Driver’s Drowsiness Detection Using Cameras

Abstract 

There are some causes of car accidents due to driver error which include drunkenness, fatigue and drowsiness. Hence, the system is needed which will alert the driver before he/she falls asleep and number of accidents can be reduced. In the proposed system, a camera continuously captures movement of the driver. To determine whether a driver is feeling drowsy or not the head position, eye closing duration and eye blink rate are used. Using this information, the drowsiness level is determined.

As per the drowsiness level the alarm is generated. A night vision camera is used to handle different light conditions.

Introduction

Almost all the statistics have identified driver’s drowsiness as a high priority vehicle safety issue. Drowsiness refers to feeling sleepy, tired or unable to keep your eyes open. Due to drowsiness, the driver can't concentrate while driving, eye blink rate is decreased or increased and unable to keep eyes open. Fall-asleep crashes are very serious in terms of injury severity and may result in death. Drowsiness affects mental alertness and decreases an individual’s capability to handle a vehicle safely.

A driver is unable to predict when he or she will have an uncontrolled sleep onset. The advancement in technologies develops an interest in driver’s safety and comfort, increases traffic flow and reduces accidents. This paper introduces an alerting process when the driver falls asleep. It calculates the level of drowsiness depending on the head position, and eye blinking rate. If the level exceeds the limit from a threshold, then the alarm is generated. Different sound alarms such as ‘Take a break’ or ‘Have a coffee’, etc., must be given at a particular level of drowsiness. So that before falling asleep driver will get alert.

Related Work

Khunisuth et al. discussed a paper in which drowsiness is detected based on various factors such as titling of the head, blinking of eyes and eye blink rate. Image is captured through camera then localization of head is done. It is followed by localization of eyes and then titling of head angle is detected. Different templates such as both eyes closed, right eye closed, left eye closed, and both eyes open are used to detect eyes. A drawback of this paper is that it is not steady in all light conditions.

The accuracy of 99.59% is achieved for only stable light conditions. In the proposed system, the night vision camera is used to handle different light conditions. Ahmed et al. discussed a paper in which drowsiness is detected based on only eyes. First, the location of the driver’s eyes is located and then it is decided whether the eyes are open or not. The captured image is in binary form through which the location of the edges of the face is detected. This ultimately gives the location of the eyes. Various frames are captured. If the driver’s eyes are found to be close in five successive frames, then the system assumes that the drives is feeling sleepy and an alert gets generated accordingly. The drawback of this paper is that it may happen that some people have a habit of blinking of eyes more than that the normal rate. The result which will get generated will be ultimately wrong. Only eyes cannot give accurate output to determine drowsiness.

Tadesse et al. discussed a paper that mainly focuses on facial expressions for drowsiness detection. Many previous papers are specifically focused on eye closure and blinking of the driver. The facial expressions of the driver are analysed through Hidden Markov Model. Every time presetting of window size to fixed value is needed for different parameters. Abtahi et al. discussed a paper in which drowsiness is detected on basis of yawing of the driver. A camera detects the face and eyes of the driver. After detection of the eye, the mouth is detected and then successively yawning is detected. On basis of only yawing, an alert is generated. Yawing is detected in two steps. In the first step, the yawn component is detected and in the second step, mouth location is used to verify the validity of the detected component. This paper uses a number of algorithms so they are insensitive to changes in light conditions, and skin types. Various verification techniques are used to reduce the false rate. The limitation of this paper is that it detects the drowsiness on basis of only yawning and uses number algorithms.

Saini et al. discussed a way to find drowsiness which uses ECG and EEG, Steering Wheel Movement (SWM), Local Binary Pattern (LBP), and Optical Detection. This paper also discusses eye blink-based technique, head nodding, and yawning-based technique. Driver’s Drowsiness Detection Using Image Processing 711.

3 Proposed Work

The main concept of ‘Driver’s Drowsiness Detection is to capture a driver’s face using a camera and accurately calculate the level of drowsiness. The proposed system consists of a camera pointing at the driver. A camera continuously captures images of the driver. There are main five stages of processing: The first stage is to capture an image using the camera. The second stage is the localization of the head and checking the head position. The third stage is a calculation of the eye blink rate. The fourth stage is a calculation of eye closing duration and the fifth stage is to generate the alert. At different levels of drowsiness, different alerts will get generated.

Capture Images from the Camera

The camera is used for continuously capturing video. The camera is placed in such a way so that it will capture all head movements and eye movements of the driver. Due to different light conditions, the noise can be introduced in the image. To handle different light conditions including nighttime, the night vision camera can be used. The camera will give the whole picture of the driver with background details but we are interested only in the head position area. Myron and OpenCV library is used to capture images and find out areas of interest. The captured image is an RGB image and it is transformed into a greyscale image for processing. As the image is stored temporarily and after processing it is discarded, it does not take a large space.

