Creative Commons

This dissertation is a part of the MSc in Data Science at the International Hellenic University. The main topic of this research was the identification of crosswalks through images in order to use the results for further decision making . The idea has emerged from the ever - evolving car industry, in which more and more manufacturers are leaning towards the autonomous driving. Driverless vehicles were just a dream a few decades ago, but now they are closer to become reality than ever before . There are already autonomous systems, like Cruise Control and Emergency Braking system , which give drivers the opportunity to already enjoy some privileges of driverless driving. Nevertheless , the target is to reach an autonomous level such that no human intervention will be necessary. For that reason, crosswalk identification is also a crucial aspect in autonomous driving since crosswalks are places where the possibility to meet an obstacle is really high. Thus, a system that identifies them should be implemented in order to make decisions itself such as whether to slow down or even stop the vehicle without the drivers’ help . Briefly , this dissertation contains the process of knowledge extraction from images in Matlab as well as the classification procedure in Weka , in which different classifiers have been trained and tested regarding the presence of a crosswalk. In addition, the dataset that was used in the experiments was a primary material which contained images from roads across Thessaloniki, the second largest city of Greece.