I.J. Education and Management Engineering, 2016, 1, 18-31
Published Online January 2016 in MECS (http://www.mecs-press.net)
DOI: 10.5815/ijeme.2016.01.03
Available online at http://www.mecs-press.net/ijeme
A Novel Vehicle Classification Model for Urban Traffic Surveillance
Using the Deep Neural Network Model
Kamini Goyal, Dapinder Kaur
Research Scholar, CGC COE Landran, Mohali, India
Assistant Professor, CGC COE Landran, Mohali, India
Abstract
The vehicle detection is the backbone of the urban surveillance systems, which is used to obtain and identify
the various statistics of the urban vehicular mobility. Also the urban surveillance systems are used for the
vehicle tracking or vehicular object classification. The proposed model has been designed for the purpose of
the urban surveillance and vehicular modelling of the traffic. The proposed model has been designed for the
vehicle position identification as well as the vehicle type classification using the deep neural network. The
proposed model has been tested with a standard dataset image for the result evaluation. The experimental
results has been shown the effectiveness of the proposed model, where the proposed model has been found
successful in detection and classification of all of the vehicles in the given image.
Index Terms: Deep Learning, Deep Neural Network, Urban Surveillance, Vehicle Classification, Vehicle
Detection.
© 2016 Published by MECS Publisher. Selection and/or peer review under responsibility of the Research
Association of Modern Education and Computer Science.
1. Introduction
1.1. Vehicle Detection
Vehicle detection is defined as detecting the vehicles on the basis of parameters such as color, shape and size.
Vehicles are detected usually by extracting the vehicle queues from the satellite images. The vehicles can be
detected with the help of neural network i.e. convolutional neural network. The complete system is trained in
order to classify, locate and detect the objects in images. Hence this can improve the accuracy of classification,
detection and localization. The network can be applied at multiple locations in the image using the sliding
window technique. Then the system is trained to produce prediction of the size and location of bounding box. A
technique is defined to perform object localization with convolutional network based segmentation. The central
* Corresponding author. Tel.:
E-mail address: