Automatic detection of abnormal situations and behavior in parking lots
Parking management system
950 million € (2016)
Number of employees
- Reinforce user safety in car parks
- Create smart video surveillance camera networks
INDIGO is a French Group, world leader in parking and individual mobility, employing 20,000 people in 10 countries and more than 750 cities. It is the only operator in the sector to be present on three continents and on all types of parking: roadways, shared private parks. INDIGO operates more than 5600 car parks worldwide, with more than 2.3 million parking spaces and more than 3000 km of roads. INDIGO supports local authorities and private actors (hospitals, shopping centres, stations, airports, universities, etc.) and develops innovative mobility solutions to facilitate users’ mobility in the city.
In June 2016, INDIGO launched OPnGO, the only application that offers a new parking experience, including route searching and then the best place at the best price. With OPnGO, the customer journey evolves towards more freedom, more choice and more personalization. Finally, in October 2017, INDIGO took a new step and launched INDIGO Weel, becoming the only French player to offer self-service bicycles without a docking station, intended for the countries in which the Group operates. INDIGO is a global player with a local presence. In a booming mobility market estimated at €100 billion in 2025, INDIGO is a key player in the mobility of the future. INDIGO. Time for the future. www.group-indigo.com
Car park operators have the task of managing both increasing security pressure and the gradual redefinition of the mobility market. In a context marked by a feeling of insecurity among citizens, car parks are engaged in creating a relationship of trust with their users and an atmosphere of security. On the other hand, as actors of mobility, and in a globalized environment where connected, electric and soon autonomous transport is developing, car parks must be able to integrate into this smart environment and know how to interact with the modes of transport of tomorrow.
Today, the installation of video surveillance cameras in sites such as car parks is common practice, but it is impossible for remote video surveillance operators to look at all screens at once to identify an abnormal situation. Indigo has 650 car parks in France, each with 20 to 30 cameras. To ensure the safety of people, parking operators must rely on additional technical resources to manage exceptional situations. When an abnormal behavior or situation is detected and an alarm is triggered, an Indigo employee present on site must remove any doubts. To be effective, it is therefore essential that the alarm accuracy rate be close to 100%, with as few false positives as possible, and operators must be able to have the information in real time.
Indigo therefore wants to equip its fleet of video surveillance cameras with AI so that they are able to analyze and understand camera images. For this purpose, Deepomatic has developed a video recognition system capable of detecting:
- Fraud and threats: violent behavior, masked people, possible presence of weapons and attempted thefts or break-ins
- A camera occlusion
- Pedestrians and vehicles
- Emergency situations: people waving their arms, people lying down
Due to the large flow of video data to process, the algorithms developed were deployed directly on site, within Indigo’s infrastructure, in order to control the costs associated with video processing.
How it works in production:
When a video surveillance camera detects an abnormal situation, an alarm goes up in priority and in real time to the “Connect Park” control room where an operator can decide to clear the doubt himself via the video stream, or alert an Indigo employee on site. Depending on the risks involved, the necessary measures can be taken instantly, such as sending a priority support team to the operating site.
- Improved levels of monitoring and assistance
- Strengthening the reliability of results with more efficient and accurate solutions
- Increased productivity with faster and more productive action possibilities
- Improved user experience and sense of security in parking lots
We are working with Deepomatic to develop a decision-making tool. We are building a customized visual recognition system that helps us detect a danger very quickly and minimize it. The Deepomatic solution adapts to our business and integrates itself into the existing video surveillance system. The purpose of our cameras is not only to monitor, but also to keep us informed in order to detect the right information at the right time for us to act in the most effective way possible.