Traffic Monitor And Analyser

Project Title

Traffic Monitor And Analyser

Project Area of Specialization Artificial IntelligenceProject Summary

Construction of unplanned roads without analyzing the traffic flow, lead to traffic jams, waste of country resources and money. There is no device to analyze when a certain road needs reconstruction or expansion. Furthermore, the tracing of the bike drivers is necessary to prevent fatal accidents. Many bike riders show negligence in helmet and triple riding. They are spotted by the police officer himself. This requires a lot of man-work and uncertainty in accuracy. Some run away and some even try to assault the officer to prevent fine.

We will be providing a prototype which consists of a Camera attached to Microcontroller, connected to internet using Wi-Fi module.for anlysing the traffic flow. We will be using machine learning approach for the detection and classification of the vehicles and couting mechanism. Also for other detections too (helmet detection & triple rider counting) we will be using the machine learnng techniques. The result will be shown in the form of graphs in a site. this part will be implemented using the pyflask. languages used will be python, html, css. 

Project Objectives

Prevent the wastage of resources, due to construction of unplanned roads. We will analyse the traffic flow on the road. The analysing will be done on live video. Thus producing records of number of vehicles passing. The statistics will be displayed in the form of graphs on a dashboard. Officials can use this data to decide how much expansion is necessary, resulting in optimum use of resources.

Ensure the safety of bike riders. this will be achieved by using helmet detection. So they can be penalized and thus ensuring their safety. Triple rider detection will also fullfill the objective of bike rider safety.

Get knowledge regarding traffic in a glimpse of an eye. This will be achieved from the statistics that will be displayed on the screen.

Ensure safety of police officers and prevent assaults during persuits.

Project Implementation Method

The machine systems used for detection and classification from images and video require features. Feature extraction manually is an invincible work. So, convolutional neural networks (CNN) have obtained its importance in achieving good accuracy for image classification in few years back. CNN learns the whole image by extracting features using feature map and has proven to obtain better detection and classification. Large data with a good machine hardware is required by CNN to train from beginning. Although with these sources it also needs a proper configuration. Also, small amount of data, may overfit or underfit the data based on the given type of learning. So, transfer learning is used on CNN model, Yolov3 Darknet trained in advance on the COCO dataset. This information has encouraged in acquiring high precision in order. We have used google colab for training the custom dataset.

for the desktop website we have used pyflask. We have implemented the front end in html, css. and the back end is coded in python flask.

We will also custom train our model of yolov3 to detect the triple rider and helmet presence.

Benefits of the Project

• Prevent the wastage of resources, due to construction of unplanned roads.
• Ensure the safety of bike riders using:
  • helmet detection
  • triple rider detection
• Get knowledge regarding traffic in a glimpse of an eye.
• Ensure safety of police officers and prevent assaults.

Technical Details of Final Deliverable

We will be providing a prototype which consists of a Camera attached to Microcontroller, connected to internet using Wi-Fi module.

This project is capable of:

• Detecting the moving vehicle on the roads.
• Differentiating between different type of vehicles. (e.g. Trucks, Cars, Motorcycle)
• Detecting their Helmet by image processing and machine learning.
• Counting the number of vehicles passing from that road.
• Keeping the record of number of persons on a bike and how many of them wear helmet, type of vehicles and vehicle counter on the database and can be accessed by required authority.
• Providing monthly, semiannually, yearly statistics of the vehicles and their types on a user/admin dash board.
• Providing detailed records to civil force so they can penalize the faulty person.

Final Deliverable of the Project HW/SW integrated systemCore Industry TransportationOther Industries Finance , Manufacturing , Health , Security Core Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Good Health and Well-Being for People, Responsible Consumption and ProductionRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	15550
NodeMcu v3	Equipment	1	550	550
Aducam mini model	Equipment	1	12000	12000
Wires ,5v charger	Miscellaneous	1	3000	3000