A PROJECT REPORT

ON

“TRAFFIC SIGNS RECOGNISATION”

Submitted in the partial fulfillment of the requirements for
The degree of
BACHELOR OF ENGINEERING IN COMPUTER
ENGINEERING

By

1) SAKSHI SINGH(62)
2) KOMAL LOGADE(42)
3) KARTIK WAGHELA(76)

UNDER THE GUIDANCE OF

Prof. Monali Deshmukh

DEPARTMENT OF COMPUTER
ENGINEERING SARASWATI
COLLEGE OF ENGINEERING,
SECTOR-5, KHARGHAR, NAVI
MUMBAI-410210 UNIVERSITY OF
MUMBAI

ACADEMIC YEAR 2020-2021

 CERTIFICATE

This is to certify that the requirements for the project report entitled “Traffic
Sign Recognition” have been successfully completed by the following students:

Roll numbers Name

62 Sakshi Singh
42 Komal Logade
76 Kartik Waghela

In partial fulfillment of Sem –VI, Bachelor of Engineering in Computer

Engineering, Saraswati college of Engineering, Kharghar, University of
Mumbai, during the academic year 2020-21.

Internal Guide

(Prof. Monali Deshmukh)

Vision:

To be universal accepted as autonomous center of learning in Engineering Education

and Research .

Mission:

➢ To educate students to become responsible and quality technocrats to fufill

society and industry needs .

➢ To nurture students creativity and skill for taking up challenges in allfacets of

life.

Department of Computer Engineering

Vision:

To be among renowned institution in computer engineering education andresearch by

developing globally competent graduates

Mission:

➢ To produce quality Engineering graduates by imparting quality training, hands on

experience and value education.
➢ To pursue research and new technologies in Computer Engineering and across
interdisciplinary areas that extends the scope of Computer Engineering and benefit
humanity.
➢ To provide stimulating learning ambience to enhance innovative ideas, problem
solving ability, leadership qualities, team-spirit and ethical responsibilities.
 DEPARTMENT OF COMPUTER ENGINEERING
PROGRAM EDUCATIONAL OBJECTIVE’s

1. To embed a strong foundation of Computer Engineering fundamentals to

identify, solve, analyze and design real time engineering problems as a
professional or entrepreneur for the benefit of society.
2. To motivate and prepare students for lifelong learning & research to
manifest global competitiveness.
3. To equip students with communication, teamwork and leadership skills to
accept challenges in all the facets of life ethically.
 DEPARTMENT OF COMPUTER ENGINEERING
PROGRAM OUTCOMES
1. Apply the knowledge of Mathematics, Science and Engineering
Fundamentals to solve complex Computer Engineering Problems.
2. Identify, formulate and analyze Computer Engineering Problems and
derive conclusion using First Principle of Mathematics, Engineering
Science and Computer Science.
3. Investigate Complex Computer Engineering problems to find appropriate
solution leading to valid conclusion.
4. Design a software System, components, Process to meet specified needs with
appropriate attention to health and Safety Standards, Environmental and
Societal Considerations.
5. Create, select and apply appropriate techniques, resources and advance
engineering software to analyze tools and design for Computer Engineering
Problems.
6. Understand the Impact of Computer Engineering solution on society and
environment for Sustainable development.
7. Understand Societal, health, Safety, cultural, Legal issues and
Responsibilities relevant to Engineering Profession.
8. Apply Professional ethics, accountability and equity in Engineering Profession.
9. Work effectively as a member and leader in multidisciplinary team for a
common goal.
10. Communicate effectively within a Profession and Society at large.
11. Appropriately incorporate principles of Management and Finance in one’s
own Work.
12. Identify educational needs and engage in lifelong learning in a Changing
World of Technology.
 DEPARTMENT OF COMPUTER ENGINEERING
PROGRAMME SPECIFIC OUTCOME (PSO)

1. Formulate and analyze complex engineering problems in computer

engineering (Networking/Big data/ Intelligent Systems/Cloud
Computing/Real time systems)
2. Plan and develop efficient, reliable and secure system and customized
application software using cost effective emerging software tools ethically.
 TRAFFIC SIGNS RECOGNISATION
ABSTRACT

This paper deals with object recognition in outdoor environments. In this type of environments,
lighting conditions cannot be controlled and predicted, objects can be partially occluded, and
their position and orientation is not known a priori.
The chosen type of objects is traffic or road signs, due to their usefulness for sign
maintenance, inventory in highways and cities, Driver Support Systems and Intelligent
Autonomous Vehicles.
A genetic algorithm is used for the detection step, allowing an invariance localisation to
changes in position, scale, rotation, weather conditions, partial occlusion, and the presence of
other objects of the same colour.
A neural network achieves the classification. The global system not only recognises the
traffic sign but also provides information about its condition or state.

