Czech scientists have presented a new system that works in drones – it taught devices to detect abnormal behavior. The method will help the police respond more quickly to emergencies.
The researchers noted that law enforcement agencies in many countries are using technology to investigate crimes or improve response times. Most of them use drones to monitor large groups of people or large areas with a limited number of people.
While this technology is very useful, it is limited in one aspect: the ability to decide what is normal and what is not. They can only transfer the footage to the researcher, who decides what action to take. Therefore, a group of Czech scientists decided to endow the devices with the ability to detect suspicious behavior. To do this, they have developed a surveillance system that analyzes material captured by drones.
Developed as part of a joint research program between the Brno University of Technology and the Czech Police, the new system uses neural networks to decode data. In addition to observing and controlling crowds, scientists say it can be used to control traffic.
The material captured by the drone is divided into smaller cells. Then the system analyzes and creates an overall picture of what is happening. After that, she develops a model of behavior in a standard environment, and then compares the anomalies – if any, the device informs the researcher about it.
The big advantage of this system is that it learns and executes the program in real time so that the police don’t have to waste time making decisions.
During the tests, scientists asked the system to observe the players on the field. Then some of them were ordered to lie on the ground. The system immediately alerted the observer to the anomaly, the report said.