한국어 English Tiếng Việt bahasa Indonesia Unisem
한국어 English Tiếng Việt bahasa Indonesia Unisem
Subject

TRAFFIC LAW ENFORCEMENT WITH UNITRAFFIC: TECHNOLOGIES AND ENFORCEMENT …

Name Admin
Date 21-07-30 15:51
View 161

Contents

Traffic Law Enforcement with UniTraffic:
Technologies and Enforcement Principles

This month's issue is going to be the introduction of the technologies and principles implemented in UniTraffic for traffic law enforcement.  

Let's start with the core technology, Deep Learning, the most pervasive form of Artificial Intelligence (AI) these days. Using the Deep Learning technique, we “train” our VAS (Video Analytics System) software to detect and identify the object from an image and classify it according to its training database and improve its performance by continuous “training” and database expansion. Deep Learning also is used to train LPDR (License Plate Detection and Recognition) software to “read” vehicles’ license plates.

Another important technology that makes UniTraffic multifunctional is Multiple Object Tracking (MOT), which allows tracking multiple moving objects. MOT facilitates accurate speed estimation and object trajectory analysis.  This technology is the foundation for the detection of violations made by traffic subjects. 

The principles of violation detection include vehicle tracking and “rules” designed in accordance with traffic regulations. For example, to detect lane violations we apply lines that separate road lanes as shown in Picture 1. MOT tracks all detected vehicles and, if any of them crosses the designated line, then the solution recognizes the vehicle as a violator, and related data including its license plate information is being stored on the server with necessary images as proof. The same principles are used for banned turns


Picture 1. Lane Violation Detection Example


In case of traffic light violation, there is an additional rule that verifies the traffic light status. As it is shown in Picture 2, the restriction line(rule) is placed on the road where is the stop line, and when any vehicle passes that line and enters the intersection, the solution sends violation data to the control room.


Picture 2. Traffic Signal Violation Detection Example 

No loops, radar, or other detection devices are being used in UniTraffic. The camera is used both as a detector and a surveillance device which makes the enforcement system structure much simpler, and, since UniTraffic detects a broad range of violations and provides traffic statistical data at the same time, city investments more efficient.  

Want to learn more contact us via sales@unisemiot.com.

 
PRESS RELEASE
2021.07.06
Asia Business Daily
UNISEM selected as implementing agency for ‘AI Convergence National Safety Security and Rapid Response Support Project’ by the Ministry of Science and Technology

Article in Korean ▷
2021.07.21
Financial News
UNISEM gets more attention in the semiconductor industry with its eco-friendly equipment - Hana Financial Investment 

Article in Korean ▷
UNISEM'S IOT PRODUCTS
  
Facebook
Twitter
LinkedIn

Copyright © 2021 UNISEM, All rights reserved.
 

Subscribe to our Newsletter

I have read and accepted Privacy Policy.