Keywords:
Genetic algorithms,
Neural networks,
Traffic sign recognition,
Driver support systems,
Intelligent vehicles,
Intelligent transportation systems.
.
 TABLE OF CONTENT
CHAPTER NO. TITLE PAGE NO.

1. INTRODUCTION 1

2. HARDWARE SOFTWARE REQUIREMENT 2

3. DETAILED DESIGN 3

4. SCREENSHOTS OF PROJECT 7

5. CONCLUSION 8
 CHAPTER 1
INTRODUCTION:
Traffic sign recognition is a multi-category classification problem with unbalanced class
frequencies. It is a challenging real-world computer vision problem of high practical relevance,
which has been a research topic for several decades. Many studies have been published on this
subject and multiple systems, which often restrict themselves to a subset of relevant signs, are
already commercially available in new high- and mid-range vehicles. Nevertheless, there has
been little systematic unbiased comparison of approaches and comprehensive benchmark
datasets are not publicly available.
There are several different types of traffic signs like speed limits, no entry, traffic signals, turn
left or right, children crossing, no passing of heavy vehicles, etc. Traffic signs classifications is
the process of identifying which class a traffic sign belongs to. In this python project example,
we will build a deep neural network model that can classify traffic signs present in the image
into different categories. With this model, we are able to read and understand traffic signs
which are a very important task for all autonomous vehicles.

OBJECTIVE AND PROBLEM STATEMENT:

We chose this project due to its wide range of applications that a system with capabilities
provides. This is an attempt to make self learning system that can itself understand and
recognize/interpret traffic signs.

The system has to be able to detect traffic signs independently of their appearance in the image.
Because of that, it has to be invariant to:
1. Perspective Distortion
2. Lighting Changes
3. Partial Occlusions
4. Shadows

In addition, it has to provide information about the presence of possible problems:

• Lack of Visibility
• Bad condition
• Bad placement

1
 CHAPTER 2

HARDWARE SOFTWARE REQUIREMENT :

HARDWARE SPECIFICATIONS :-
1. RAM : 512 MB RAM
2. Hard Drive : 40 GB Hard Drive
3. Processor : Intel Core 2 Processor
4. camera module (Webcam)
5. Projector
6. Colour Markers

SOFTWARE REQUIREMENTS
1. Python (3.7.4 used)
2. IDE (Jupyter used)

2
 CHAPTER 3
DESIGN DETAILS :
Frameworks required are:-

1. Numpy (version 1.16.5)

2. cv2 (openCV) (version 3.4.2)
3. Keras (version 2.3.1)
4. Tensorflow (Keras uses TensorFlow in backend and for some image preprocessing)
(version 2.0.0)
5. Matplotlib (version 3.1.1)
6. Pandas (version 0.25.1)

ALGORITHM:
The algorithms used in character recognition can be divided into three categories: Image Pre-
processing, Feature Extraction, and Classification. They are normally used in sequence – image
pre-processing helps makes feature extraction a smoother process, while feature extraction is
necessary for correct classification.

Pre-processing :-
The second method called as Pre-processing is the entry method for recognition of character
and very important in deciding the recognition rate. Preprocessing works to normalize the
strokes and also remove variations that can reduce the rate of accuracy. Preprocessing mainly
works on the various distortions like the irregular text size, points missed during the pen
movement, jitters, left-right bend and uneven spaces.

3
Segmentation :
Segmentation is used to convert input image consisting of many characters into the individual
characters. The techniques used are word, line and character segmentation. It is generally
performed by dividing single characters from the word picture. Moreover, a content is
processed in a way that is tree like. In initial scenario, row histogram is used to segment the
lines. Then after, every level, characters are retrieved by technique called histogram and then
finally getting it retrieved.

Feature Extraction :
The aim of feature extraction is to allow the extraction of pattern which is most important for
the classification. Some of the Feature extraction techniques like Principle Component
Analysis (PCA), Scale Invariant Feature Extraction (SIFT), Linear Discriminant Analysis
(LDA), Histogram, Chain Code (CC), zoning and Gradient based features can be applied to
extract the features of individual characters.
All of these features are used to train the given system. Each of the segmented image is taken
of some pixel of dimension 28 * 28.
This can be performed to be as a big array of numbers. There by flattening the array into a
vector of 28*28 = 784 numbers. Thus, the image now converges to a minimal bunch of arrays
in a 784-cell dimension of a highly efficient structure. The image now becomes a tensor of n
dimensional array.

