Detection of Drowsiness and Fatigue level of Driver

Transcription

1 IJIRST International Journal for Innovative Research in Science & Technology Volume 1 Issue 11 April 2015 ISSN (online): Detection of and Fatigue level of Driver Shreya P. Patel Madhu Sharma Bhumika P. Patel Nisha Shukla Hinaxi M. Patel Abstract Driver s inattentiveness is one of the main causes for most accidents related to vehicle crashes. Safe driving is the major concept in the societies to reduce accidents. Long distance driving cause driver fatigue due to which heavy accident occurs of vehicles like cars, aero planes, trucks etc. So it is very necessary to prevent the accidents all over the world. This paper presents the procedure of driver s drowsiness detection system which measures the drowsiness and fatigue level of driver. An initial buzzer is used to alert the driver after the system detects that the driver is in drowsy condition. In this system a camera is placed in front of driver s face in car which records the video of driver. The MATLAB will be used to detect the eye of driver that is conducted using the process of extraction of face image from the live video and in case if it detects the eye closure for certain threshold time period then it beeps the alarm to aware driver. We also describe the procedure to measure the open and close of eye so from that we can accurately detect the drowsiness and fatigue level. The main criteria of this system must be that it is highly non-intrusive and the driver should not be responsible to give any kind of feedback to system. Keywords: Inattentiveness, fatigue, drowsiness, threshold, detection system I. INTRODUCTION In today s world where Science and Technology has made amazing advances, the innovations in automobile industries over hundred years should be considered as it have made our vehicles more powerful, easier to drive and control, more-safer and environmentally friendly[1]. But the increasing number of road vehicle accidents is the crucial problem in the world. One of the most important factor of vehicle crashes is the driver fatigue and monotony. Furthermore, the accidents related to the driver s drowsiness or fatigue is more dangerous and serious than other types of accidents, as the sleepy driver do not take the correct prior action against the collision. Generally the 20% of crashes and 30% of fatal crashes are occur due to driver s drowsiness and lack of concentration. Recent statistics shows that more than 1,500 deaths and around 80,000 injuries are due to fatigue related accidents. As per the survey reports of Road Traffic Injuries (RTI) the road accident ranked fourth among the leading causes of death in the world. Nearly 1.3 million people [2] die every year on the world's roads and 20 to 50 million people suffer non-fatal injuries, with many sustaining a disability as a result of their injury. According to forecasting of statistics the number of road accident will increase to 5 million in 2020 [3]. Driver s inattention might be the result of lack of alertness while driving due to the driver s drowsiness and distraction. So the prevention of such crucial crashes is the main focus effort in the field of active safety research. The advancement in the field of computer technology has provided the means for building the intelligent drowsiness and fatigue detection system. Driver drowsiness detection system is one of the best application of intelligent vehicle system [4]. The main aim of this paper is to develop a prototype [5] of driver s drowsiness and fatigue detection system with a warning alarm. Our whole focus will be placed on designing the drowsy detection system that will accurately monitor the open and closed state of the driver s eye in real time [1]. By continuously monitoring the eyes, it can be seen that the symptoms of driver s drowsy behavior can be detected early enough to prevent the crashes. Fig.1 shows the overall flow of our system. Based on the image acquisition of video from camera which is placed in front of the driver s face perform real time processing of video stream in order to get frames from which we detect the driver s face for the further detection of eyes from it [6]. So that we can track the All rights reserved by 133

2 Detection of and Fatigue level of Driver blinks of driver from eyes and after as we get driver eyes closed in continuously 4-5 frames then we can say that driver is in drowsy state and the warning alarm beeps to alert the driver and thus it avoid crucial accidents. Fig. 1: System Flow II. PROPOSED SYSTEM A. Face Recognition: Face detection is a process that aims to locate a human face in an image. Face recognition is a necessary step in all face processing system, and its overall performance of drowsiness detection systems. To begin tracking a face, we have to first detect it. We use vision.cascadeobjectdetector to detect the face in video frame. By this command only the face is detected from video frame. The best Face detection is done by Viola-Jones method developed by Paul Viola and Michael Jones. Fig. 2: Detected face from a input video Object Detection by Paul Viola and Michael Jones can be done method like Simple rectangular features named Haar-like features, an Integral image, AdaBoost machine-learning and Cascaded classifier [16]. 1) Haar-Like Features: This feature considers rectangular regions at a location in detection window. The advantage of haar-like features is its calculation speed. Due to haar-like feature any size can be calculated at constant time. Haar-like features can be expressed by two or three joins-black and white. Rectangle feature can indicate whether the border lies between dark region and light region [16]. All rights reserved by 134

