VEHICLE DETECTION, COUNTING AND

CLASSIFICATION USING DEEP LEARNING.

Under the Guidance of Presented By:

Mrs. Anitha Padigapati Batch No : D13
(Assistant professor) M.SUMUKHESH GUPTA
(16K91A05L7)
P.AKHILESH
(16K91A05P8)
S.NIKHILESH
(16K91A05N2)
CONTENTS
1.Introduction
2.Existing system
3.Proposed system
4.ModulesSystem
5.Requirements
6.Conclusion
7.Future Enhancements
INTRODUCTION
• Vehicle detection and counting in aerial images have become an
interesting research focus since the last decade. It is important for
a wide range of applications, such as urban planning and traffic
management.
• Many methods have been introduced for solving this problem.
These methods are either based on shallow learning or deep
learning approaches. However, these methods suffer from
relatively low precision and recall rate.
• The main purpose Vehicle detection and counting from video
sequence method is to Tracking and counting vehicles for High
quality videos and also identify the type of each vehicle
EXISTING SYSTEM
• Vehicle detection and counting is a challenging task due to many
reasons such as: small size of the vehicles, different types and
orientations, similarity in visual appearance of vehicles and some
other objects and detection time in very high resolution images is
another challenge that researchers need to take in consideration.

• Disadvantages of Existing System:

• Unable to detect in poor visual video streams.
• There is no vehicle classification.
PROPOSED SYSTEM
• We propose to solve vehicle detection and counting problem by a
fully convolutional regression network (FCRN).
• During training, an input image and its corresponding ground truth
are given to the FCRN where the goal is to minimize the error
between the ground truth and predicted output.
• During inference, the output of the trained model goes under an
empirical thresholding after which a simple connected component
algorithm is used for returning the count and the location of the
detected vehicles.
 MODULE DESCRIPTION

• Region of Interest selection:

In the very first frame of the video, We define a ROI by drawing a
close line on the image. The goal is to recognize that ROI in a later
frame, but that ROI is not a salient vehicle. It is just a part of an
vehicle, and it can deform, rotate, translate and even not be fully in
the frame.
• Vehicle Detection:
Active strategy to choose a search window for vehicle detection
using an image context was proposed a deep CNN framework
(AttentionNet) to capture the vehicle by sequential actions with top-
down attention.
• DNN Training:
One of the compelling features of our network is its simplicity: the
classifier is simply replaced by a mask generation layer without
any smoothness prior or convolutional structure. However, it
needs to be trained with a huge amount of training data: vehicles
of different sizes need to occur at almost every location.
• Vehicle Tracking:
 Visual tracking solves the problem of finding the position of the
target in a new frame from the current position. The proposed
tracker dynamically pursues the target by sequential actions
controlled by the Adnet’s.
 The ADnet predicts the action to chase the target moving from
the position in the previous frame.
• Location updating
The Adnet is designed to generate actions to find the location
and the size of the target vehicle in a new frame. The Adnet
learns the policy that selects the optimal actions to track the
target from the state of its current position.
• Vehicle Counting:
In this module detected vehicles will be counted and these
counted results will be updated frequents based on vehicle
detection, results will be printed streaming video using OpenCV.
SYSTEM REQUIRMENTS
HARDWARE REQUIREMENTS:
• Processor : Intel i3 and above
• RAM : 4GB and Higher
• Hard Disk : 500GB: Minimum

SOFTWARE REQUIREMENTS:
• Programming Language / Platform : Python
• IDE : PyCharm/Jupiter
CONCLUSION:
FUTURE ENHANCEMENTS
• Usage of Frequency:
In further versions of the system we can use the frequency of the
vehicles, determined by the deep learning system, to identify or to
predict the amount of traffic present in the region of interest.
• Determining the speed:
In further improvements in this system we can develop a method to
determine the speed of the vehicle that is travelling in the region of
interest. This could be futher used to identify vehicles which are
travelling over the speed limit.
THANK YOU!!