A Deep Learning Tool to Classify Vehicles in Real Time

The problem of vehicular congestion is transversal to all the cities worldwide. To deal with it, it can be done from the construction of new roads or the management of these as in the case of exclusive roads for public transport. In the latter case, information is necessary. Currently, in Chile this...

Full description

Saved in:
Bibliographic Details
Main Authors: Farias, Humberto, Solar, Mauricio, Ortiz, Daniel
Format: Conference Proceeding
Language:English
Subjects:
Deep learning
Optical fibers
Real-time systems
Roads
Smart cities
smart city
smart mobility
Streaming media
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The problem of vehicular congestion is transversal to all the cities worldwide. To deal with it, it can be done from the construction of new roads or the management of these as in the case of exclusive roads for public transport. In the latter case, information is necessary. Currently, in Chile this collection process is manually done, so it does not respond to current information needs, because the long time the process takes. The use of deep learning models for automatic counting has shown outstanding results, but based on a scheme for sending video streaming to remote servers. The present proposal aims to not depend on fiber optic connectivity. It is based on a network of modules based on Edge Artificial Intelligence (Edge AI) in order to build a scalable, real-time data capture platform.
ISSN:2573-1998
DOI:10.1109/ICEDEG52154.2021.9530932