1 of 22

“SkullRaksha”-Helmet detection-based vehicle with SMS alert and fall detection�

Member Details:

Rituraj Vijay Sharma, Prem Shejole, Mahesh Bhimrao Sathe

​

Team Phantom

2 of 22

  • Introduction
  • Literature Survey
  • Research Gap�Existing Systems
  • Novelty
  • Methodology
  • Performance Metrics
  • Advantages and Limitations
  • Results�Conclusion

2

  • Contents

3 of 22

INTRODUCTION:

3

  • 87% of Bike Deaths in accidents due to non-wearing of helmets.
  • Due to comfort (29%), cost (13%), perception of law (16%), and habit (22%), a significant portion of people opt not to wear helmets.
  • Excessive speed is a leading cause of motorcycle accidents.
  • Swift response crucial for timely medical aid, minimizing injury impact.

​

​

​

4 of 22

4

  • LITERATURE SURVEY

5 of 22

6 of 22

7 of 22

7

  • Research Gap

Opportunities for further accuracy improvements

Lack of integrated hardware-software systems

Lack of systems for active enforcement and rider safety

Limited real-world deployment and application

8 of 22

8

  • Existing Systems

Used RF receiver transmitter to establish connection between bike and helmet.

Few systems allow only detection of the helmet.

9 of 22

9

  • Novelty

Compact System

Good Accuracy

If a helmet is worn, the speed can exceed 25 kmph.

Real-time detection

SMS Alert and Fall Detection

10 of 22

10

  • METHODOLOGY

11 of 22

System Block Diagram

11

12 of 22

Dataset and Preprocessing:

  • Dataset consists of 720 images.
    • 504 – Training data
    • 144 – Validation data
    • 72 – Testing data
  • Every image is resized to 400x400.
  • Performing data augmentation on a dataset to increase its size to 1720 images.

12

13 of 22

Tech Stack:

  • Software
    • OpenCv(library)
    • Roboflow
    • Yolov5
    • CNN
    • OpenCageDecode(library)
    • Twilo API

​

​

13

  • Hardware
    • Rasberry Pi 3 Model B
    • External WebCam
    • Accelerometer Sensor(ADXL345)
    • DC Motor 12V
    • DC Motor Driver(L293D)
    • 12 V Rechargeable battery

​

​

​

14 of 22

SkullRaksha Flow Diagram:

14

​

1. Check if the driver is wearing a helmet.

If yes, proceed to step 2.

If no, maintain current speed.

​

15 of 22

How we evolved?

15

Before

After

SPECIFICATIONS

BEFORE

AFTER

Microcontroller

Arduino

Raspberry Pi 3 Model B

Functionality

No keyhole access

Speed limit threshold

Fall detection

​

​

SMS alert

​

​

Sending location

co-ordinates to emergency contacts

​

​

16 of 22

16

Performance Metrics

Performance Metrics

The performance metrics display readings across various x, y, and z coordinates, highlighting the instances when alerts were triggered and when they remained inactive.

Note: g(gravitational constant)

17 of 22

17

Advantages

  • Slow processing
  • No Night vision external webcam
  • VNC Server connections timeout

​

Limitations

  • Real time detection

​

  • SMS alert within 15 seconds

​

  • High accuracy: 84.04%

​

  • Cost effective

18 of 22

Results

With helmet

Without helmet

Speed limit:25kmph

Speed limit:100kmph

19 of 22

Conclusion

​

  • Can accurately detect with or without helmet.

​

  • Vehicle speed can drop to 25 km/hr if the helmet is removed during the journey or not worn at the start.

​

  • SMS Alert and Fall Detection

​

  • Accuracy achieved: 84.04%.

​

  • Fitted and made compact in external casing.

​

  • Lightweight, efficient and cost effective solution.

20 of 22

20

  • References:
  1. Dasgupta, Madhuchhanda, Oishila Bandyopadhyay, and Sanjay Chatterji. "Automated helmet detection for multiple motorcycle riders using CNN." 2019 IEEE Conference on Information and Communication Technology. IEEE, 2019.
  2. Silva, Romuere, et al. "Automatic detection of motorcyclists without helmet." 2013 XXXIX Latin American computing conference (CLEI). IEEE, 2013.
  3. Desai, Maharsh, et al. "Automatic helmet detection on public roads." International Journal of Engineering Trends and Technology (IJETT) 35.5 (2016).
  4. Saponara, Sergio, Abdussalam Elhanashi, and Alessio Gagliardi. "Real-time video fire/smoke detection based on CNN in antifire surveillance systems." Journal of Real-Time Image Processing 18 (2021): 889-900.
  5. Mazzia, Vittorio, Aleem Khaliq, Francesco Salvetti, and Marcello Chiaberge. "Real-time apple detection system using embedded systems with hardware accelerators: An edge AI application." IEEE Access 8 (2020): 9102-9114.

21 of 22

21

  1. C. A. Rohith, S. A. Nair, P. S. Nair, S. Alphonsa and N. P. John, "An Efficient Helmet Detection for MVD using Deep learning," 2019 3rd International Conference on Trends in Electronics and Informatics (ICOEI), Tirunelveli, India, 2019, pp. 282-286, doi: 10.1109/ICOEI.2019.8862543.
  2. Singh, A. K., & Tiwari, A. K. (2018). Helmet use among motorcycle riders and associated factors in a developing country: A case study from India. International Journal of Injury Control and Safety Promotion, 25(3), 322-327.
  3. Aerdehanyan, L., & Sosdian, M. E. (2016). Factors influencing motorcycle helmet use in a hot climate: A review of the literature. International Journal of Injury Control and Safety Promotion, 23(4), 380-387.
  4. Sullman, M. J., & Taylor, P. J. (1984). Peer influence and adolescent risk-taking behaviors. Pediatrics, 74(4), 436-441.
  5. Ozkan, Zehra, Erdem Bayhan, Mustafa Namdar, and Arif Basgumus. "Object detection and recognition of unmanned aerial vehicles using Raspberry Pi platform." In 2021 5th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT), pp. 467-472. IEEE, 2021.

22 of 22

THANK YOU

We would like to extend our sincere gratitude and thank you for giving us the opportunity to present our work.

​