The design of a Smart Parking system requires the ability to estimate the number of free parking spaces and the number of occupied spaces in a car park, whether it is on the street (as is the case of a city centre for example) or within a perimeter, i.e. delimited by an entrance and an exit gate.

A.I. Tech is able to satisfy the various requirements for counting and monitoring parking spaces. With AI-Road3D it is possible to count the number of vehicles at the gates, and with AI-LPR it is possible to estimate not only the number of vehicles but also the average time they spend in the car parks. Finally, with AI-PARKING we can monitor the specific parking space, and determine whether it is free or occupied.

A special feature of this solution is that it can be used on multiple hardware platforms: edge side (on cameras), server side, or on the AI Appliance platform. Their integrability is practically unlimited thanks to the support of most of the communication protocols used today, such as Modbus/TCP, FTP and HTTP.


AI-PARKING is the video analytics app that can automatically detect the occupancy status of a parking spot, identifying whether it is free or occupied. Thanks to the use of advanced algorithms based on deep neural networks, AI-PARKING-DEEP can be used both in indoor and outdoor environments and only requires the vehicle to be partially visible inside the parking space.


AI-LPR-DEEP is a video analytics app that utilises an advanced artificial intelligence algorithm to perform licence plate detection and recognition. Thanks to the use of an innovative engine based on semantic technologies, it also enables automatic correction of license plates based on the specific nationality of the plate (*).

The solution can detect vehicles up to a maximum speed of 230 km/h (depending on the chosen hardware platform) and can be used both indoors (e.g. for monitoring car parks) and outdoors (e.g. for monitoring city streets).

* Countries for which the semantic engine is currently available: Italy.


AI-ROAD3D is a video analytics app that allows counting and classification of vehicles passing by a virtual sensor in a given direction. Three vehicle classes are considered: motorbikes, cars and trucks. The app also identifies the colour and average speed of each vehicle, and generates an alarm if this speed exceeds a certain threshold chosen by the operator. It is also able to assess traffic volume in real time. AI-ROAD3D combines an advanced 3D calibration and reconstruction mechanism of the scene with the most advanced artificial vision and artificial intelligence algorithms.

The application is available in two versions: AI-ROAD3D uses the latest deep learning algorithms to classify moving objects; whereas AI-ROAD3D-DEEP uses these algorithms for both object detection and classification, guaranteeing high accuracy even in extremely complex scenarios, such as in tunnels or crowded city streets, at night or in severe weather conditions.