E-CHALLAN USING AUTOMATIC

NUMBER PLATE RECOGNITION

PROJECT SYNOPSIS
OF MAJOR PROJECT (KCA-451)

MASTER OF COMPUTER APPLICATION

SUBMITTED BY
ANANTA AGARWAL (2100680140008)
NITESH TALIYAN (2100680140061)
JAI TALIYAN (2100680140038)

Year: 2022-23

SUPERVISOR DEPARTMENT HEAD

Dr. Satish Kumar Soni Dr. Kapil Kumar

DEPARTMENT OF MASTER OF COMPUTER APPLICATION

Meerut Institute of Engineering & Technology, Meerut,

(U.P.)-250005
 TABLE OF CONTENT

S Title Page Remarks

No. no.
1. Abstract 3

2. Motivation 4

3. Related Work 5

4. Proposed Work 6

5. Methodology 8

6. Facilities Required for Proposed Work 11

7. Expected Outcomes 12

8. Bibliography/References 13

Page 2
 1. ABSTRACT

Generation of E-challan using Automatic Number Plate Recognition(ANPR) is a

potential future system for effective traffic rule violation management. With the
burgeoning population and more vehicles on road, authorities find it difficult to
manage and monitor traffic violations and issue respective penalties. Traditional
methods for checking the vehicle documents manually, verifying them, monitoring
and detecting for violations not only require manpower and time but also slow down
the traffic. The objective of this paper is to discuss a method for generating automated
E challan to the traffic violators using ANPR system using image processing and
Optical Character Recognition(OCR) to read the vehicle number plate with the help
of installed cameras and CCTV systems. The vehicle registration number so detected
is looked up in the database for getting the type of vehicle and owner's information.
This information is used to generate e-challan in the name of the person who owns the
vehicle and details of challan are also sent to the owner as a text message. These
systems can operate 24*7 and therefore results in effective and easy detection of traffic
rule violation and E challan management thus providing an aid to the officials with
less user intervention. Keywords: E-Challan, ANPR ,Traffic rule violation, Image
Processing, OCR.

Page 3
 2. MOTIVATION

The main motivation for this project was the slow and inefficient traditional traffic
checking system especially in tier 2 or down categorised cities and towns. So, why
not make it automated, fast and much efficient for them in the way similar to the tier
1 or metro cities.

Recently, image processing which extracts useful information from a digital image
plays a unique role in the advent of technological advancements. It focuses on two
tasks i.e. improvement in pictorial information of human interpretation, loading of
image data for storage, transmission and representation for autonomous machine
perception. Also people have started the usage of image capturing devices never as
before with the advent of smart phones and closed circuit television.

In the same expect it is also mandatory to check for regular movement of vehicles
around us for many purposes like safety, security, authentication, etc.

In western countries it is very common to the availability of CCTV cameras all over
the road and on traffic signals.

Also, availability of all the data online anywhere may reduce the time taken and
handling the traffic to move smoothly and also reduces the human efforts.

Page 4
 3. RELATED WORK

Various cameras are installed over the traffic signals and roads to check for the vehicle
over their speed and traffic violence and they are basically surveillance cameras
which only capture for some of the vehicles break regulations.
Most of them are not automated and connected to the databases that are required to
obtain information.

Yogiraj Kulkarni, Shubhangi Bodhke, considers moving object for license plate
detection. The proposed method uses a neural network with an improved Optical
Character Recognition (OCR). This method uses K-Nearest Neighbors (KNN) for
classifying the features retrieved from region properties in an image. This system uses
a decision tree classifier with Ada-Boost for detecting a partial and complete helmet.
For achieving accuracy, they have also used CNN for classification of the pretrained
model. For subtracting the background from a video for getting moving objects,
various features of computer vision have been used like Hough transform, LBP, HOG,
SIFT, etc. For identifying geometric shapes like circles, ellipses in an image Hough
transform is used. The distribution of edge directions or intensity gradients is used for
describing the shape and its appearance of a local object within an image. For
describing the local texture structure and for capturing key-points in an image the
LBP and SIFT have been used respectively. The author gets an accuracy of 98.72%
for traffic videos which was earlier as 96.36% using Tesseract OCR on the test images
[Yogiraj Kulkarni, 2018].

MAJOR FINDINGS :
 Less availability.
 Not an up to the mark software.
 Scalability issues.

Page 5
 4.PROPOSED WORK

Automatic Number Plate Recognition (ANPR) is a system which itself is capable for
license plate detection without human intervention. It uses an image captured with
supporting illumination.