3 Detection of and Fatigue level of Driver Fig. 3: Haar-like features 2) Integral Image: Integral image is also known as summed area table. It is used to compute Haar-like rectangle feature. It is an array containing sums of pixel. Calculating a feature is very fast and efficient. The integral value of each pixel is sum of all pixel to its top left [16]. The formula for integral image is given below: ii (x, y) = where ii(x,y) is the integral image at pixel location (x,y) and i(x,y) is the original image. To compute the sum of any rectangular area is extremely efficient using the integral image. Fig. 4: Summed area of integral image 3) Adaboost-Machine Leaning: "Adaptive Boosting" is a machine learning method given by Yoav Freund and Schapire. In AdaBoost learning method, the sums of pixel in white boxes are subtracted from the sum of pixel black areas. The machine creates a set of Haar-like features [16]. B. Eye Tracking: The experimental results on eye detection are based on the assumption that eye regions eye regions from video frame are corrected located. PERCLOS Percentage of Eye Closure over time) is most useful method for measuring eye blinking. We can determine that if eyes is open, then the condition is normal and if the eyes are closed then a alarm signal is generated to alert the driver. We need to detect the eye region because we have to examine eye to know whether the person feels sleepy or not. Fig. 5: Tracking eyes from a input video C. Eye Template Creation: The eyes to be located at a proper place in the taken video frame so that the size of template is created accurate tracking. The system will track user's open eye located in template. Eye blinking is the time difference between the beginning and the end of the blink. Once the eye is located, a timer is started to generated whether the driver's eyes are open or closed. Fig. 6: Eye Template All rights reserved by 135

4 Detection of and Fatigue level of Driver D. Blink Detection: Eye blink is a quick action of closing and opening of the eyelids. Eye blink detection has a wide range of applications in human computer interface (HCI) systems. It can also be used in driver s assistance systems. This proposed method detects visual changes in eye locations using the proposed horizontal symmetry feature of the eyes. Our new method detects eye blinks via a standard webcam in real-time. We give overview of different techniques to the problem and describes best possible methods of eye blink detection techniques. The main propose system is the motion analysis method and finding frame difference used for tracking intentionally blink of eyes. We are using some Parameters to describe the blink behaviour: Blink Interval Blink Duration 1) Blink Duration: A common definition of Blink Duration is the time difference between the beginning and the end of the blink, where the beginning and end points are measured at the point where half the amplitude is reached. A better definition of blink duration is the sum of half the rise time and half the fall time in the blink complex. 2) Blink Interval: A common definition of Blink Interval is the time difference between the beginnings of two successive blinks. Fig. 7: Blink Templates III. CONCLUSION In this, a driver s drowsiness detection system has been proposed based on fatigue detection. The proposed system is based on eyes closer, blinking rate of eye detection of the driver. The system continuously captures the image of the subject on site and detects face region, then eyes are detected in the face under consideration to determine if eyes are closed or open, if eyes found to be closed for 4-5 consecutive frames or blinking rate of is found to be abnormal for long duration, for 3-4 consecutive frames then it is concluded that the subject is falling asleep or having state of drowsiness therefore fatigue is detected and a warning alarm issued. The proposed system can be applied in applications like in software industries where developers works continuously for hours to detect their vigilance level, for computer operators, operating critical operations on distant machines, hands free interaction with computational devices/machines, controlling heavy machineries like cranes etc. IV. FUTURE WORK The first and foremost improvement that can be made is to make the system work in real time where the system takes the input video directly from the camera and detect whether the driver is drowsy or not. If the driver is drowsy it gives the voice alert to the driver. If the driver is not drowsy it will continue taking the video as input and perform the calculations. The further improvement that could be made to the project is to determine a fixed threshold such that any face in any illumination i.e in both light and dark could be given as input and the blink could be detected efficiently. Another improvement that can be made is to reduce the computation time by designing an efficient and more accurate algorithm. A. Appendix: APPLICATION AREA Table -1: Various Techniques of Detection PUBLICATION AUTHOR TITLE CENTRAL IDEA YEAR Mitesh Patel, Sara For Driver s eyes detection the iris is Fatigue Detection Lal, Diarmuid detected from the eyes using binary Using Computer Kavanag, Peter MAY 2010 image. Face detection takes place using Vision Rossiter[7] skin segmentation technique. Cha Zhang and A Survey of Recent The Viola Jones face detector technique All rights reserved by 136

