Introduction

Why Important?

As transportation become a essential part of daily life, human is now empowered to move in high speed unprecedentedly. However, human makes mistakes, such as inappropriate driver behaviour or wrong decision-making. One major cause of car accidents is drunk-driving. With self-driving car, people can rely on the computer and reduce all these human factors accident. Driving after drinking will no longer be a life threatening behaviour with the technology. To make self-driving car feasible, it is essential to train the self-driving car understand the surroundings, and this is based on object detection system with Computer Vision.

What We Did?

Object detection system in autonomous vehicle can basically split into divisions: detecting where lanes are, figuring out if following traffic rules (e.g., traffic light, speed limit, etc.), and preventing from bumping into obstacle (e.g., pedestrians, cars, etc.). Our goal is to train a model to distinguishes the obstacles.