"It is time to move from the question What is happening? to the question Why is it happening?"
When the need arises to move from documenting phenomena to analyzing their causes, this is the time to engage advanced analysis functions.
The difference between the classical approach and the use of advanced analytical methods is that in the classical approach, the system is set up to track a particular phenomenon. It processes this, analyses it according to the given methodology and gives the result.
Using these methods, it is possible to solve tasks that are difficult to solve with classical methods or only with extremely high costs.
Examples:
- It is common that a camera can categorize objects very different in appearance - a truck from a pedestrian. A passenger vehicle and a cyclist. But it is extremely difficult to distinguish objects that are similar, e.g. a pedestrian from a person on roller skates, skateboard or scooter. A cyclist from a motorcyclist or a person on a scooter. In this case, it is more efficient to teach the AI to distinguish between these objects than to develop a specific application that can distinguish between them.
- The camera monitors the speed of the vehicles and detects an infraction if they are speeding or if there is a small distance between vehicles. Advanced analysis tools can distinguish what behaviour is potentially dangerous and react to it. It looks for reasons why dangerous situations occur somewhere and not others, even if conditions are similar.
The difference between the classical approach and the use of advanced analytical tools is therefore mainly that in the first case the system follows a set algorithm. In the latter, it collects data, learns from it and looks for connections itself.