Department of Information Technology

Mini Project

Automated Vehicle Registration

Recognition Using Computer Vision
Tools
Presented by
P.ABHINAV VARMA 160922737039
SIMEEN FATIMA JAMEEL 16092273703
Under the Guidance of MOHD IRFAN UDDIN 160922737003
Ms. BHARGAVI.B
Assistant Professor
 CONTENT

• INTRODUCTION
• ABSTRACT
• EXISTING SYSTEM
• PROPOSED
SYSYTEM
• TECHNICAL
APPROACH
• ALGORITHMS USED
• TECHNOLOGY USED
• SYSTEM DESIGN
• FUTURE SCOPE
 INTRODUCTION
In today’s rapidly evolving urban landscape, the number of vehicles on the road has
significantly increased, leading to various challenges in traffic management and law
enforcement. To tackle these challenges, intelligent transportation systems (ITS)
have gained importance.

A key component of ITS is Automatic License Plate Recognition (ALPR), which uses
computer vision, image processing, and machine learning to detect and
recognize vehicle license plates from images or video feeds.
This project focuses on implementing and comparing two key license plate detection
methods:
•Sobel Edge Detection
•Morphological Gradient Detection

Both techniques are implemented using Python and OpenCV, enabling automatic
detection and segmentation of vehicle license plates for further character recognition.
 ABSTRACT
This project presents the design and development of an Automated Vehicle Registration
Recognition system using computer vision and machine learning tools. The primary objective
is to accurately detect and extract license plate information from vehicle images.
Two distinct detection techniques are implemented:
Sobel Edge Detection and Morphological Gradient Detection
•Sobel helps to catch fine edge details.
•Morphological gradient smooths and enhances the plate boundaries, improving robustness.
•This can lead to better accuracy in detecting the license plate area, especially in complex or noisy
images like cars with cluttered backgrounds.

The system is developed using Python, with Flask as the web framework and MySQL for data
handling. Character recognition is achieved using a Support Vector Machine (SVM) classifier.
The solution is modular, featuring an admin panel for dataset upload, model creation, and testing
functionalities.

The proposed system achieves over 90% accuracy, demonstrating reliability and efficiency. It lays
the groundwork for future enhancements like video stream integration and real-time deployment in
smart traffic systems.
 EXISTING SYSTEM

ional traffic monitoring systems rely on manual observation or basic camera footage.
se plate detection, if present, is often inaccurate or inconsistent.
ulty in tracking vehicles in real-time or detecting unauthorized entries.
dependency on human effort for vehicle identification.
ptimized for high-traffic environments or diverse lighting/angle conditions.
ations:
ot process large-scale vehicle data automatically
ctive in handling tilted or blurred license plates
scalability and real-time responsiveness
PROPOSED SYSTEM
•An automated system using computer vision and machine learning to
detect and recognize vehicle license plates.
•Implements two robust detection methods:
•Sobel Edge Detection – effective for front-facing, clearly visible plates.
•Morphological Gradient Detection – suitable for tilted or slightly
obscured plates.
•Developed using Python, OpenCV, Flask, and MySQL.
•Features an admin panel for:
•Uploading datasets
•Model training (using SVM)
•Testing predictions
Advantages:
•High accuracy in plate detection
•Works on various image types and angles
•Easily extendable to live video streams
•Reduces human involvement in vehicle monitoring
TECHNICAL APPROACH
🔹 1. Vehicle Detection
Detect the presence of a vehicle in the input image using basic object localization
🔹 2. Image Acquisition
Capture or upload vehicle images for processing.
🔹 3. Image Preprocessing
• Convert to grayscale
• Apply thresholding
• Use filters for noise removal
• Resize and normalize image size
🔹 4. License Plate Detection
Apply either of the following techniques:
• Sobel Edge Detection
• Morphological Gradient Detection
These methods identify the license plate region in the image.
🔹 5. Character Segmentation
Segment the license plate region into individual characters.
🔹 6. Character Recognition
Use Support Vector Machine (SVM) to recognize and extract alphanumeric characters.
ALGORITHMS USED
Sobel Edge Detection for License Plate Detection
Purpose:
Detect high-contrast horizontal and vertical edges to locate license plate regions.
Workflow Steps:
1.Image Preprocessing
•Convert to grayscale
•Apply Gaussian blur to reduce noise
2.Sobel Edge Extraction
•Apply Sobel operator in X and Y directions
•Emphasizes vertical edges (common in plate characters)
3.Candidate Region Extraction
•Threshold the gradient image
•Use morphological operations to highlight plate-shaped regions
4.Bounding Box Detection
•Search for rectangular regions likely to contain the license plate
•Filter by size/aspect ratio