Classification :
The decision making is done in the classification phase. For recognizing the characters, the
extracted features are used. Different classifiers like SVM and Neural Networks are used. The
classifiers sorts the given input feature with reserved pattern and find the best matching class
for input, for which Soft Max Regression is used.Soft Max regression assigns the probability to
each result thus classification becomes easy. It basically first adds up all the evidences it gets
by the below formula and then convert that into the possible probabilities .

WORKING:
Traffic sign detection is usually based on the shape and color attributes of traffic signs, and
traffic sign recognition is often used with classifiers, such as convolutional neural networks
(CNNs) and SVM with discriminative features. It is not difficult for human beings to
distinguish traffic signs from a background, so for a computer detection system, color
4
information is also an important feature.

Extracting a potential traffic sign from an image:

Traffic signs are designed such that they appear unique and easily identifiable to the
human eye. Traffic signs in the United States of America are of 3 main colors: Red,
White, and Yellow. Other colors like orange and blue are also used. In our approach we
concentrate on Red, White, and Yellow traffic signs. Since the color of a traffic sign is
unique in a background we can use the color information to narrow down our areas of
interest (parts potentially containing the traffic sign). Since RGB colored images are
susceptible to variations in lighting, we use HSV (Hue, Saturation, and Variation)
images. Once we have the HSV image our next goal is to define our areas of interest
(i.e. range of Yellow, Red and White) so that we can segment our HSV image based on
these 3 colors. The next step is to use these color ranges and create binary masks for
each of the 3 colors. For Example, the red binary mask will have 0 assigned to all the
regions which are not in the red range and 1 assigned to all regions which are in the red
range. We know that traffic signs are usually occur in different closed shaped like
rectangles, triangles, diamonds etc. We can use this property to extract closed shaped
from each of the 3 binary masks. This can be done by using ‘Topological Structural
Analysis of Digitized Binary Images by Border’ [5]. We used the OpenCV
implementation of tis algorithm.

Predicting the type of Extracted traffic sign:

From the extracted areas of interests in the previous step we want to determine if it is a
sign or not and if it is a sign we wish to know what the type of sign it actually is. For
this purpose, we can train a convolutional neural network. The data used to train and
test the CNN was obtained from http://cvrr.ucsd.edu/LISA/lisa-traffic-signdataset.html.
It had about 6000 frames and 49 different types of traffic signs. For each frame, the
coordinate positions for the traffic sign in the image was given. From these positions the
traffic signs were cropped out to use for training the CNN. A CNN is basically inspired
by the connections between the neurons in the visual cortex of animals. [7]Since traffic
signs have unique shapes inside them like arrows, words, circles and so on. It is useful
to convert the traffic sign into a more useful form by using a Laplacian operation on the
traffic sign.

Now the traffic sign is fed into CNN architecture:

5
6
Once the CNN has been trained, it is used to predict the sign of the contours obtained in step 1.
Each of these contours are assigned the sign with the maximum probability which is the output
of the CNN.
As result the following table gives the Accuracy, Precision, Recall and F1 score metrics on the
test set. The test set was obtained by splitting the whole dataset into 70% train data and 30 %
validation and test data. Out of the 30% validation, 15% was the test data.

CHAPTER 4
SCREENSHOT OF PROJECT:

7
 CHAPTER 5

CONCLUSION:
The algorithm that has been used for traffic signs it can be generalized to deal with other kinds
of objects. The known difficulties that exist for object recognition in outdoor environments
have been considered. This way the system is immune to lighting changes, occlusions and
object deformation being useful for Driver Support Systems.
Due to this knowledge of the sign status, it is believed that the system is useful for other
applications such as maintenance and inventories of traffic sign in highways and or cities.

FUTURE SCOPE:
Future improvements can be made for extracting signs from test images by using advanced
segmentation methods.

REFERENCE:
[1] https://bartlab.org/Dr.%20Jackrit's%20Papers/ney/3.KRS036_Final_Submission.pdf [2]
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.695.3606&rep=rep1&type=pdf [3]
http://cvrr.ucsd.edu/LISA/lisa-traffic-sign-dataset.html [4]
http://yann.lecun.com/exdb/publis/pdf/sermanet-ijcnn-11.pdf [5] Suzuki, S. and Abe, K.,
Topological Structural Analysis of Digitized Binary Images by Border Following. CVGIP 30 1,
pp 32-46 (1985)
[6]http://docs.opencv.org/2.4/doc/tutorials/imgproc/shapedescriptors/find_contours/find_con
tours.html [7] https://en.wikipedia.org/wiki/Convolutional_neural_network

8
9
10
11