Detekce obsazenosti parkoviště ze stereo obrazů pomocí metody grafových řezů

Abstract

The aim of this work is to study and to try possibilities of using Graph Cuts for detection of Parking Lot Occupancy from Multiple Images and creating an application using this method. Gaph Cuts are used to minimize energy function from stereo matching. There are described algorithms solving the graph cuts and the steps that are needed to achieve the goal in this work. The work also describes the necessary pre-processing images and its implementation. Programming language used here is C++ and used libraries are OpenCV 2.4.3 and gco-v3.0. The resulting algorithm is tested on images taken with the camera above the parking lot VŠB in Ostrava-Poruba.