5 Detection of and Fatigue Level of Driver Zhengyou Zhang[10] JUNE 2010 Advances in Face Detection Huong Nice Quan[15] Hariri, B. Abtahi, S., Shirmohammadi, S, Martel, L.[19] Staszek, K. Wincza, K. ; Gruszczynski, S.[20] D.Jayanthi, M.Bommy[17] Ashish Bodanwar, Rahul Mudpalliwar, Vikrant Pawar, Kaustubh Gaikwad[5] Ijaz Khan, Hadi Abdullah and Mohd Shamian Bin Zainal[13] K.Subhashini Spurjeon, Yogesh Bahindwar[14] Mohamed A. Mohamed, A. I. Abdel-Fatah, Bassant M.El-Den[8] NOVEMBER 2010 AUGUEST 2011 MAY 2012 OCTOBER 2012 APRIL 2013 JUNE 2013 DECEMBER 2013 AUGUST 2014 Vikash, Dr. N.C. Barwar[3] MAY 2014 Detection for car assisted driver system using image processing analysisinterfacing with hardware Demo: Vision based smart in-car camera system for driver yawning detection Driver's drowsiness monitoring system utilizing microwave Doppler sensor Vision-based Realtime Driver Fatigue Detection System for Efficient Vehicle Control Drowsy Driving Detection System Efficient eyes and mouth detection algorithm using combination of Viola Jones and skin color pixel detection Tracking of Eyes to detect the Driver's Fighting Accident Using Eye Detection for Smartphones Monitoring of Driver Vigilance Using Viola- Jones Detection of and Fatigue level of Driver is used in this research paper. Viola Jones technique contains three basic idea, The integral image, classifier learning with AdaBoost, and the attentional cascade structure. The purpose of this paper is to detect drowsiness in driver accidents. The vision-based system have been widely used. Feature-based, image-based, template matching approaches are used. In this paper we will present a visionbased smart environment using in-car cameras that can be used for real time tracking and monitoring of a driver in order to detect the driver's drowsiness based on yawning detection. System is built on the top of an embedded platform, called APEX. In This sensor utilizes multiple receiver architecture providing electronically steered beam of a receiving antenna. A SSB modulation of the transmitted signal has been applied to allow for beam steering at the intermediate frequency. Investigations of the output signal have been carried out in order to detect and measure eye blink duration and eye blink frequency. In this paper the face and eyes are tracked from the captured video. In this paper Eye-tracking is done using the method of Dynamic template matching. Sensors like alcohol sensor, IR sensor, accelerometer is used for drowsiness detection and consumption of alcohol by driver. This paper presents the Viola Jones improved algorithm for face, eyes and mouth detection in the image form a input video frame. Viola Jones and Skin color pixel technique is used. Driver's fatigue and drowsiness are the major causes of accidents. PERCLOS (Percent of Eyelid Closure) technique is used. Detection is based 2 parts: 1) Based on physiological measures 2) Based on the physical measures. System is implemented in combination with android environment. Using of eyes blink rate of driver for detection. They used Haar cascade library to detect facial features. The All rights reserved by 137

