Drowsiness Detection System
Nowadays there has been an increase in the rate of road accidents due to drowsiness of a driver while driving which leads to enormous deadly accidents. Drivers who do not take regular breaks when driving long distances, they often run a high risk of becoming drowsy. The driver loses his total contro
2025-06-28 16:32:13 - Adil Khan
Drowsiness Detection System
Project Area of Specialization Artificial IntelligenceProject SummaryNowadays there has been an increase in the rate of road accidents due to drowsiness of a driver while driving which leads to enormous deadly accidents. Drivers who do not take regular breaks when driving long distances, they often run a high risk of becoming drowsy. The driver loses his total control over the wheel when he falls asleep and accident happens. This is because when the driver is not able to control his vehicle at very high speed on the road. According to study around one quarter of all major accidents are because of sleepy drivers in order to take rest, and from it we can analyze that drowsiness causes more road accidents than drink-driving. This project is aimed to generate a model which can prevent such accidents. It is a real-time system that can detect driver fatigue and distraction using computer vision approaches.
When, the driver is detected drowsy then the location of the car is sent to the concerned authorities and these coordinates are sent using a raspberry pi and a GPS module.
Project ObjectivesThe project objectives can be described briefly as follows:
- Live video frames will be processed using camera.
- Application will communicate to the server for drowsiness detection.
- Algorithms will be implemented to detect the sleepy behaviors.
- The detection of drowsiness using face images.
- The images are enhanced and noises are then reduced using filters.
- If such conditions are observed the system will generate certain actions like alarm.
- Server will be sent the real time location from the GPS module of the car.
- Sever then conveys the coordinates to the mobile application
- The Mobile application is used by road safety authorities.
Our system (DDS) receives an input from a live pi-camera which is placed in front of the driver’s face on the dashboard and processes the collected frames, using our raspberry pi module by live streaming of pi-camera for the detection of the state of drowsiness. Our DDS system is composed of a pi-camera, raspberry pi and an android application that continuously checks the eye of the driver to detect the eye blink duration. We use the algorithm named Viola Jones for detection of the face using the face detector that is available in the OpenCV library. We used the neural network-based eye detector that is available in the STASM library to orients the positions of the pupils. The STASM is a variation of the Active Shape Model of Coote’s implementation. We derived only the Rowley’s eye detection code for real-time speed constraints from the STASM library which is a group of neural networks that provides eye positions.
We define three conditions for the driver’s drowsiness as provided in Table 1. Considering the Caffier’s study, the normal eye blink duration is less than 400ms on average and 75ms for minimum. For this reason, we used TDrowsy=400ms and TSleeping=800ms.
| Drowsiness Level | Description |
| Awake | Blink durations < TDrowsy. |
| Drowsy | Blink durations > TDrowsy and Blink durations < TSleeping |
| Sleeping | Blink durations > TSleeping |
Drowsiness Level
Awake
Drowsy
Sleeping
Benefits of the ProjectThe Direct beneficiaries are the drivers who are driving the vehicle on long routes. They are needed to remain alert of any harmful situation that may relate drowsiness state. The indirect beneficiaries can be the corporates that perform logistics of huge materials via transport. They need to keep track of their logistics movement. So, they can be alerted of any driver in such a state.
Technical Details of Final DeliverableThe Final product is divided into three modules. Firstly we have the Detection Module that consists of a Raspberry Pi, a GPS Module and a Pi camera. This module process for detection of any drowsy driver that is driving the car and sends the live coordinates of that vehicle. The second module is an android application that is used by the road safety authorities. It will receive the alerts of any drowsy state driver as the live location received through the coordinates. The third module is a web application that is used as a central control monitoring station. It will also be used by the safety authorities. It will be used to monitor all the activities of the system.
Final Deliverable of the Project HW/SW integrated systemType of Industry IT Technologies Artificial Intelligence(AI)Sustainable Development Goals Good Health and Well-Being for People, Life on LandRequired Resources| Elapsed time since start of the project | Milestone | Deliverable |
|---|---|---|
| Month 1 | 1st Milestone (Project Initiation) | SRS, Project charter, Literature view, Gantt Chart, GUI (tentative), Data acquisition, Android application UI, chapters 1, 2, 3, Prototype. |
| Month 2 | 2nd Milestone (Process & Planning) | Algorithm implementation, Database Design, UML, Chapter 4, 5 and 6, Android Application without drowsiness system. |
| Month 3 | 3rd Milestone (Execution & Testing) | Testing Chapter 7 and 8 |
| Month 4 | 4th Milestone (Project Closing & Deliverables) | Complete Project, Chapter 9 and 10 |