Optical Sensors Using Artificial Intelligence Monitor Pedestrian Movement In Brno
Thanks to artificial intelligence, special optical sensors are helping to map the movement of pedestrians on namesti Svobody in Brno, the number of cyclists using cycle paths, and the number of vehicles passing through the Želešice municipality. Photo credit: VisionCraft
Brno, June 29 (BD) – The small device tracking the movement of pedestrians, powered by public lighting, collects anonymised data; the device does not send any videos or photos, but only information about the type of object and its behaviour.
The optical sensor is a small device equipped with a computing unit, a camera, GPS and a GSM modem. “From a software perspective, the sensor uses computer vision and artificial intelligence algorithms,” said Robert Pinkas, CEO of VisionCraft, the Brno-based developer of the device. “The software for object detection runs directly in the sensor, so it is not necessary to transmit huge image data over the GSM network, but only information about the type and behaviour of the object.”
Besides the involvement of artificial intelligence, the sensor’s main innovation is its battery operation and the possibility of charging from public lighting. “This makes it ideal for urban and municipal environments. Within minutes, the sensor can be installed on virtually any lighting pole without complicated installation. During the day, it is able to operate on the energy stored in the battery, and at night it is recharged,” said Michal Jukl, ICT Director of Technical Networks Brno (TSB), which supports the technical operation of the sensors and supplies the service for the City of Brno.
From its position on the lighting mast, the sensor monitors the surroundings and detects types of moving objects, including vehicles, people, cyclists, and others. It then sends information about the type of object and its movement in the scene to the cloud. Real-time data processing then depends on the purpose of the sensor. No configuration and learning of the operator software is required from the end user. The data and alerts respond to the customer’s existing systems. The data is also stored historically in the cloud, so analyses of object movement can be carried out at any time after the sensor is installed.
“Due to the versatility of the sensor and the possibility of deploying different detections, the applications are almost limitless. Currently, the sensors are used to count pedestrian passage, cyclist movement, checking the correct passage of vehicles – including the differentiation of cars, vans and trucks -, counting the capacity of street passage or parking areas,” explained Pinkas. He added that the sensors can also be used to count objects in closed locations, such as parks, rest areas or cultural events.
A concrete example is the monitoring of pedestrian flows on Brno’s namesti Svobody, where sensors detected how many people entered and left the square via Masarykova. They are also used in the square during cultural events, where the data is used to provide an accurate estimate of the number of visitors.
“For the City of Brno, sensors will be deployed to detect the number of cyclists in selected locations, where it is also possible to check their equipment, such as helmets. The effectiveness of campaigns to promote cycling safety and the impact of changes in weather and traffic on the number of cyclists passing through will then be measured,” says Jukl.