Real-Time Automatic Traffic Data Measurement by Deep Learning, YOLOv3 Algorithm and Moving Observer Method

This research aims to propose an automatic traffic data acquirement method denotes with the acronym MOM-DL, based on the moving observer method (MOM), computer vision, deep learning and the YOLOv3 algorithm. MOM-DL method allows the vehicles detection into a certain traffic stream and the estimation of traffic variables (flow q, space mean speed v and vehicle s density k) for two-lane undivided highways in stationary and homogeneous traffic conditions. Experiments have been conducted on the “SS624 Palermo-Sciacca” highway, in Italy. The research shows that tMOM-DL.