Docsity
Docsity

Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity


Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan


Guidelines and tips
Guidelines and tips

Object detection in detectron2 using pytorch on google colab. Reuse already trained model or import existing trained model and predict the objects, Exercises of Gastroenterology

6. Effective Storage management includes a) Securities b) Backups c) Reporting d) All of the mentioned View Answer Answer: d

Typology: Exercises

2019/2020

Uploaded on 10/20/2020

nikhil-patil-5
nikhil-patil-5 🇮🇳

4 documents

1 / 39

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
A Project Report
On
“SMART TRAFFIC CONTROL USING DEEP
LEARNING
Submitted to
SHIVAJI UNIVERSITY, KOLHAPUR
In Partial Fulfilment of the Requirement for the Award of
BACHELOR OF ENGINEERING
(INFORMATION TECHNOLOGY)
BY
Chintamani Naik 23
Yash Mudgal 24
Nikhil Patil 25
Sandeep Pandita 26
Vivek Patil 27
UNDER THE GUIDANCE OF
(Mr.J.S.Pujari)
DEPARTMENT OF INFORMATION TECHNOLOGY
KIT‘s COLLEGE OF ENGINEERING KOLHAPUR
2019-2020
AFFILIATED TO
SHIVAJI UNIVERSITY, KOLHAPUR
pf3
pf4
pf5
pf8
pf9
pfa
pfd
pfe
pff
pf12
pf13
pf14
pf15
pf16
pf17
pf18
pf19
pf1a
pf1b
pf1c
pf1d
pf1e
pf1f
pf20
pf21
pf22
pf23
pf24
pf25
pf26
pf27

Partial preview of the text

Download Object detection in detectron2 using pytorch on google colab. Reuse already trained model or import existing trained model and predict the objects and more Exercises Gastroenterology in PDF only on Docsity!

A Project Report

On

“SMART TRAFFIC CONTROL USING DEEP LEARNING”

Submitted to

SHIVAJI UNIVERSITY, KOLHAPUR

In Partial Fulfilment of the Requirement for the Award of

BACHELOR OF ENGINEERING

(INFORMATION TECHNOLOGY)

BY

Chintamani Naik 23

Yash Mudgal 24

Nikhil Patil 25

Sandeep Pandita 26

Vivek Patil 27

UNDER THE GUIDANCE OF

(Mr.J.S.Pujari)

DEPARTMENT OF INFORMATION TECHNOLOGY

KIT‘s COLLEGE OF ENGINEERING KOLHAPUR

AFFILIATED TO

SHIVAJI UNIVERSITY, KOLHAPUR

A

Project Report

On

“SMART TRAFFIC CONTROL USING DEEP LEARNING”

SUBMITTED TO

Shivaji University, Kolhapur

In Partial Fulfilment of the Requirement for the Award of

Bachelor’s Degree In

INFORMATION TECHONOLOGY

Submitted By

CHINTAMANI NAIK 23

YASH MUDGAL 24

NIKHIL PATIL 25

SANDEEP PANDITA 26

VIVEK PATIL 27

Under The Guidance Of

Mr. J.S.Pujari

DEPARTMENT OF INFORMATION TECHNOLOGY

KIT‘s College Of Engineering(Autonomous), Kolhapur

DECLARATION

We hereby declare that the project work report entitled "Smart Traffic Control

Using Deep Learning” is being submitted to K.I.T’s College of Engineering ,

Kolhapur affiliated to Shivaji University, in partial fulfillment of Bachelor of

Engineering (Information Technology) course, is a bonafide report of the work

carried out by us. The material contained in this report has not been submitted to

any university or institution for the award of any degree.

Submited By

Mr. Chintamani D. Naik

Mr. Yash J. Mudgal

Mr. Nikhil A. Patil

Mr. Sandeep Pandita

Mr. Vivek B. Patil

Place: KITCOEK, Kolhapur

Date: / /

Acknowledgements

We are highly grateful to Mr. T. B. Patil , HOD IT, KIT’s College of Engineering,

Kolhapur, for providing this opportunity to carry out the Project at the IT depart-

ment. We would like to expresses our gratitude to other faculty members of the

IT department for providing academic inputs, guidance encouragement through-

out this period. We would like to express a deep sense of gratitude and thanks to

Mr. J.S.Pujari without whose permission, wise counsel and able guidance, it would

have not been possible to carry out our project in this manner. Finally, we express

my indebtedness to all who have directly or indirectly contributed to the successful

completion of our project.

