On the streets, at airports, at state offices, at schools, inside the vehicles that are part of a business fleet, etc… all you have to do is look up to see the cameras.
With time, images turn into allies for monitoring all kinds of activities. However, this strategy brings a difficulty. Although the figures differ from source to source, it is estimated that cameras associated with closed circuit television generate 24,000 million hours of video per day.
How to obtain valuable information from such a large amount of data? What is the use of accumulating hours of video if no one can watch the material later? Is it possible to have as many people as there are cameras? New technologies bring us a concept to solve all these questions: “video analytics”.
Video analytics allows the enormous number of image capture devices distributed throughout the world to be valued, since it uses machine learning and artificial intelligence to train cameras that can detect unusual events or behaviors, for example, alert about a potential security problem.
The main benefit is that in this way the historical paradigm in terms of surveillance is broken: there is no longer a reaction to an event that has occurred -and which, as a consequence, could already have generated costs, risks or reputational problems-, but rather it anticipates to the facts before they materialize. On the other hand, you don’t need to check monitors continuously: video analytics highlights only unusual events, which are a minority in the huge volume of recording hours.
As one of the characteristics of artificial intelligence is continuous learning, the results will be better as time goes by.
Analysis of objects and people
Video analytics is used for different functions. For example, one of the most common uses is facial recognition, an application that is used for security control in areas where crowds converge (such as immigration barriers at airports) and to authorize access in companies, among others.
But in addition, the algorithm may be able to detect a particular object and track its trajectory (it can even warn security personnel if it crosses a certain line), or follow a person through multiple cameras.
With the pandemic, an application that quickly gained popularity emerged: counting people in an environment to ensure that the permitted capacity is not exceeded. In many cases, video analytics was even combined with temperature sensors to verify that none of those attending a place had a fever.
While security is the first big application segment, it’s far from the only one. The global video analytics market could grow from $1.6 billion in 2021 to $7.5 billion in 2027, according to AllTheResearch.com. This is explained by the notable increase in video analytics use cases in the business world.
In manufacturing it increases its quality control in the same production line, quickly detecting anomalies or deficiencies in the products. In logistics, meanwhile, it can be applied to inventory management and optimization of storage spaces.
Retailers can use it to track customer behavior, to redistribute space, improve the consumer experience and increase sales.
Video images contain a huge volume of data. Video analytics is the key to converting this into timely information for decision making, improving security levels and even increasing profitability.