Surveillance cameras are installed in many places such as airports, parking lots, train stations and banks. In the past, the video imagery has been mainly used as a forensic tool after an event. To take advantage of the video in real time
A human must monitor the system continuously in order to alert security officers if there is an emergency.
One person can only observe approximately four cameras at a time with good accuracy of event detection.
Humans are limited in terms of attention span and cognitive processing. We could discuss reduced attention and processing capability – per video feed, as the cognitive load increases, visual attention span and focus is divided amongst multiple scenes, the brain simply receiving more information than it’s capable of processing, overloading the short term memory capacity, and decision fatigue. It’s enough though to say that humans are not a reliable means of monitoring live video feeds.
Humans too, are an expensive resource, requiring one operator for a maximum of four video channels.
I once wrote an article on the evolution of video surveillance, from it’s inception in the german war bunkers – to the point where humans could no longer effectively monitor the large number of cameras in play today. In the evolution of video surveillance, AI and machine learning make the biggest impact in this area. It provides an accurate live monitoring capability, of multiple video feeds – if it’s overwhelmed, you add compute power and memory.
While many sales people write about AI and video surveillance, and the many capabilities it possesses – it is here, in the field of live monitoring, of a large number of cameras, that AI video analysis was born.
These then are the real benefits of AI video analytics.
Enhanced Detection and Analysis
AI algorithms can maintain focus on multiple scenes, detect movement, recognize faces, identify objects and analyze patterns of behavior without tiring. They can achieve this in real time. As the systems detects events, it throws out alerts to a much reduced number of human operators. It doesn’t stop monitoring – as a human would, when engaged with a detection. The human operators can now investigate the detected event with all their attention.
Improved Accuracy
Alongside eliminating the issue of fatigue, the number of fatigue induced errors is reduced or eliminated. The accuracy of the system is maintained 24/7.
Machine Learning models (I have posted another article somewhere, detailing the difference between AI and ML) can learn from new data. It uses this data to improve it’s detection capabilities over time, and adapts to new threats or unusual activities.
Scalability
Unlike the human, whose mental capacity remains fixed – you can upgrade! As the video complexity grows, the business adds compute power and memory. AI is extremely flexible, It can be trained to detect the clues and features presented to identify theft, unauthorized access, or tracking the flow of people or goods – without missing one of the features or clues.
Cost Efficiency
As already discussed, significantly fewer, expensive human resources are needed to monitor a large number of cameras – reducing costs. The AI doesn’t need breaks to ensure continuous surveillance. There is no need for double the number of operators, and multiple shifts.
Case Studies and Examples
The use cases are many but some businesses that benefit significantly from AI are:
Retail: Detecting shoplifting, unauthorized access, analyzing traffic patterns can be achieved effectively, with significantly reduced number of security personnel.
Transportation Hubs: Managing security in crowded areas, and ensuring the smooth flow of passengers.
Public Safety: Monitoring public spaces, detecting criminal activity and monitoring vehicle traffic.
In summary, AI is a distinct evolutionary marker for video surveillance. Addressing the critical limitations of human monitoring, offering cheaper scalability, higher efficiency and accuracy. This transformation brings to businesses a high level of accuracy and efficiency, at reduced cost.
Gensix Technology takes this a step further. Our machine learning extensions continuously improve the performance of AI as it learns from non obvious cues.
Speak to a gensix representative to explore the potential of reduced costs of live monitoring – service@gensixtech.co.za