KIT’s College of Engineering, (Autonomous), Kolhapur Page ii

List of Figures

List of Tables

3.3.4. Software Quality Attributes KIT’s College of Engineering, (Autonomous), Kolhapur Page iv

Table of Contents

Title Page No.  Certificate ii  Acknowledgement iii  Abstract iv  List of Figures v

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 1

1. Introduction

The problem of traffic congestion is increasing with the increasing number of vehicles and a solution is required which can adapt to the changing traffic. The traditional situation of handling traffic was with the help of a traffic policeman or officer. Later the traffic signal system was used but they have fixed time allocation techniques. Both of this solutions fail to provide the answer to ever changing incoming traffic handling. Therefore there is need for smart traffic control which can adapt to changing traffic and provide time allocation technique for all lanes at a single signal and alert other connected signal of incoming traffic ahead of time. Vehicle detection allows the use of various applications of artificial intelligence system for several purposes as intelligent transportation, automatic monitoring, autonomous driving, and driver safety guarantee. In this work, we focus on the detection and recognition of vehicles in a video stream. For this reason, we have used the convolutional neural network technique (CNN) and a dataset that contains images to enable recognition and classification of vehicles.Convolutional neuron network efficiency depends to a large extent on the quality of the training data set; the network will produce good results only if the training data used contain sufficient important characteristics so that they can produce new predictions. 1.1 Problem Statement The traffic handling schemes that are in use today are fixed time allocated traffic signal which do not change on incoming traffic or fail to provide time allocation scheme over changing traffic. Where traffic signals are not present this work is carried out by traffic policeman or traffic officer who is able to assess the incoming traffic and provide handling of traffic based on changing traffic. Both of this technique are not suitable where heavy traffic is present or its constantly changing. A solution is required to the traditional traffic signal which can provide better handling of incoming traffic and alert ahead of time of incoming traffic. Thus there is need to make smart traffic control system which can identify types of vehicles in a video frame belonging to categories of car, truck, bikes and buses along with number of vehicles present to control traffic by adjusting traffic signal timing for each individual lane and send this data to its connected signals and alert them of incoming traffic to calculate respective time allocation for each individual lane by using deep learning algorithms and object detection. 1.2 Purpose The purpose of this project is develop a desktop, stand-alone application for of Smart Traffic Control Using DL. An application which makes automatic allocation of time for traffic signal according to different conditions of traffic. The different software requirements of Smart Traffic Control Using DL are described in the Document which will outline all the technical aspects regarding development of Smart Traffic Control Using DL. It will illustrate the purpose and complete declaration for the development of system. It will also explain system constraints, interface and interactions with other external applications. This document is primarily intended to be proposed to a customer for its approval and a reference for developing the first version of the system for the development team. The software requirement specification in this document is an all time reference for the basic requirements for proceeding with the development.

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 2

1.3.Scope

The project deals with the development of desktop, stand-alone application for the purpose of Smart Traffic Control Using DL. The application is supposed to change the way the manual signal assignment is carried out for different situations. The project can help government to achieve increase in efficiency of signal allocation scheme as defined and used till date by the traffic regulations authorities. We focus on the detection and recognition of vehicles in a video stream. For this reason, we have used the convolutional neural network technique (CNN) and a dataset that contains images to enable recognition and classification of vehicles. The input video stream is CCTV at signal which provides live images. To identify types of vehicles in a video frame belonging to categories of car, truck, bikes and buses along with number of vehicles present to control traffic by adjusting traffic signal timing for each individual lane and send this data to its connected signals and alert them of incoming traffic to calculate respective time allocation for each individual lane by using deep learning algorithms and object detection. The video is converted into frames for processing and further calculations are carried out accordingly and results are sent to traffic signal timer which updates the different types of signals for different lanes. Also to make the system capable of carrying out its work on live CCTV footage. 1.4 System Analysis 1.4.1 Existing System

  1. Traffic Officer - a single officer in charge of all lanes of traffic
  2. Traffic Signal - a fix time allocation based traffic signal The traffic handling schemes that are in use today are fixed time allocated traffic signal which do not change on incoming traffic or fail to provide time allocation scheme over changing traffic. Where traffic signals are not present this work is carried out by traffic policeman or traffic officer who is able to assess the incoming traffic and provide handling of traffic based on changing traffic. Both of this technique are not suitable where heavy traffic is present or its constantly changing. 1.4.2 Limitations of Existing System
  1. The traffic officer has to manually allocate handling of traffic.
  2. The traffic officer may not always give the best time for all traffic lanes.
  3. The traffic signal has fixed time for all lanes which does not change.
  4. Traffic signal fail to adapt to different situations of incoming traffic.

