Number Plate Recognition (NPR) is a computer vision technology that captures images of vehicles using a camera. It extracts the vehicle's number plate to identify the owner's details by matching it to a database. The system works by capturing images, preprocessing them, detecting the number plate using YOLO, recognizing the characters, and outputting the results to a database. It has benefits like saving time, reducing errors, and aiding in tracking criminals. Potential future improvements include enhancing plate recognition for different fonts/sizes and speeding up the system.

1. Number Plate Recognition (NPR)
Created By : -
Abhishek Choudhary

2. Number Plate Recognition
Number Plate Recognition (NPR) is a computer vision-based
technology in which we capture the image of vehicle . After
capturing the image of the vehicle we extract the number
plate of a vehicle to identify the details of the owner of the
vehicle. This project aims to design an efficient automatic
authorized vehicle identification system by using the vehicle
number plate.

3. Requirements : -
Hardware Requirements :-
1) Camera
2) Computer
3) Infra Red Detector (for night images)
4) GPU (graphical Processing Unit)
Software Requirements :-
1) Number plate recognition software
2) Database
3) A backend system for provide the details of number of Car

4. Working : -
Working of this system can be break down in five steps which are
follows : -
1) Capturing of image from camera
2) Pre processing of image
3) Number plate detection from image (using YOLO)
4) Character recognition from plate
5) Output of character to the Database

5. Capturing and Preprocessing of the image
At first we capture the image of the vehicle from the camera and store it
into the PC. After Capturing of the image we first detect the car in the
image if there is any car present in the image than we will pass it to the
further model otherwise it will neglect the photo. If there is car in the
image than we first remove the background of the car for this we can use
any of pre trained API like (remove.bg) for removing the background of
the image and adjust the light exposure in the image . After this we pass
the image to the plate detection model.

6. Number plate detection and character
Recognition
Now as the image is passed to plate detector. Here in plate detector we
use machine learning model YOLO to detect the number plate in the
image. After this we pass the detected plate to the Optical Character
Recognition system where it will detect the character from the image
and provide the number plate of the car as a string to the Database

7. Database
As now we get the number of the car So now we can pass this to a API which will
provide all the details of the car like owner name of vehicle, model of the vehicle.
After getting all the details of the vehicle it will store these details in a entry
Database of the place So that it can be use in further time at the requirement.

8. Similar solution
1.RFID(Radio Frequency Identification):- This technology could be a variety of wireless
communication during which we have a tendency to use magnetic or electric coupling within the
frequency portion of the spectrum to unambiguously establish an individual or object.
This is a system that has primarily 2 components:1. Tags and 2.Readers.
2.DSRC(Dedicated Short Range Communication) :-This technology is based on bidirectional data
transfer radio communication between fixed gadgets which are Installed at roadside and vehicles.
3.There are some other technologies used like GNSS which determine geophysical position of a
vehicle, distance travelled on specific road for road charges and other information.

9. Outcome and Benefit of the Project
Outcome of this proposal-
1) Save time
2) Errors reduction
3) Ease in tracking the criminals
This model can be used at following places : -
1) Automatically open a gate or barrier into a secured area for authorised
members
2) Helping in finding the stolen vehicle
3) Traffic flow management.

10. Future Scope : -
1) The number plate recognition system can be enhanced to record all types of
number plates whatever their size and font type may be.
2) The woking of this model can be made faster as the technique which are
implemented at this time are slow they usually take around 2 second to provide
complete result .
3) Could detect the fake Number plate and take the image of the driver and store it
in the Database.
4) Could detect the vehicle which run fast if they are breaking the traffic rules than
we could put fine on the vehicle driver.

11. Conclusion
By number plate recognition system management, storage and sharing of
vehicle data becomes very easy and fast.
It will shorten the vehicle entry and exit time thereby hastening the process
of parking.
This system will give instant additional information about the vehicle as the
database will also contain vehicle RC information i.e registration record from
the government agencies.

12. Thank You