International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 08 | Aug 2024
p-ISSN: 2395-0072
www.irjet.net
Fatigue Surveillance and Shield Pavithra Moorthy1, Thyagarajen T2, Hakesh M3 *12Formerly, Dept. of Computer Science and Engineering, Rajalakshmi Engineering College, Tamil Nadu, India
3 Student, Dept. of Computer Science and Engineering, Anand Institute of Higher Technology ,Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Traveling has become our daily essentials, like
recorded worldwide, i.e., approximately for every 25 seconds, an individual out there will experience a road crash.
oxygen, water, and food. Vehicles have evolved along with human needs and desires, as well as risk and danger. It is natural for a human to doze off while driving for a long time. Driver drowsiness or fatigue is one of the main reasons behind road accidents. Continued surveillance with the required preventive measures through an efficient model will be provided in this project. This project comprises two phases. In phase one, with the help of a camera module, a computer vision model can be used to monitor and detect the fatigue state. The proposed system incorporates a digital library. The Face Landmark Detection algorithm offered by the Digital Library is an implementation of the Ensemble of Regression Trees (ERT) and HOG descriptors for object detection. The EAR (Eye Aspect Ratio) value is determined, followed by the threshold frame count to endorse the fatigue state. Detection of drowsiness or fatigue: an alert is displayed on the screen along with an alert alarm. Simultaneously, the achieved value is uploaded in real-time to Firebase (cloud database storage). In phase two, a simulated car with a Raspberry Pi as its base control unit is established to take input from the firebase. For every specified threshold value, the actions of the car are subjected to certain changes. As a preventive measure, the motor slows down the vehicle’s wheels as the fatigue threshold increases, and finally, the car is stopped. As an emergency measure, the car is integrated with a GSM and GPS module, which in turn sends an alert message and the location of the vehicle to the nearest hospital and one’s respective emergency contact.
The current evolved vehicle generation does not have the proper system to prevent accidents caused by drowsiness. A proper fatigue monitoring system should be implemented to prevent road crashes and save lives. A model that just detects fatigue and alerts the driver alone will not be sufficient to prevent road crashes from happening. A shield system that simultaneously synchronizes with the fatigue monitoring phase to save lives and avoid road accidents must be incorporated. These phases together will form an efficient barrier to road accidents. Continuous surveillance with the help of computer vision and implementing the changes concerning the observed changes will enhance vehicle safety and ensure the prevention of road crashes. Fatigue surveillance and shield systems lay driver protection and road safety as their foundation. This model aims to receive the facial landmarks of an individual as input, which is processed to generate a result that will indicate the state of fatigue. The further proceedings of the model depend on that state and execute the necessary actions to achieve its goal. If an individual is observed to be drowsy or inattentive while driving, the drowsiness is detected through the camera, and the alert system will be triggered, mentioning that the driver is drowsy. This result will be updated in cloud storage for later. If a condition prevails where the driver remains drowsy even after the alert alarm and the indication message goes on, the momentum of the vehicle will be reduced step by step with the help of the value attained from the cloud for every threshold frame until the car halts and is prevented from collapsing. Once the car is halted, a caution text and the location of the fatigued individual will be delivered to their respective emergency contact.
Key Words: digital library, histogram of oriented gradients, eye-aspect ratio, ensemble of regression trees, GSM, GPS.
1.INTRODUCTION The evolution and growth of the automobile industry in the present period are impeccable. Increases in vehicle automation and self-sustainability are witnessed every day. Despite all the innovations, the road safety and lives of the people on board a vehicle are still intriguing. Every year, thousands of people lose their lives due to road accidents. There are many causes of accidents, for example, speeding and being drunk.
2. LITERARY SURVEY The authors of [1] propose a fundamental algorithm that detects the individual differences of a driver while a deep cascaded convolutional neural network is constructed to detect the face region and also where the eye aspect ratio concept is introduced to evaluate the drowsiness of the individual in the present frame. The use of deep cascaded CNN avoids the problem of poor accuracy in artificial feature extraction.
driving, but still, the foremost factor, especially on rural roads, is the driver’s fatigue and monotony. According to the data collected in 2017 from the World Health Organization, an approximate number of 1.25 million deaths have been
© 2024, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 541