6 Detection of and Fatigue level of Driver Vinay K Diddi, Prof S.B.Jamge[18] A.N.Shewale, Pranita Chaudhari[16] NOVEMBER 2014 DECEMBER 2014 Technique Head Pose and Eye State Monitoring (HEM) for Driver Detection: Overview Real Time Driver Detection System programming is done using Open CV. This paper presents visual analysis of eye index (EI), pupil activity (PA), and head pose(hp) for alertness of vehicle driver. The system uses Viola Jones algorithm which detects images. Haar-like fetaure technique is used to detect face and eye. REFERENCES [1] Singh Himani Parmar, Mehul Jajal, Yadav Priyanka Brijbhan Drowsy Driver Warning System Using Image Processing [2] Ruikar M. National statistics of road traffic accidents in. J Orthop Traumatol Rehabil [3] Vikash, Dr. N.C. Barwar Monitoring of Driver Vigilance Using Viola- Jones Technique International Journal of Application or Innovation in Engineering & Management (IJAIEM) Volume 3, Issue 5, May 2014 [4] Monali V. Rajput, J. W. Bakal, PhD. Execution Scheme for Driver Detection using Yawning Feature International Journal of Computer Applications Volume 62 No.6, January 2013 [5] Ashish Bodanwar, Rahul Mudpalliwar, Vikrant Pawar, Kaustubh Gaikwad Drowsy Driving Detection System International Journal of Engineering and Advanced Technology (IJEAT), Volume-2, Issue-4, April 2013 [6] Vandna Saini Driver Detection System and Techniques: A Review (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 5 (3), 2014 [7] Mitesh Patel, Sara Lal, Diarmuid Kavanag, Peter Rossiter Fatigue Detection Using Computer Vision INTL Journal of Electronics and Telecommunications, 2010, VOL. 56 [8] Bassant M. El-Den Fighting Accident Using Eye Detection for Smartphones Int. Journal of Engineering Research and Applications Vol. 4, Issue 8( Version 2), August 2014 [9] Dr.Suryaprasad J, Sandesh D, Saraswathi V, Swathi D, Manjunath S, "Real Time Drowsy Driver Detection Using Haarcascade Samples", PES Institute of Technology-South Campus, Bangalore, [10] Cha Zhang and Zhengyou Zhang, "A Survey of Recent Advances in Face Detection", June [11] Phillip Ian Wilson, Dr. John Fernandez-Texas A&M University Corpus Christi 6300 Ocean Dr., Corpus Christi, TX ( ), "Facial Feature Detection Using Haar Classifier", JCSC 21, 4 (April 2006) [12] Yasaman Heydarzadeh, Abolfazl Toroghi Haghighat, "An Efficient Face Detection Method Using Adaboost and Facial Parts", Computer, IT and Electronic department Azad University of Qazvin Tehran, Iran [13] Ijaz Khan, Hadi Abdullah and Mohd Shamian Bin Zainal, "Efficient Eyes And Mouth Detection Algorithm Using Combination Of Viola Jones And Skin Color Pixel Detection", University Tun Hussein Onn Malaysia, June Volume 3, No. 4 [14] K.Subhashini Spurjeon, Yogesh Bahindwar, "Tracking Of Eyes To Detect The Driver's ", Department of Electronics and Telecommunication SSGI, BHILAI, Chhattisgarh, INDIA, Volume 1, Issue 5, December 2013 International Journal of Research in Advent Technology [15] Huong Nice Quan, " Detection for car assisted driver system using image processing analysis-interfacing with hardware", Faculty of Electrical & Electronics Engineering University Malaysia Pahang, November 2010 [16] A.N.Shewale, Pranita Chaudhari, "Real Time Driver Detection System", International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, Vol. 3, Issue 12, December 2014 [17] D.jayanthi, M.bonmmy, "Vision-based Real-time Driver Fatigue Detection System for Efficient Vehicle Control", International Journal of Engineering and Advanced Technology (IJEAT) ISSN: , Volume-2, Issue-1, October 2012 [18] Vinay K Diddi, Prof S.B.Jamge, "Head Pose and Eye State Monitoring (HEM) for Driver Detection: Overview", IJISET - International Journal of Innovative Science, Engineering & Technology, Vol. 1 Issue 9, November 2014 [19] [20] All rights reserved by 138