computer vision based drivers drowsiness age and seat belt recognition 15055

Project TitleProject Area of Specialization Artificial IntelligenceProject Summary

Computer Vision-based Driver’s Drowsiness, Age and Seatbelt Recognition is a computer vision-based project which reduces road accidents using dashcam and the detection system of human facial expression, age, and seat belt detection through which buzzer and ignition systems start and stop. This project can be commercialized as a startup and our product can be installed in cars for assisting drivers. Like tracking, this small and indigenous product can be installed in cars and different tracking and motor companies may purchase this product for attracting customers. This project focuses on the elimination or reduction of road accidents by assisting the drivers and is a part of the Advanced Driving Assistance System (ADAS). The project provides the detection of age, drowsiness, and seat belt and alarms the buzzer if any of the conditions aren’t fulfilled. Thus providing a safe method for driving will result in a reduction in accidents

Project Objectives

• Age detection

• Seat belt detection

• Drowsiness and fatigue detection

• jetson Nano controlled stopstart Ignition

• Buzzer Alarm

• Synchronizing these all  conditions with Car

Project Implementation Method

•Collection of different images of yawning, seat-belts wearing, eye closure, different aged people from Dashcam.

•Labeling and making text files of images using labeling.

•Training of dataset on Yolo-v3 model for detection of Seat belt wearing, Age detection, and Drowsy driver.

•Testing on different scenarios and evaluation of trained model using mAP etc.

•Integrating the model with real-world hardware using Jetson Nano.

•Automatic switching of the car if any condition is not fulfilled.

Benefits of the Project

• Long distances and motorway driving can be tiring and monotonous. In recent years, drowsiness and fatigue have become the supreme reasons for causing severe road accidents worldwide

•Driver drowsiness detection is a car safety technology that prevents accidents when the driver is getting drowsy.

•Various studies have suggested that around 20% of all road accidents are fatigue-related, up to 50% on certain roads.

• This system is able to determine the driver's state under real day conditions using a dash camera.

 To make the product more user-friendly, it can be integrated with GSM for location tracking and sending driver status to remote entities. To improve model accuracy, it can be trained for more datasets with different images from a different perception

Technical Details of Final Deliverable

In the very first step, we will collect picture dataset. Then label the dataset with the help of an automatic labeling tool. Train Yolo model by installing and setting darknet. These all steps will be taken at the online platform named Google colab after that with the help of online Google colab GPU train the model and then test the model after training is done. Once these steps are done, connect the buzzer with JetsonNano and then guide the driver with the help of detected/ tested results.

softwares :

• Python3
• Darknet
• OpenCV
• Google colab
• Labellmg
• Numpy
• Matlabplotib

hardware : 

• ·Dash cam

• Nvidia Jetson nano

• SSD

• ignition system

• battery

• buzzer

Final Deliverable of the Project HW/SW integrated systemCore Industry TransportationOther Industries Medical , Manufacturing , Others , Security Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Good Health and Well-Being for People, Industry, Innovation and Infrastructure, Life on LandRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	70000
dash cam	Equipment	1	13000	13000
battery	Equipment	1	3000	3000
ignition system	Equipment	1	5000	5000
Nvidia: Jetson Nano Developer Kit B01	Equipment	1	34500	34500
SSD	Equipment	1	14500	14500