Sitting behind the video wall of the Pisciac urban supervision center (CSU), they see everything. Since the creation of the equipment by the City in 2015, there are seven video-operators to carefully scrutinize, 365 days a year, the images captured by the hundred cameras installed throughout the city. A tedious job that requires constant vigilance since in the event of an accident, a traffic offense or a person, these little hands of the CSU are sometimes the first witnesses and often the first link in the municipal security chain. To help them and improve the efficiency of its video surveillance system, the City has been using artificial intelligence for the past year to immediately identify certain offenses or suspicious behavior on its streets.
In Poissy, the idea of integrating artificial intelligence into the video surveillance system was born in 2021 when the former mayor, Karl Olive (Together), was looking for a way to fight more against heavy goods traffic. in transit in the municipality. Although their circulation was prohibited in the city center, offenders still too often took liberties to the taste of the municipality. Because even if the offending trucks can be filmed by pisciacaises video surveillance cameras, the operator still has to have his eye on the right camera at the right time to then warn the municipal police officers on the ground who will intercept the driver.
This is how the City turned to the XXII company (pronounced twenty-two) which proposes to integrate into the video stream of the CSU operators, a decision-making aid tool operating with alert feedback. Very concretely, in the case of truck traffic, if a heavy goods vehicle passes in front of a camera connected to one of the artificial intelligence (AI) licenses purchased by the City, the operator will instantly receive a notification on his screen, even if the latter was busy monitoring other sectors of the commune.
“The idea is to help operators make decisions: the algorithms will shift part of the human processing that is complicated to do on many cameras to tell the operator to pay attention to the event in progress”, explains Souheil Hanoune, director of research and development of the XXII group.
“It allows us to have a little greater responsiveness in the field”, explains the interest of such technology the brigadier chief, Patrick Ployart, who is responsible for the urban supervision center of Poissy. The AI is not on all our cameras, because it has a cost, it is distributed according to our needs. Today we have ten licenses that we can use for different skills such as parking, heavy-vehicle traffic or, for example, the prolonged presence of certain people on the public highway”.
Of the hundred cameras available to the city, 78 perched on public roads and around thirty in municipal buildings, only ten can therefore use this alert mechanism simultaneously. “The advantage is that you can switch licenses from one camera to another because they are not associated with a specific camera,” emphasizes Souheil Hanoune.
Likewise, each camera that uses artificial intelligence is not limited to a single skill. A camera installed on an urban boulevard can, at the same time, report information on heavy vehicle traffic, detect the presence of vehicles parked in prohibited locations and count the number of pedestrians circulating in the area. area.
This last aspect was used in particular by the CSU during the health crisis, when traffic on the public highway was very restricted by periods of confinement or when the rules of social distancing were still strict. “For example, we could know at the moment T if the public gauge had been exceeded in the town hall”, explains Patrick Ployart. Person detection can also prove to be particularly useful in the event of an intrusion into a public building. “When we want to do a replay, we will have a bookmark directly on the time tape, which is a very significant time saver since it allows us not to have to review the entire tape,” adds the chief brigadier.
“Security remains a priority, underlines the mayor of Poissy, Sandrine Dos Santos (LR). Our majority initiated upon their arrival at the head of the community in 2014 the deployment of video protection cameras in all neighborhoods, with the support of the Pisciacais. It is also with the support of the inhabitants that we have set up this innovative urban supervision center, which operates 7 days a week, 24 hours a day. This system has, through its effectiveness, reduced crime in Poissy . Today, the implementation of XXII’s AI contributes to strengthening and optimizing our security strategy, allowing greater responsiveness when an abnormal situation arises. On the strength of this experience, it is natural that we will be able to move towards a smart city strategy”, specifies the city councilor.
Inspired by the needs of municipalities to create new skills for their artificial intelligence, the XXII company is now working on an algorithm to detect illegal deposits, a real scourge in the Seine Valley. If the City of Poissy already uses a first version of the tool, which only detects van stops in “sensitive” areas, the head of the CSU is impatiently awaiting a more extensive update. In his office, Souheil Hanoune does not fail to reassure him: “We are working on a new version to do anomaly detection, that is to say we are going to teach the algorithm what is normality in a street, there are never green or black things blocking or cluttering the sidewalk, and therefore issuing an alert on something new that remains in time, it can also be someone who has fainted on the road! »
We would love to thank the author of this post for this outstanding material
Doped with artificial intelligence, nothing escapes video surveillance cameras
Check out our social media accounts as well as other pages related to it.https://www.ai-magazine.com/related-pages/