Find Head Position

For finding head position (Area of Interest) ‘Haar-Cascade’s algorithm is used. Myron library will remove an unnecessary portion (background details) from the captured image and give the area in which the head is present. On which ‘Haar-Cascade’ algorithm is applied. As mentioned in the algorithm JMyron library returns a vertex point of the frame (ROI) in which the head is present. Using this point, height and width of the frame, the centre point of the frame is calculated which is ultimately the centre point of the head. The standard centre point is already defined to indicate the position of the head when the driver is not feeling drowsy. The centre point of a new frame is compared with the standard centre point and the difference between these two points is calculated. If the difference is greater than the threshold value then an alarm is generated. By using the difference, a level of drowsiness can be determined. If the difference is greater then, the drowsiness level is high and if low then, the drowsiness level is low. The threshold value is approximately 100 or 150 pixels.

Eye Blink Rate Calculation

After localization of the eye, the eye blink rate is calculated. The average eye blink rate of human beings is approximately fifteen to twenty times in one minute. For comparing with the current eye blink rate two values for threshold are set.

  • High_threshold_binkrate 25
  • Low_threshold_binkrate 10

The system captures the eye blink rate for one minute. If the current eye blink rate is greater than high_threshold_value or less than low_threshold_value then an alarm is generated. Template matching is not used to detect eyes are open or not. If the eye blink rate is too fast or low then, the drowsiness level is high.

Eye Closing Time Calculation

The system captures eye closing time by checking successive frames. If the eye closing time is greater than the threshold value, an alarm is generated. The approximate threshold value is 3s.

Alarm Generation

This module is responsible to alert drivers. Depending upon the level of drowsiness different alarms are generated. This can be a buzzer or voice message such as ‘You need a break’, ‘Take a cup of coffee’ etc. Another way is to include hardware like a vibrator to alert the driver.

Conclusion

In this paper, the way to identify the level of drowsiness is specified and according to that different alarms can be given to the driver. The night vision camera is used to 714 P. Gilbile et al. handle different light conditions. The algorithm to find the head position is specified. In the algorithm, the drowsiness level is determined depending upon head position, eye blink rate, and eye closing duration. The proposed system detects drowsiness levels and helps a driver to stay awake while driving. This system will reduce car accidents to great extent.

References

  1. Ahmad R, Borole JN (2015) Drowsy driver identification using eye blink detection. Int J Comput Sci Inf Technol. 6(1):270–274
  2. KhunpisuthO, Chotchinasri T, Koschakosai V,HnoohomN (2016)Driver drowsiness detection using eye-closeness detection In: Signal-Image Technology& Internet-Based Systems (SITIS), 2016 12th International Conference on, pp. 661–668. IEEE, 2016
  3. ParmarSH, JajalM,BrijbhanYP (2014)Drowsydriverwarning systemusing imageprocessing. Int J Eng Dev Res, IJEDR1303017
  4. Kuo Y-C, Hsu W-L (2010) Real-time drowsiness detection system for intelligent vehicles. Proceedings of the 5th Symposium on Smart Life Science and Technology (Part 1)
  5. Ahmed J, Li J-P, Khan SA, Shaikh RA (2015) Eye behavior based drowsiness detection system In: Wavelet Active Media Technology and Information Processing (ICCWAMTIP) 2015 12th International Computer Conference on, pp. 268–272. IEEE, 2015
  6. Tadesse E, Sheng W, Liu M (2014) Driver drowsiness detection through hmm based dynamic modelling In: Robotics and Automation (ICRA) 2014 IEEE international conference on robotics and automation (ICRA), pp. 4003–4008. IEEE, 2014
  7. Abtahi S, Hariri B, Shirmohammadi S (2011) Driver drowsiness monitoring based on yawning detection In: Instrumentation and Measurement Technology Conference (I2MTC), pp. 1–4. IEEE, 2011
  8. Saini V, Saini R (2014) Driver drowsiness detection system and techniques: a review. Int J Comput Sci Inf Technol. 5(3):4245–4249
  9. Pamnani R, Siddiqui F, Gajara D, Gupta A, Pandya K Driver drowsiness detection using haar classifier and template matching. Int J Adv Res Eng Technol 3(IV), April ISSN 2320–6802
  10. Nguyen TP, Chew MT, Demidenko S (2015) Eye tracking system to detect driver drowsiness In: Automation, Robotics and Applications (ICARA), 2015 6th International Conference on, pp. 472–477. IEEE, 2015
  11. Assari MA, Rahmati M (2011) Driver drowsiness detection using face expression recognition In: Signal and Image Processing Applications (ICSIPA), 2011 IEEE International Conference on, pp. 337–341. IEEE, 2011
  12. Flores MJ, Armingol JM, de la Escalera A (2010) Real-time warning system for driver drowsiness detection using visual information. J Intell Robot Syst 59(2):103–125
29 April 2022
close
Your Email

By clicking “Send”, you agree to our Terms of service and  Privacy statement. We will occasionally send you account related emails.

close thanks-icon
Thanks!

Your essay sample has been sent.

Order now
exit-popup-close
exit-popup-image
Still can’t find what you need?

Order custom paper and save your time
for priority classes!

Order paper now