Traffic Statistics Deep Learning

Traffic Statistics Deep Learning Analytics provides real-time verifiable counting data for pedestrians and vehicles using Artificial Intelligence and Deep Learning technology.

Products

Product overview

Thanks to the use of video based deep learning algorithms, the Citilog Traffic Statistic Deep Learning solution offers multimodal counts in a single system.

It provides statistical data of pedestrians, bicycles and motor vehicle users. Traffic flows as well as bicycle and pedestrian ones can be quantified, in increasingly complex road spaces, where traffic lanes stand alongside dedicated bicycle and pedestrian pathways. The counting of these vulnerable road users is efficient either in a dedicated lane or in mixed traffic.

There is also the possibility to visualize, in real time, the trajectories of passing vehicles. This is a strong tool to learn on road users’ behavior and be able to react to new mobility trends.

ASK FOR A QUOTE

Citilog Traffic Statistics with Deep Learning, a state-of-the-art traffic analytics system. A flexible and cost-efficient alternative to traditional technology.

Play Video

Identification

Unlike many other technologies, the Citilog Traffic Statistics solution can categorise many sorts of vehicles, even if they have similar dimensions and shape, for example detecting the difference between motorcycles and bikes in complex situations.


Vulnerable road users

The solution offers multimodal counts in a single system with a focus on vulnerable road users counting (pedestrians and bicycles) in addition to motor vehicle users. This is perfectly in line with new mobility trends all over the world where vulnerable road users take a much more significant place in the public space. Traffic flows as well as bicycle and pedestrian ones can be quantified, in increasingly complex road spaces, where traffic lanes stand alongside dedicated bicycle and pedestrian pathways. The counting of these vulnerable road users is efficient either in a dedicated lane or in mixed traffic.

Flexible and non-intrusive

The Citilog Traffic Statistics solution is a durable, flexible and scalable solution. It can be easily reconfigured if the traffic lane scheme is altered. No additional maintenance is required in comparison to video-surveillance cameras, and there is no need for road closure to manage it. Moreover, if the camera view slightly moves, the configuration can be adjusted remotely in just a few clicks, avoiding expensive on-site service. The solution does not require any specific calibration and is ready to use as soon as it is deployed, delivering detailed traffic data 24/7.

Architecture

Traffic Statistics Deep Learning Analytics is compatible with centralised architecture and on-site light architecture. This enables it to be easily deployed in cities equipped with fiber connectivity where all video streams are sent to a centralised server. The solution can also be deployed at unique measure points, where the system is directly installed on the chosen site.

The Traffic Statistics solution can differentiate between the following vehicle classes:

  • Car
  • Truck
  • Bus
  • Motorcycle
  • Bicycle
  • Pedestrian

The Traffic Statistics Deep Learning solution developed to manage complex situations, such as:

  • Congested traffic
  • Harsh weather conditions
  •  Night-time
  •  Various road and intersection configurations
  •  Mono or Bi-directional traffic flows

Use cases

Generated traffic data can be downloaded through remote access, or be automatically pushed to third-party systems, such as the TagMaster VDA-net platform. This gives the user direct access to trends and data over days, weeks, and months. The user can also search historic data, also in from a legacy system. That enables the user to directly identify unusual periods where the traffic is particularly heavy, and then link it to specific events in the managed zone.


The Citilog Traffic Deep Learning solution is designed to be used in various situations of traffic data collection and monitoring. Thanks to traffic reports based on aggregated or individual data, its users can handle general traffic studies, or more specific ones focused on pedestrians and bicycles. This helps building a more relevant urban planning, as well as providing actionable insights to help evaluating its effectiveness.


Cameras

Traffic Statistics Deep Learning solution is designed and optimized for cities with the ambition to generate real time, actionable data from new cameras or existing CCTV-systems.

The solution is camera agnostic, any camera with Real Time Streaming Protocol (RTSP) support can be used. So, this solution can retrofit existing video surveillance infrastructure or be deployed together with new cameras.

Build your Traffic Statistic solution

The Traffic Statistics Deep Learning solution is built around an analytics module CT-TS DL Analytics, that processes, in real time, video streams from cameras to automatically generate reliable traffic statistics; a configuration software CT-Center, to centralize counting data coming from the analytics and provide it either to the on-premises Citilog operator interface CT-Center Client or to a 3rd party system.



ASK FOR A QUOTE