm or security system, if it wakes you up in error 9 out of 10 times?
The Quest For Accurate Intruder Detection
As a security provider, it’s really frustrating and embarrassing to deploy the next best thing and find out that it too, is subject to more false alarms than actual alerts.
That’s not providing our customers with security – rather just increasing their anxiety.
The problem is mostly present with outdoor detection – rain, wind, sun and shadows play havoc with all these systems.
I have really pursued the objective of accurate detection. I thought I had the answer when major manufacturers like Hikvision and Dahua started talking about their AI algorithms – that could distinguish between humans and vehicles, and ignore leaves and shadows.
Hours of remotely trying to fine tune customers systems, and more embarrassment ensued. Rain and tree shadows where detected as human, birds hovering in front of the cameras, driveway lights reflecting off the wet driveway detected as vehicles. An endless stream of push notifications every few seconds – right back where we started. So much for artificial intelligence.
Getting Closer To A Solution
I did however find something useful. Ironically it was a technology – a video analytic, that I had started using way back in 2012 with success on estate perimeters. It wasn’t video motion detection, Human or vehicle detection – but the simple trip-wire analytic.
At last – I could conveniently place my detection in areas away from trees and shadows, along perimeters, and it wouldn’t be activated by rain, wind and lights. (Mostly that is)
As promised, here is how we will greatly reduce false alarms, and have the ability to instantly verify an alarm when one does occur.
(1) We are going to use the AI Camera technologies indoors only, This is a good environment for human detection and it’s generally pretty accurate indoors.
(2) We will deploy outdoor video tripwires only. No motion detection, No Human Detection and no Vehicle Detection. If these technologies improve substantially, we should reconsider.
For now, with the current state of technology, this is about as close as I can get in the quest for reliable security detection that should allow my customers to sleep easy, feeling secure and protected.
gensixtech.co.za
