Automatic Engine Locking System for Drunk Drivers thru Iris Recognition Pattern
Marinela Villareal, Iris Santos, Gio Guevarra, Dan Gamaya, Rommel M. Anacan, Nelor Jane L. Agustin · 2022
This proposal uses an iris recognition system that confirms if the driver is the car owner and would also detect if he/she is drunk. The system extracts the iris images through Hough transform in MatLab to compare the input image with the stored iris image. If it matches the stored iris image, a signal will be sent through a relay circuit that would ignite the car's engine. The Hough transform is used as a standard computer vision algorithm that determines the parameters of geometric objects in an image like the lines and circle. The goal is to find the location of lines in images and compare the sizes and conditions to the reference value. In this manner, the owner's identity will be recognized, and the possible drunkenness of the driver will be determined. A bypass system is also an option that can be established if ever the iris recognition system detects that the driver is drunk and still insists on driving the car. It can also be used for emergency purposes. The results show the time (in seconds) it takes for the system to identify the owner and detect the driver's drunkenness. There were four images used, and the data shows that time of identification and detection varies approximately from 11sec to 15second. Suggestion to improve the results includes changing the camera from a CCD camera to a more advanced model. In addition, filters can be added to reduce noise. Lastly, it is recommended to use the same size and scale of the iris as the reference to have a precise and accurate scale.