WRONG LANE DRIVING DETECTION BASED ON MACHINE LEARNING

Abstract

Road traffic safety has always been a hot topic for years. With many technological innovations, types of transportation are becoming increasingly diverse, suitable for many users and purposes. However, crowded traffic also leads to many safety problems when the awareness of some traffic participants is not high. One of the biggest problems is driving in the wrong lane, which is illegal and can lead to serious consequences. Therefore, the introduction of systems and software that allows identifying vehicles that are traveling in the wrong lane is necessary. Currently, the fields of Deep Learning and Machine Learning have made many advances and are always researched every day and every hour. In addition, open source software tools also allow people to build their own models for fields that require machine learning support. Based on these factors, this thesis is built and developed to identify vehicles traveling in the wrong lane, with the hope of somewhat minimizing traffic insecurity.