The two approaches have been used by an Automated License Plate Recognition (ALPR)
for performing license plate detection (LPD) and perform license plate recognition
(LPR). The LPD uses character segmentation and optical character recognition (OCR)
for identifying vehicle location and license plate detection. As an alternative of using a
general object detector method like Faster-RCNN or SSD300, the authors have used
segmentation free OCR evaluation approach for developing a precise system for
detecting license plates. The detection in the feature map pyramid is performed by object
detectors for detecting objects at different scales. As the deep network needs multiple
samples, the new data augmentation technique has been developed which has the
potential to train and evaluate.

The proposed method uses a recursive training method for the character segmentation
network and for detecting the license plate. The same data have been used for training
the YOLO framework. This network has proven for fast object detection in a real-time
scenario.

A recursively region-growing procedure with some heuristic approach is used by Region

Growing Algorithm (RGA). This also focuses on the symbols present on the license
plate. The background of the license plate consists of region dimensions, and position,
width, and height (aspect ratio), and colour and typestyle size. The proposed algorithm
is used for identifying the Region of Interest (ROI).

For Number plate recognition, Super resolution and OCR classification is used. Once
the license plate is detected for a particular vehicle, an email is generated and sent using
the SMTP service to the registered email-id.

Following are the steps considered in this complete process:

1. Input image - For the image to be used as input we will process image frames from
the CCTV footage and camera. The captured image will be in RGB format and it will
be passed for image processing.

Page 6
2. Image Pre-processing - The input image from the camera needs to be processed for
plate detection using closed contour. We also need to perform ratio check ok and plate
cleaning.

3. Plate Localization - Here we will be working on the number plate which is detected
using the localization technique. We will perform crop operation for character extraction
in next step.

4. Character Recognition - In this method we will be using tesseract and OCR to filter
out the text from the localized number plate. each character is detected using optical
character recognition technique. The characters are recognized individually and then are
combined to get the alphabets and numbers present on the actual number plate taken
from the image.

5. Challan Maker - The number plate characters extracted using the optical character
recognition technique are then matched with the database available for getting the details
of the vehicle and the owner information. The details so found are verified and checked
and accordingly challan will be generated for those found guilty on the basis traffic rule
guidelines and unavailability of various documents including pollution certificate
insurance certificate valid driving license and so on.

6. Web Application – User can access web application through internet. User need to
create a new account for first time visit. Registered user can login and access their
dashboard where user can find his challan by entering their vehicle registration number.
User can make payment and track their payment records. User can also raise dispute in
case there is some payment issue or invalid challan charged.

Page 7
 5. Methodology

The aim of this project is to propose a method for real time generation of e-challan
using automatic number plate recognition system. For this we will be working on
ANPR system that includes capturing image of vehicles, detecting number plate using
localization, character segmentation and at last extracting the characters from the
number plate using optical character recognition. The number plate is thus recognized
for the vehicle that breaks and violates the traffic rule and then we compare the
vehicle number of such violators from database for knowing the type of vehicle,
owner's information and verifying their details. Then instantly an appropriate fine
message is sent to the owner via SMS and e challan is generated using a web based
application.

Page 8
System Architecture is:

Page 9
Page 10
 6. Facilities Required for Proposed Work

HARDWARE REQUIREMENT:

PROCESSOR INTEL PENTIUM IV

RAM 4 GB
HDD 200 GB

SOFTWARE REQUIREMENT:

Software Type/Platform Version

Operating System Windows Windows 7 or more
Python - Python 2.7 or above
Open CV - Open CV 3.2.0 or above

Page 11
 7. Expected Outcomes

 Improves quality of images using image super-resolution techniques.

 Will detect the number plate of the vehicle.
 The rule violator can be easily detected, and the e-challan will be sent on their
registered phone number as well as on the email address.
 8. Bibliography

Reference –

 Abhay Singh, Anand Kumar Gupta, Anmol Singh, Anuj Gupta, Sherish
Johri, “Vehicle number plate detection using image processing”,).
 Jaskirat Singh, Bharat Bhushan, “Real-time automatic license plate
recognition through deep multitask network”.
 Praveen Ravirathinam and Arihant Patamari, “Automatic license plate
recognition for indian roads using faster-RCNN”.
 Ms. Sushama H. Bailmare, Prof. A.B. Gadicha, “A Review paper on
Vehicle Number Plate (VNPR) Using Improved Character
Segmentation Method” International Journal of Scientific and Research
Publications, Volume 3, Issue 12, December 2013 1 ISSN 2250-3153.
 Dinesh Bhardwaj, Sunil Mahajan, “Review Paper on Automated
Number Plate Recognition Techniques” International Journal of
Emerging Research in Management &Technology ISSN: 2278-9359
(Volume-4, Issue-5).