Cities are first of all built for people, and in today’s world, science and technology hold the power to transform the places we live in into highly efficient, technologically driven havens – the so-called smart cities. The Smart City Paradigm is a key for making the growing number of cities around the globe more tech-savvy, more connected, more secure, more ecological – and with all that, they can guarantee the top quality of life for urban dwellers.

Abto Software has already contributed to “smarter” use of the integral part of modern cities – the transport networks – having developed Automatic Vehicle Detection and License Plate Recognition Software and invested in the means of using CCTV cameras for car theft detection. These technologies are some of the many that make up the Intelligent Transportation System (ITS) – an advanced application that provides transport solutions by utilizing state-of-the-art approaches and innovative techniques. Aforementioned solutions incorporate:

  • Vehicle Detection & Traffic Management;
  • Automated Vehicle Counting & Classification;
  • Real-Time Traffic Light Control System;
  • Automated Vehicle Identification & License Plate Recognition;
  • Parking Management & Enforcement;
  • Automated Toll Collection;
  • Real-time Passenger Information System;
  • Public And Road Cleaning Transport Scheduling;
  • Traffic Law Enforcement & Safety Systems;
  • Travel Assistance & Navigation.

With the aim to work in the field of Intelligent Transportation Systems Abto R&D team set the goal to eradicate one of the most annoying and time-consuming situations that happens to every driver – fuming at red lights. The operation of standard traffic lights deployed at most crossroads is based on predetermined timing schemes, which are fixed during the installation and can be adjusted only by manual resetting. The timing used is an over-simplified sequence that fails to adapt to and efficiently manage real-time traffic flow conditions that are changing throughout the day and varying for every road junction and intersection.

Objective

Our computer vision engineers decided to develop an Intelligent Crossroad Traffic Management System (ICRTMS). The first task undertaken is to calculate the number of vehicles crossing the intersection in each direction. At this stage, the focus is on developing the real-time traffic flow measurement system meaning counting the overall number of vehicles while not taking into consideration their type (e. g. is it a car, bus, motorcycle or truck). Such categorization would be a part of one of the next stages, namely development of Automated Vehicle Counting & Classification System.

Abto’s Approach to Traffic Flow Measurement

Conventional though not wide-spread methods of Traffic Flow Measurement rely heavily on the usage of RFID technology and traffic sensors that is both expensive and hard to maintain or replace – hence limited installations. Applying Abto extensive experience in Computer Vision and, in particular, object counting to vehicle detection and counting overcomes these drawbacks in the blink of an eye – no pun intended. The computer vision technology allows measuring traffic flow from the standard CCTV camera stream making our solution non-intrusive, fully wireless and easy to install or adjust.

We reached the desired result by carrying out the next operations:

  1. Vehicle detection that can be solved either by a straightforward approach or machine learning and deep learning algorithms. The first option proved to be more efficient for the ordinary detection while we plan to apply machine learning during the next stages, namely Automated Vehicle Classification.
    Thus, we defined as objects of interest everything that is not a background which we estimated through image processing. The bounding boxes of detected vehicles on the demo video below are highlighted in different colors – depending on which intersection border they crossed first. The intersection area is also highlighted and its boundaries are given the numbers that will allow us later to fill the traffic count matrix.
  2. Trajectories definition for which we extended the feature space we track the vehicles in by adding extra dimensions to the traditionally used X/Y coordinates of the detected cars. Such approach was developed and proven to be accurate during the development, surprisingly, of Abto’s Real-time Hockey Player Tracking Technology.
  3. Traffic count matrix calculation that is carried out on the basis of defined trajectories which we compare to the two counting lines, that is the borders of the crossroad numbered from 1 to 4. The (i, j) element of the square traffic count matrix M[4, 4] represents the number of < i | j > vehicle transitions – vehicles that entered the road intersection through the i-th boundary and left through the j-th boundary.
    For example: if the vehicle first crosses the line number 1 (for demonstrative purposes the car will also become highlighted in the navy blue color assigned to that line) and then leaves the road intersection through the line number 2, the (1, 2) element of the traffic transition matrix increments by one.

The one minute 60fps demo video below shows how Abto Traffic Flow Measurement Technology processes the CCTV footage of this crossroad you can check out on the Google Maps with 100% accuracy rate. The calculated traffic count matrix allows to conclude the next:

  • no vehicles made a U-turn as all the diagonal values are equal to zero;
  • no vehicles violated traffic rules by leaving the road intersection through the third boundary;
  • the biggest traffic flows are observed in the next three directions: 1) from the first intersection border to the second (14 cars during the observed time period); 2) from the third border to the first one (10 cars during the observed time period) and 3) from the fourth border to the second one (9 cars during the observed time period).

We plan to extend the functionality of our Traffic Flow Measurement system so it can be used for comprehensive crossroad traffic management by adding:

  • lane-by-lane vehicle speeds and gaps;
  • stopped-vehicle notifications;
  • wrong-way vehicle alarms.

Vehicle Counting Software Technologies

The described prototype of the Abto software solution for Traffic Flow Measurement was developed in MATLAB using the next technologies:

  • Computer vision;
  • Image & video processing;
  • Object & motion detection;
  • Background estimation & subtraction;
  • Multiple object tracking;
  • Trajectory definition & binding.

Areas of Application & Benefits

The developed Traffic Flow Measurement System can be applied for:

  • Adaptive Traffic Light Control System: the lane-by-lane occupancy readings and vehicle counts can be the input for the Intelligent Traffic Light Control System based on real-time traffic flows;
  • Traffic Law Enforcement: the Abto solution can detect traveling in the wrong direction by sending wrong-way vehicle alarms;
  • Intelligent Mapping of Public Transportation: additional public transport routes can be added to the transport network to facilitate the traffic conditions in the busiest directions or during the detected peak hours ;
  • Urban Planning: the statistics collected should be considered for smart city development plans.

The integration of Abto Traffic Flow Measurement System into urban infrastructure allows to:

  • Improve traffic flow;
  • Speed up responses to traffic conditions;
  • Lower transportation costs;
  • Ease out traffic stress;
  • Reduce pollution and fuel consumption.

Contact Us

To find out more about Abto Software expertise, request a quote or get a demo of your custom solution.

  • Clicking this button, I agree to the processing of my personal data.
Insert math as
Block
Inline
Additional settings
Formula color
Text color
#333333
Type math using LaTeX
Preview
\({}\)
Nothing to preview
Insert