 The basic idea of proposed system is to provide alertness to the driver about the presence of traffic signboard at a particular distance apart. It generates a warning to the driver in advance of any danger. The warning allows the driver to take appropriate actions in order to avoid the accident.The system takes continuous video input from the console monitor or camera installed on the car's bonnet. The underlying algorithm extracts the features of the input image and matches them with an existing library of traffic sign. The output is fed to the driving assistance system and in turn drives the car accordingly. We developed this intelligent system using Machine Learning.This device will take camera feeds and upgrade the system instantaneously.

1. Traffic
Sign Board
Classification
ImplementationReview
Under Guidance of :
YSowjanya
Assistant Professor
IT department

2. Abstract
The universe is governed by a combination of several
laws which are environmental, physical and many
more. Likewise, mankind has created a set of traffic
rules, to guide the people travelling and to regulate
the traffic flow.
Traffic Sign boards are a great source of avoiding
accidents, when observed and followed properly.
It is very difficult for a driver to notice all the sign
boards and act accordingly.
An automatic recognition system is proposed to
recognise the sign boards and alert the driver by a
voice message.
The project can be extremely useful for autonomous
vehicles as it detects signs and helps drivers take the
necessary actions.

3. Existing
System/
Methodology
Summary of
some
reportedTSR
applications:
A TSRvendor support process can aid the operators by alerting
forward road signparticulars, along with prohibitions, warnings and
restrictions.
TSRsystems are avery crucial part of driverless cars getting them
aware of the current publicroad traffic regulations.
By sensingthose types of signsforward,TSRcanreduce energy intake
by finding ideal traffic signsof velocity, reducing the useof breakage.
The
drawbacksof
existing
systemare:-
During internet connectivity issuesor in unchartered terrain.
Smallfuzzy traffic signsand high-resolutionpictures. During bad
weather and innights.
Colordetection in RGB.
Costlier installation.
*TSR –Traffic Sign Recognition

4. Proposed
System
The basic idea of proposed system is to provide alertness to the
driver about the presence of traffic signboard at aparticular
distance apart. It generates awarning to the driver in advance of
any danger. The warning allows the driver to take appropriate
actions in order to avoid the accident.
The system takes continuous video input from the console
monitor or camerainstalled on the car'sbonnet.The underlying
algorithm extracts the features of the input image and matches
them with anexisting library of traffic sign.
The output is fed to the driving assistance system and in turn
drives the car accordingly.Wedeveloped this intelligent system
using MachineLearning.
This device will take camerafeeds and upgrade the system
instantaneously.

5. Functionaland
Non-
Functional
Requirements
System RequirementStudy

6. Functional
Requirements
Preprocessingwill checkcontrast,brightness,and clarity.This block will
makesure the image is readyto have imageprocessingdone to it.
The application of processing algorithms shall take the
preprocessed image and findcolors of interest and look forshapes
relating to the sign or signs we aresearching for.
The classify sign block shall take the regions of interest passed from
the algorithms block.These regions will be analyzed and used to
compare to ‘templates’ of known signs.
The highlight image subsystem shall create some sort of
distinguishing box or highlight aroundthe actual sign.
The recommend appropriate action subsystem shall give a
recommended action as an output based on the type of sign
encountered.

7. The software to be developed must:
1. Detect only road sign boards.
2. Ignore all other objects except road sign boards.
3. Recognize the road signs correctly.
4. Display the road sign in textual format.
5. Convert the text output to voice output.
Non-
Functional
Requirements

8. Design Elements

9. 04
Traffic Sign
Recognition
Detected sign is extracted and fed to
the classifier model to classify the sign
into one of the 43 trainedsigns.
05
Text-to-Speech conversion
Recognized traffic sign is sent to TTS
module for getting voice alert through
car speakers.
03
Traffic Sign Detection
Localize the sign board in the frame
and extracting it as a singlesign.
01
Model building
CNN model is built on GTSRB and
tested with 98% accuracy. 02
Image Input
upload the real-time image and extract
the patterns.
MODULES

10. Unified Modelling Language
(UML) Diagrams

11. CLASS DIAGRAM

12. COLLABRATION DIAGRAM

13. Use Case
Diagram

14. STATE CHART DIAGRAM

15. Sequence Diagram

16. COMPONENT DIAGRAM

17. DEPLOYMENT DIAGRAM

18. System
Architecture

19. IMPLEMENTATION SCREENSHOTS

25. Experimental Results

26. Model
Summary

27. Epoch
Summary

28. Accuracyplot v/sLossplot

30. S.N
o.
Meta Sign Sign Actual Predicted Test
1 General Caution General Caution Pass
2 Children Crossing Children
Crossing
Pass
3 Road Work Road Work Pass
Test
Cases

31. S.N
o.
Meta Sign Sign Actual Predicted Test
4
Round About
Mandatory
Round About
Mandatory
Pass
5 No passing No passing Pass
6 Turn Left ahead Turn Left ahead Pass
Test
Cases

32. This system is used to savethe valuable life by preventing
accidents due to the negligence of traffic signsboards.
At present 40%of deaths that aretaking place these days
aremainly due to the road accidents.
People die in these road accidents which is agreat loss for
the family. Our project provides maximum efficiency and
is userfriendly.
This project mainly focuses on majority of the society who
travel especially the night travelers and it also helps traffic
police to reduce the traffic issues.
The main idea for this project is from the road accidents
that take place due to driver’s ignorance of traffic signs.
Conclusion

33. The Project should be extended to implement real-time.
Traffic sign extraction from the video input is the next work
to be done in this project.
Response time should be improved to a greater extent.
An efficient voice alert should be developed after
classification of the sign label.
FutureScope

34. References
• Aparna A. Dalve, Sankirti S. Shiravale “Real Time Traffic Signboard Detection and
Recognition from Street Level Imagery for Smart Vehicle” International Journal of
Computer Applications (0975 – 8887) Volume 135 –No.1, February 2016.
• POONAM.S.SHETAKE, S.A.PATIL, P.M JADHAV,“REVIEW OF TEXT TO
SPEECH CONVERSION METHODS” International Journal of Industrial Electronics
and Electrical Engineering, ISSN: 2347-6982.
• Anushree. A. S , Himanshu Kumar , Idah Iram , Kumar Divyam , Rajeshwari. J
“Automatic Signboard Detection System by the Vehicles” International Journal of
Engineering Science and Computing, May 2019.
• Yuan Yuan, IEEE, Zhitong Xiong, and Qi Wang “An Incremental Framework for Video-
Based Traffic Sign Detection, Tracking, and Recognition” IEEE TRANSACTIONS ON
INTELLIGENT TRANSPORTATION SYSTEMS, VOL. 18, NO. 7, JULY 2017.
• Safat B. Wali, Mahammad A. Hannan, Aini Hussain, and Salina A. Samad “An
Automatic Traffic Sign Detection and Recognition System Based on Colour
Segmentation, Shape Matching, and SVM” Hindawi Publishing Corporation
Mathematical Problems in Engineering Volume 2015, Article ID 250461, 11 pages
http://dx.doi.org/10.1155/2015/250461.

35. Thankyou
• G.Harivardhan Reddy –19H65A1201
• Sandeep Diviti –19H65A1203
• N.Southen Kumar –19H65A1206