Real time license plate detection for non helmeted motorcyclist using CNN

Project Title

Real time license plate detection for non helmeted motorcyclist using CNN

Project Area of Specialization Wearables and ImplantableProject Summary

This idea is related to real-time object detection. This will be beneficial in a variety of sectors where things must be recognised or images must be classified.

So, this is a project titled as Real-Time License plate detection for non helmeted motorcyclist using CNN (i.e. convolution neural network) architecture.It is hardware and software based project.

The main hardware part we are using in this project is a pan tilt hat module.
Pan tilt hat is a device that has two servos or motors that can help rotate from left to right as well as upward and downward. We will mount this module on Raspberry pi which is used as a microcontroller.

Alongside with this, we are integrating PI camera  and the camera will help in detecting motion or movement of the object in any direction and the camera which is mounted over pan tilt hat and integrated with raspberry pi will move in that direction along with the movement of object and will also classify the object in the frame.

This image will be classified on the bases of the data set that we will define in the code.
The coding is based on Python using tensor flow which is an open source framework for machine learning or deep neural learning.

In addition to this, we will use SSD or single shot detector to create bounding boxes.

Project Objectives

Nowadays, two-wheelers are the most popular means of transportation since they are affordable to individuals of all socioeconomic levels. As the number of motorcycle riders grows, so does the frequency of motorcycle accidents caused by reckless riding. Motorcyclists' recklessness in not wearing a helmet is a major problem, and it frequently contributes to the biker's head injury.

The study's major goal is to create an algorithm allowing non-helmeted motorcyclists to automatically detect LP (License Plate). To complete the task, a single convolutional neural network is used. The detected licence plate can be used in ANPR (Automatic Number Plate Recognition) technology to identify the LP characters, which can then be used in data analysis.The suggested technology can detect the licence plate of riders who are wearing a hood or cap.

Project Implementation Method

The implementation of the project begins by setting up the raspberry pi and then we will install the software on the system. After that we will assemble our pan tilt HAT kit by attaching the pan-tilt module and connecting the servos and also attaching the camera to the module. Further, we will connect the Pi camera between the USB and HDMI modules on raspberry Pi and then enable it. Now, we will test our Pan-tilt HAT module. If the HAT installed correctly, both the servos moves in a smooth sinusoidal motion while the test is running. Then we will test pi camera and if it is installed correctly, the footage from the camera rendered to the HDMI or composite display. Finally, we will code it and integrate the coding with the hardware to make it work.

Benefits of the Project

The benefits of this technology are:

• Helps traffic copes to generate challan against the vehicle that violates the law.
• It will also help the police in browsing and collecting data of the suspicious vehicles that are involved in crime.
•  It can also identify potential witnesses to specific incidents by identifying vehicles in the location at the time of an incident.

Technical Details of Final Deliverable

The final deliverable will have the following components:

1. A software integrated with a pan-tilt HAT module that will move in the direction of the object and detect and identify the person if he or she is helmeted aur not and help the copes to take action.
2. A software that uses an algorithm for object detection.

Final Deliverable of the Project HW/SW integrated systemCore Industry TransportationOther Industries Legal , Security Core Technology Wearables and ImplantablesOther Technologies RoboticsSustainable Development Goals Industry, Innovation and Infrastructure, Sustainable Cities and CommunitiesRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	65000
Raspberry Pi 4	Equipment	1	25000	25000
Raspberry Pi camera V2	Equipment	1	9000	9000
Pan-Tilt HAT kit	Equipment	1	20000	20000
Micro SD card 16+ GB	Equipment	1	2000	2000
Micro HDMI cable	Equipment	1	1000	1000
12	Equipment	1	1000	1000
Documentation and printing	Miscellaneous	1	4000	4000
Travel expenditure and utilities	Miscellaneous	1	3000	3000