SBRR MAHAJANA FIRST GRADE COLLEGE(A)

Jayalakshipuram, Mysuru-12, Karnataka, India.

Affiliated to University of Mysuru

Re-Accredited by NAAC with ‘A’ Grade, College with Potential for Excellence

DEPARTMENT OF CO MPUTER APPLICATION

PROJECT TOPIC :
Accident Prevention System and Vehicle Tracking System
for Driver Safety.

Team Members

Name: R Karthik Name: Prajwal K K Name: Darshan Srikant

1
Abstract:
Drowsy driving is a major cause of accidents on the road, and efforts to prevent it
are essential. In recent years, there has been increasing interest in the use of
technology to improve driver safety and reduce the number of accidents caused
by drowsy driving. One such solution is the Drowsiness Detection and Vehicle
Tracking System. This system uses advanced technology, such as facial
recognition, to detect signs of drowsiness in drivers. It analyses the driver's facial
features and head movements to determine if they are showing signs of fatigue or
drowsiness. If the system detects signs of drowsiness, it can alert the driver
through audio and visual signals, reminding them to take a break and rest. In
addition to detecting drowsiness, the system also uses GPS and wireless
technologies to track the vehicle's location and speed. This information is
transmitted to a central server, which can be accessed by authorized personnel. In
the event of an emergency or accident, the system can send alerts to the central
server and other authorized personnel, providing real-time updates on the
vehicle's location and status.

Introduction:
Drowsiness detection plays a critical role in driving safety as drowsy driving
poses significant risks on the road. Fatigue and drowsiness can impair a driver's
cognitive abilities, reaction time, and decision-making skills, increasing the
likelihood of accidents. According to studies, drowsy driving is comparable to
driving under the influence of alcohol in terms of impairment and crash risk.

Motivation:
The motivation behind developing a drowsiness detection system using dlib and
OpenCV stems from the critical need to address the alarming issue of drowsy
driving and its potential consequences. Drowsy driving poses a significant risk to
road safety, leading to accidents, injuries, and even fatalities. By developing an
effective drowsiness detection system, we aim to mitigate these risks and
contribute to the overall goal of reducing drowsy driving-related accidents.

2
Objectives:
➢ The objective of this project is to develop a system that can identify signs of
drowsiness in individuals such as
a. Drooping eyelids
b. Yawning
➢ Alert the driver with Alarm system to prevent accidents
➢ If the driver did not respond for Alarm system then
a. Initiate emergency slowing down of the vehicle
b. And send the location of the vehicle to the emergency number
provided

Methodology:
➢ Driver drowsiness detection using image processing typically involves the
following methodology:
a. Data Collection
b. Preprocessing
c. Feature Extraction
➢ In this methodology, we have implemented drowsiness detection using
Python, OpenCV and Dlib Library
➢ OpenCV is a popular computer vision library
➢ Dlib is a toolkit containing machine learning algorithms.

1. Data Collection:
• Obtain a dataset of images or videos of individuals exhibiting different
levels of drowsiness, such as open or closed eyes, blinking, yawning, and
nodding
• This dataset can be obtained from various sources, such as publicly
available databases or by capturing images or videos using a camera.
2. Pre-processing:
Pre-process the images by converting them to grayscale, resizing and
cropping them, and applying various image enhancement techniques such as
contrast and brightness adjustment, edge detection. This can be done using
OpenCV functions.
3. Feature Extraction:
Extract relevant features from the pre-processed images using various
techniques, such as Haar wavelets, Histogram of Oriented Gradients (HOG).
• These features can include eye opening ratio, eye aspect ratio, mouth
aspect ratio, and facial landmarks.
3
4. Facial Landmarks:
Face landmarking, defined as the detection and localization of certain key
points on the face.

5. Threshold Values For EAR, MAR:

4
6. Threshold Values For EAR, MAR:
Mouth Aspect Ratio(MAR):

Methodology for Alert System:

Block of the Proposed Methodology for Alert System

5
Proposed Methodology:

Figure. Flow Chart of the Proposed Methodology

Proposed System:
The proposed Accident Prevention System and Vehicle Tracing System is an
advanced solution aimed at enhancing driver safety and preventing accidents on
the road. Leveraging cutting-edge technologies and innovative approaches, this
system combines various components to monitor vehicle behaviour, detect
potential risks, and provide real-time alerts and assistance to drivers. Additionally,
it integrates a vehicle tracing feature to track vehicle location and enable efficient
recovery in case of theft or emergency situations.

6
Conclusion:
In conclusion, the proposed Accident Prevention System and Vehicle Tracing
System represents a significant advancement in driver safety technology. By
combining state-of-the-art sensors, intelligent algorithms, and comprehensive
monitoring features, this system has the potential to significantly reduce the
number of accidents, promote safer driving behaviours, and provide enhanced
security for vehicles.

Hardware and Software Requirements:

Hardware Requirements :-
• Arduino UNO
• Sim 800A GSM
• NEO 6M GPS Module
• Relay Module
• Geared Motor

Software Requirements :-
• PyCharm IDE
• Arduino IDE