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 4

2. The Overall Description

2.1 Product Perspective For development of the proposed system we have used the convolutional neural network technique and a dataset that contains images to enable recognition and classification of vehicles. Convolutional neural network efficiency depends to a large extent on the quality of the training data set; the network will produce good results only if the training data used contain sufficient important characteristics so that they can produce new predictions. The proposed solution will have an object detection model which can identify different types of vehicles and it is built in Tensorflow. To develop an object detection model we need to give training images and apply pre-processing then pass the dataset to CNN algorithm. The model will be used for identifying different vehicles. The proposed project is supposed to work on as a desktop application or standalone as far as the project is concerned. 2.2 Product Functions  The system will take input as a live video stream and apply object detection on it.  The system will display the different types of objects detected in a single frame and label them as cars, trucks, bikes, etc.  The system will initialize the counter of signal which is depend on number of object detected.  The system will be send the count of the objects to next signal.  The system will initialize the counter of the signal on basis of the data received from the previous signal or on its own 2.3 User Characteristics User of the system in this case is the field officer who should be able to login using the ID provided. The ID is unique to the field officer who will be monitoring the traffic signal. The field officer can login to application and should be able to perform the following operations  Login into the system against his ID to keep at track of his login status o Select input CCTV footage o Observe and monitor vehicles detected o Monitor traffic signal timers o Get results for traffic analysis and signal timer o Login using a single window on the server o Check the details of the current traffic o Check if proper lane assignment is carried out

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 5

3. SPECIFIC REQUIREMENTS

This section contains all the software requirements at a level of detail, that when combined with the system context diagram, use cases, and use case descriptions, is sufficient to enable designers to design a system to satisfy those requirements, and testers to test that the system satisfies those requirements. 3.1 External Interfaces The Smart Traffic Control using Deep Learning will use the standard input/output devices for a personal computer. This includes the following:  Keyboard  Mouse  Monitor  Web camera/CCTV 3.1.1 User Interfaces Registration New user registration Login Log into the system Home This Contains Services that user can access Control Panel Observe traffic and time allocation schemes Table 1: User Interfaces 3.1.2 Software Interfaces OpenCV It mainly focuses on image processing, video capture and analysis including features like object detection. Tenserflow It is an open source artificial intelligence library, using data flow graphs to build models. Anaconda Spyder IDE for python application development Anaconda For compiling python files as we will be using python native to desktop application. Table 2: Software Interfaces

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 7

3.2.Functional Requirements

1 User Module: 1.1 User Registration  User Register with Name, RTO ID, Phone Number, E-mail address. 1.2 User Login  User Login with Username and Password.  Forgot Password facility for Retrieving Password 1.3 Product Functions The major functions of the product to be developed are shown in the block diagram below. Figure 1 : Block diagram of application

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 8

3.3.Nonfunctional Requirements

3.3.1 Performance Requirements  The application should not crash under any circumstances. Care must be taken to avoid breaks in input stream.  The application should be compatible across various operating versions of Windows, Linux  Easy login and logout sessions for officers  Application should process images in faster manner and in low amount of time 3.3.2 Safety Requirements The make sure the data is not lost due to hard disk crashes or any other failures regular backup of collected data should be maintained. This requires regular backups of the data to be made every day to prevent any data loss of collected price. 3.3.3 Security Requirements The password should be encrypted to prevent any data hacks so that incorrect assignment or sensitive data won’t be compromised. Additionally the timer values calculated needs to be accurate. 3.3.4 Software Quality Attributes 3.3.4.1 Reliability The system will consistently perform its intended function. For e.g. The time allocation scheme for traffic. 3.3.4.2 Efficiency Unnecessary data will not be transmitted on the network. Application should process images in faster manner and in low amount of time 3.3.4.3 Reusability The system can be reused in any organization master definition under software license agreement. 3.3.4.4 Integrity Only system can change the timer values. Each user will be having rights to access the control panel for observing but not changing the data.

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 10

4. SOFTWARE MODEL

For this project we have used prototyping model. Prototype is a working model of software with some limited functionality. The prototype does not always hold the exact logic used in the actual software application and is an extra effort to be considered under effort estimation. Prototyping is used to allow the users evaluate developer proposals and try them out before implementation. It also helps understand the requirements which are user specific and may not have been considered by the developer during product design. Following is a stepwise approach used to develop a software prototype. Basic Requirement Identification This step involves understanding the very basics product requirements especially in terms of user interface. We collected object detection requirement for different types of vehicles. Developing the initial Prototype The initial Prototype is developed in this stage, where the very basic requirements are showcased and user interfaces are provided. These features may not exactly work in the same manner internally in the actual software developed. We developed a working prototype for object detection with cars and trucks.

Review of the Prototype

The prototype developed is then presented to the customer and the other important stakeholders in the project. The feedback is collected in an organized manner and used for further enhancements in the product under development. Revise and Enhance the Prototype The feedback and the review comments are discussed during this stage and some negotiations happen with the customer based on factors like – time and budget constraints and technical feasibility of the actual implementation. The changes accepted are again incorporated in the new Prototype developed and the cycle repeats until the customer expectations are met.

Smart Traffic Control using DL KIT’s College of Engineering, (Autonomous), Kolhapur Page 11

5. PROJECT PLANNING

Figure 2 : Project Planning in GANNT chart Figure 3: Project Planning Schedule