THE 3
rd
INTERNATIONAL SCIENTIFIC CONFERENCES OF STUDENTS AND YOUNG RESEARCHERS
dedicated to the 99
th
anniversary of the National Leader of Azerbaijan Heydar Aliyev
121
EMERGENCY ALERT SYSTEM USING CONVOLUTIONAL
NEURAL NETWORKS
Alikram Masimzada
Baku Higher Oil School
Baku, Azerbaijan
elikram.mesimzade.std@bhos.edu.az
Supervisor: Ph.D Associate Professor Leyla Muradkhanli
Keywords:
Machine Learning, Crash analysis, Deep Learning, Convolutional Neural
Network
Introduction
Compared to previous decade, it is not complicated to find out traffic
offender by the use of both radar and surveillance cameras. However,
despite the reduced amount of traffic offender, but still they are dangerous
and statistic is 4000 deaths and more than 100000 injuries in a day by car
accidents. In modern countries, mostly all roads are controlled by surveillance
cameras, but again there are some considerably amount problems:
• As there are a lot of cameras, it is required plenty of operators (single
operator can monitor several cameras in a real time).
• If any accidents occur, at that time mostly the people around whether
think someone already called the emergency or they forget to call
emergency.
• Human being cannot be always reliable; operators will not monitor all
the cameras for all of their working hours.
These factors require to have automated solution to call emergency in
case of any traffic accident. The use of Emergency alert system in traffic by
Convolutional Neural Networks (CNN) will prevent such problems and by the
applying of that emergency alert system, the death and injury rate will
decrease.
Dostları ilə paylaş: