Helmet Deduction Using Image Processing

ABSTRACT


INTRODUCTION
The microcontroller does the massive task of embedded vision processing at very low energy consumption [4]. The project uses STM32F429 microcontroller for its processing needs that run the image processing algorithm using the images it gets from an OV2640 camera vision sensor [5]. The camera configures to output pictures in the RGB565 format at QVGA (320 x 240) resolution [6]. The helmet colour defines the threshold of the pixels in the image that is processed further [7]. This process of separating a region on interest in a digital image based on the given colour is called image segmentation [8]. Image Super-Resolution Using Wavelet Transformation Based Genetic Algorithm was proposed and described in [9]. After the picture segmentation is done, an algorithm recognises the contour of the image and its centre, which results in identifying the rider's helmet. Helmet Detection on Motorcyclists [10] proposed a system for detection of motorcyclist without a helmet. For this, we have applied the circular Hough transform and the Histogram of Oriented Gradients descriptor to extract the image attributes. Then, the Multi-Layer Perceptron classifier use and the obtained results were compared with other algorithms.

EXISTING PROBLEM
Not wearing the helmet is one of the main reasons for the high number of fatalities involving twowheeler occupants in our country. Solar panel analysis using crack detection using optimisation techniques present with fuzzy C-mean. Even though the government has made it mandatory to wear a helmet, not  The microcontroller does the massive task of embedded vision processing at very low power consumption. The project uses STM32F429 microcontroller for its processing needs that run the image processing algorithm [11] using the images it gets from an OV2640 camera vision sensor. The camera configures to output pictures in the RGB565 format at QVGA (320 x 240) resolution. The helmet colour defines the threshold of the pixels in the image that process further. This process of separating a region on interest in a digital image based on the given colour is called image segmentation. After the picture segmentation is done, an algorithm recognises the contour of the image and its centre, which results in identifying the rider's helmet.

PROPOSED SYSTEM
Automated Helmet unit incorporates earphone/speaker, inside the mouthpiece, push catch switch, GPS satellites, MP3 sound decoder, ARM Cortex-M4, GPS collector, IEEE 802.15.4/handset, miniaturised scale SD/memory and outside mouthpiece. Earphone goes about as a speaker to hear a recorded voice and mp3 tunes. The receiver is utilised to record voice or summons given by rider with the assistance of push catch switch provided in the head protector. GPS beneficiary is used to get the area of the rider through GPS satellites. IEEE 802.15.4 assists in speaking with bicycle dashboard unit. ARM Cortex-M4 controls everything and outside mouthpiece is used to screen activity sounds.

RESULT
Bike engine starts only when the helmet is brought near to bike dashboard unit. The condition ishelmet: present = engine and helmet: absent = engine: off. Hazard warning information is passed to the rider when he is a distance of 10 meters from the danger to alert him. The volume of MP3 playback is automatically adjusted to mute when critical traffic sounds are detected. Comparative study and performance analysis of automotive braking system discussed in [12], [13].

CONCLUSION
Helmet for road hazard warning designed with wireless bike authentication and traffic adaptive mp3 playback. The main aim of this project is to encourage people to wear a helmet and to prevent road accidents, which achieved. Thus traffic accidents can be limited to some extent and safety of bike riders is ensured.