Advantages:
•Simple and fast
•Effective for clear, high-contrast images
•No training required
Morphological Gradient Support Vector Machine (SVM)
Detection
Morphological Gradient Detection
highlights the edges or SVM is a supervised machine learning algorithm
boundaries of objects by used for classification and regression. It works by
calculating the difference finding the best boundary (hyperplane) that
between the dilated and eroded separates data into classes.
How It Works:
versions
Purpose: of an image.
• Plots data points in n-dimensional space
•Detect sharp transitions in intensity
(based on features).
•Emphasize object outlines
•Ideal for finding license plate boundaries
• Finds the maximum-margin hyperplane
that best separates different classes.
How It Works: • Only the closest points (support vectors)
1.Dilation enlarges bright regions.
influence the boundary.
2.Erosion shrinks bright regions.
3.The difference isolates the outer edges. Use in License Plate Recognition:
Use in License Plate Detection: Classifies image regions as plate or non-
Enhances edges plate (in detection)
Isolates plate regions Recognizes individual characters (in
Removes background noise recognition)
Refines object boundaries
TECHNOLOGY USED
Programming & Development
•Python 3.8 – Core language for application logic and ML
algorithms
•Flask – Lightweight Python web framework for building
the web interface
•PyCharm – Integrated Development Environment (IDE)
used for development and debugging
🔹 Computer Vision & Machine Learning
•OpenCV – Image processing for license plate detection
•Support Vector Machine (SVM) – Classification
algorithm for character recognition
•MySQL 5.7 – Relational database to store data
•Flask-MySQLDB – Used for integrating MySQL with Flask
🔹 Web Technologies
•HTML, CSS, JavaScript – For frontend presentation and
user interaction
🔹 Application Structure
•Blueprints – Flask feature for modularizing the code for
admin and user modules
SYSTEM DESIGN

The system is designed as a web-

based application using Flask for
the interface and Python for
backend processing. When an
image is uploaded, it undergoes
preprocessing using OpenCV—this
includes converting to grayscale,
resizing, and noise removal. Next,
the system detects the license
plate using either Sobel edge
detection or morphological
gradient techniques. The plate
region is then segmented into
individual characters, which are
classified using a Support Vector
Machine (SVM) model. Finally, the
recognized license number is
displayed to the admin. The design
ALGORITHMS
FUTURE SCOPE

 Extended model to work on live video streams

 Integrate cloud-based real-time recognition for smart cities
 Improve model accuracy with deep learning (CNN)
Applications
 Automated toll collection
 Parking lot management
 Gate acess system
 Stolen vehicle detection
 Red light violition

Conclusion
IMPLEMENTATION

a
a
a
 Conclusion
The result is a method that recognizes the characters from the photo of a car taken
from a good resolution camera and prints the number of the License plate of the
car as an output. This project serves as a stepping stone for larger scale (and more
advanced) computer vision projects, such as bulk extraction of car license plate
text from large image quantities, and even application of these concepts on video
hies or live feed. Future extension of this work is to develop character recognition
using template matching algorithms. Advances technology took Automatic Number
Plate Recognition (ANPR) systems from hard to set up, limited expensive, fixed
based applications to simple mobile ones in which “point to shoot” method can be
used. This is possible because of the creation of software which ran on cheaper PC
based and also non specialist hardware in which their no need to give pre-defined
direction, angle

you a
 a n
T h u
a