New technological solutions for Airports - The continuous development path of algorithm

NGÀY ĐĂNG 18/09/2025

Admin

Smart video surveillance today is not simply the recording of pre-set images surveillance cameras, but also the technology of recognizing and analyzing people and objects regardless of multiple groups of people moving continuously within its field of view. Built-in intelligent algorithms are the basis for the process. This modern technology application works based on the connection with other digital devices at the airport.

Bosch will integrate intelligent algorithms into airport security systems to speed up the processing passenger information

One ​​person with a backpack. One person without a backpack. One person wearing a red jacket. One person wearing a blue jacket. When Robert Borchers' team trains algorithms, it's not like the high-concept science fiction we usually think of when we think of algorithms, but It's more like explaining coloring books to a toddler. "The task we give the algorithm is to recognize regularity and repeating patterns," explains the Head of Connected Mobility and Computer Vision at Bosch. Ultimately, the software includes algorithms that have learned enough to be able to make independent judgments even if the object it recognizes is just a person. It can not only distinguish between individuals, but also between luggage and mobile devices. Although training these algorithms is time-consuming, Bochers believes it can be managed. People at the airport are more efficient thanks to algorithms.

In the future, cameras will not only observe normal images but also be able to analyze data intelligently

Boscher has "Previously, surveillance cameras were mainly used as evidence in court in emergency cases." However, the actual processing speed and analysis of conventional cameras still face many difficulties. Although the human eye can see sudden changes very well in a static environment, it has difficulty difficulty focusing on constantly moving images, such as the movement of people at airports. “When p"When watching airport video, most people can't focus on identifying details of multiple objects at the same time," Borchers explains. Meanwhile, there is software that includes algorithms that have learned to distinguish individual people in the image so that they can be recognized in the surveillance area. how many people in just a few seconds.

Analyze howí can be achieved by using multiple camera tracking

Using algorithms to count the number of passengers at an airport on a real-time basis is quite interesting. However, these algorithms also need to be combined with other complex supporting equipment at the airport to optimize its application. For example For example, with the help of this so-called multi-camera tracking feature, all the surveillance cameras of a terminal or airport can be connected to several hundred other devices at a small airport, or over 1,000 devices at international airports. “Such a hardware-based algorithm can precisely specify the number of identifiable passengers and provide each customer with any service, “according to their needs”, Borchers describes the areas in which it can be applied. Therefore, a myriad of combined calculations can arise from passenger behavior: What percentage of passengers entered the airport at 10:00 and went through security at 10:30, what was the average speed of passengers?. For a newly arrived passenger, what is the shortest waiting time for baggage handling?   “Data on all passenger movements is collected at the airport and allows predictions on a per-passenger basis,” Borchers emphasizes. “Using intelligence will speed up processes program.”

Suse software that includes algorithms and Linking airport security devices together to create technology that understands customer behavior in the near future

Customize notifications for each customer

In these ways, experts think that this modern technology can surpass the standard of conventional video analysis. The key point lies in the smartphone. At the moment when a passenger enters the airport and checks in at the system, they will scan their virtual plane ticket through their smartphone. At this point, the intelligent algorithm quickly knows which gate the passenger will be flying at, so it can accurately calculate the time the passenger must check in at that gate. At the same time, intelligent video analytics also accurately knowlocation of all passengers on the same flight. “What’s interesting is that at the exact moment of security check, the system will send each passenger a different message via their smartphone,” Borchers said. In addition, the system will recognize if the baggage check areas are more crowded than usual. In such situations, passengers of those flights will be informed, and possibly, they will be offered a cup of coffee. phˆ.

The future starts today

Borchers asserts: “We are taking on three challenges so that the surveillance system can do what I have mentioned”. The software includes algorithms that must continually learn to interpret human behavior. Borchers finds it advantageous that Bosch is not an airport application developer in the first place, but they can benefit from algorithms that have already been developed, for example, by the automotive industry. These algorithms already know what a human is, and they just have to learn more about the characteristics of airport passengers. However, they are only specialized calculations, at which point Borchers mentioned the second challenge, “The more data is analyzed at once, the more powerful the computing power of the computer must be.” Not all processors assessments are all installed at the airport, Borchers admits, “Today, the computing power of computers is transferred to the cloud, but in our experience we see that the more data that needs to be analyzed, the higher the analysis capability.” Therefore, modern camera systems often require a powerful processor to directly perform initial assessments. In addition, features such as high resolution, day and night monitoring, and recording solutions can affect the quality of individual devices operating in the system and main control algorithm.

Smart video analytics step by step

Airport managers who want to implement such systems need to start planning ahead. “We are making significant progress in deep learning, because our algorithms are constantly learning, and at the same time the hardware is getting exponentially faster,” says Borchers. The first firmware analysis programs could be available in 2018 or 2019. However, according to experts, airport solutions are still in the early stages. comprehensive functionality and allows trackingWith multiple cameras and other related modern technologies, it may take another 10 years before it can be fully popularized in life. “Such smart surveillance systems need to be implemented step by step,” Brochers admits. “First, the airport manager can test the feature at one gate, or he can introduce multi-camera surveillance throughout the airport, and A few years later, he added a fisherman's information system. access smart phone.”

Source: boschsecurity.com