IMAGE PROCESSING AND OBJECT ANALYSIS
Data in cities
A special type of city data is data collected by individual devices or entire vision systems, collecting and transmitting video information, which, combined with data on their location in the city space, constitute a comprehensive collection of city data. Proper analysis of the data collected in this way is able to provide information on road traffic, the effectiveness of urban transport, dangerous behavior, vandalism, parking zones, and also be a source of aggregated statistical data.
We have also found out that the proper use of such collected data can become an additional source of income for the city budget.
One of the elements that are part of the smart urban ecosystem is a solution for the acquisition and processing of information collected with the use of mobile multisensors. The system is also dedicated to real-time delivery of measurement results to the City Data Platform (created by MCX) which is a repository of analytical data.
Uncontrolled Parking Zones
In the case of the described solution, the analyzed data is information on vehicles parking in the unguarded paid parking zone and verification of payment of the parking fee.
From the technical side, the system's task is to recognize license plates located in the indicated zone, determine the parking spot and send the collected data to the system, which verifies the parking space and payment.
Recognition of parking vehicles and precise determination of the parking position is possible thanks to the use of technical possibilities offered by high-resolution cameras, LiDAR scanners and GPS. Measurements data is subjected to pre-treatment in order to anonymize information and adjust the format. Next the data is send via LTE network (soon 5G) to municipal systems.
To ensure mobility, the equipment was mounted on deignated electric cars, whose main task is to move in assigned zones and scan the city space. Thanks to the advanced optics used in the system's vision devices, it is possible to recognize the registration plates of vehicles parked perpendicularly, at an angle, or even parallel to the road section along which the scanning vehicle is moving. This is also possible at night and in poor visibility. The systems recognizes corretly 95% of car license plate and determines the position of vehicles with an accuracy of a few centimeters. The accuraccy of the readings is maintained with the speed of the scanning vehicle of up to 100 km per hour.
Project after 6 months
Benefits of the solution
The easiest way to talk about beneficial impact is by looking at specific examples and showing the effects. One of such examples is the implementation carried out in Warsaw, which was launched in January 2020.
Thanks to the tests carried out earlier, the project’s risk was low. The first carried try outs showed high efficiency of e-control: two electric cars with the installed equipment were able to identify about 260 vehicles within 10 minutes. For a single inspector, it takes an hour on average. In such a large area as the paid parking zone in the capital city, foot patrols were able to inspect only 40% of the allocated spots daily. Thanks to the new solution, each parking spot will be checked at least once a day.
Another advantage of the tool is the complete protection of privacy - the system automatically anonymizes the faces of passersby as well as other license plates, thanks to which personal data remains anonymous. The e-control system started to bring positive results almost immediately: after just two weeks, revenues from fees paid increased by as much as 10%, and the entire project paid off after 11 days of operation.
Within the first 6 months from the launch of the project, two cars generated an additional income to the city budget in the amount of 7.5 million PLN.
Searching for empty parking spots
Searching for vehicles parked in unauthorized places
Identifying searched for or stolen vehicles
Inventory of municipial advertising space