I know what you’re doing ‘now’
It’s past midnight on the 18th of this month, in an alleyway in Osan, Gyeonggi-do. A man walks alone down an unpopulated alleyway. He stumbles, not sure if he’s walking correctly. As soon as a person appears where there was no movement, the CCTV system detects movement – a ‘situation’.
The CCTV system pops up a screen that says to the controller, “Something’s not right here,” and the controller decides to keep an eye on the man. The next CCTV captures the parking lot, and without hesitation, the stumbling man starts his car and drives off.
Tipped off, the controller alerts the police. The license plate number is already known from the CCTV, and the dispatcher shares with the police where the car is going. The car drives for about 4 kilometers and is caught by a police car following it. The police checked the driver, who is in his 30s, and found that he had a blood alcohol level of 0.077%, which is what the dispatcher expected.
■Gender, smoke, and collapse detection…Wheelchair and baby car recognition
The place where these CCTVs are gathered is the Osan Smart City Integrated Operation Center. More than 2,400 CCTVs installed in various parts of Osan City are controlled around the clock by 50 people in shifts at the integrated center.
The basic analysis data for the control system is human movement. It can identify the gender of the person moving, and whether they are wearing long or short pants. If a person suddenly lies down and remains still for more than five seconds, the monitor automatically displays the words “situation collapse”. The controller can decide whether to call the fire department for help by watching the automatic screen.
In addition, objects such as wheelchairs, white canes, and baby cars, as well as situations such as fire smoke, are also recognized by the CCTV and displayed on the monitor. It can be used to locate lost elderly people, missing children, and more, and can even recognize and relay fires to the fire department faster than citizen reports.
■ Detecting ‘crowding’ and expanding safety applications…thanks to technological advances such as deep learning
The use of CCTV in safety areas is growing and diversifying. For example, a virtual safety perimeter can be drawn on a CCTV capturing a reservoir, and if a person is caught crossing it, a situation can be flagged. You can also make announcements from CCTV in the field.
You can also use this feature in areas that are normally crowded, such as festivals. You can calculate how many people are staying per square meter by calculating the area of the set zone. Currently, in Osan, the monitor displays 4 people per square meter as caution, 5 as warning, and 6 as severe. It can be customized by changing the settings depending on the slope or characteristics of the road.
These technologies are possible because we are adding detection objects and increasing the recognition rate through big data and deep learning.
■ Only 24% nationwide penetration rate of intelligent CCTV…’privacy violation’ concerns
If conventional CCTV is simply viewing or recording, CCTV that detects and tracks various movements and is connected to the control center as described above is called ‘intelligent’ CCTV. It’s a way to effectively control with fewer people. However, there are bound to be blind spots, and in the case of Osan, the city said that it would take 4,000 cameras to eliminate them.
In response to this concern, Kim Young-hyuk, head of the Smart Traffic Safety Division in Osan City, said, “The video recorded by CCTV is destroyed after 30 days, and when CCTV is installed, it is announced and operated for 20 days, including the purpose of use.” He added that employees are prohibited from carrying mobile phones when entering the control center.
The Ministry of the Interior and Safety announced the ‘Comprehensive Disaster Safety Plan’ after the Itaewon tragedy, and the full introduction of intelligent CCTV was a key part of the plan.
Nationwide, the penetration rate of intelligent CCTV connected to local government integrated control centers is currently only 24%안전놀이터.
■ “A key tool in the social safety net”… Safety is not ‘willpower’ alone
The Ministry of the Interior and Safety has stated that it aims to deploy intelligent CCTV control nationwide by 2027, and initially made a budget estimate of KRW 700-800 billion. In response, Heo Seung-beom, head of the Safety Improvement Division at the Ministry of the Interior and Safety, said, “If you change all the cameras to AI (artificial intelligence), it will cost KRW 700-800 billion, and it can also be done by improving the software of existing cameras,” adding, “The budget estimate should be redone.”
Kim Sung-ho, head of the Disaster Safety Management Headquarters at the Ministry of the Interior and Safety, said, “We will push for nationwide deployment of intelligent CCTV as a key tool in the social safety net, as well as overhauling laws and strengthening the capacity of control personnel.”
In the case of Osan City, which has been building intelligent CCTV for 10 years, he said that the budget has always been insufficient, and he requested active financial support from the central government. This is an expression of his experience that safety is not only a matter of will, but also a matter of budget, which means that real safety can only be achieved if the budget is supported along with the will.