A Vehicle Detection, Counting and Tracking system for Traffic Assessment Using Machine Vision and Aerial Video
View/ Open
Date
2020-12-15Author
Theophanous, Andri A.Publisher
Πανεπιστήμιο Κύπρου, Πολυτεχνική Σχολή / University of Cyprus, Faculty of EngineeringPlace of publication
CyprusGoogle Scholar check
Keyword(s):
Metadata
Show full item recordAbstract
Οι ακραίες φορτιστικές καταστάσεις και τυχηματικές δράσεις όπως είναι τα κρουστικά φορτία, η φωτιά και οι εκρήξεις μπορεί να οδηγήσουν σε τοπικές αστοχίες του φέροντα οργανισμού των κτιρίων και σε ορισμένες περιπτώσεις σε προοδευτική κατάρρευση της κατασκευής. Για παράδειγμα η ξαφνική απώλεια ακραίου υποστυλώματος ενός φορέα επιφέρει σημαντική αύξηση της έντασης των γειτονικών υποστυλωμάτων και δοκών που είναι πιθανό να αστοχήσουν και αυτά με την σειρά τους. Μερικά παραδείγματα είναι η κατάρρευση των Δίδυμων Πύργων στην Αμερική (11η Σεπτεμβρίου 2001), η πυρκαγιά σε πολυώροφο κτίριο στην Σανγκάη (15η Νοεμβρίου 2010), η πυρκαγιά στην Notre-Dame στο Παρίσι (15η Απριλίου 2019) κ.ο.κ. Currently, there is a strong drive within the construction and transportation industry to automate and integrate the multitude of data sources and to modernize the mechanisms for calculating and managing information. The continuous increase use of cars, as the primary means of transport, induces to the rapid increase of traffic congestion on the road network. This situation affects our lives, particularly in the urban areas, since people need, more and more, to move rapidly between different places. The results are traffic congestion, accidents, transportation delays and higher vehicle pollution emissions. A well-designed road mobility data management system is a significant improvement over the current management methods. It allows quick and efficient changes to face many anomalies on the road network. The main purpose of this thesis is to study the movement (traffic) of vehicles in roundabouts using a built-up detection, counting, and tracking system, through image and video processing. In addition, this dissertation aims to predict the possible areas of cracks in the road surface, since the vehicle tracking system detects the most common path of a vehicle moving in a roundabout, so the most commonly used area of this surface. This system generated in the Python programming language and the "OpenCV" library (Bradski, G., 2000) was mainly used. Ιt is noted that the code of this thesis focuses on videos that have been filmed by unmanned aerial vehicles, known as "Drones".