Περίληψη: | In this thesis was developed a system able to track cars moving in the street in the wrong direction. An algorithm in order to develop the tracking system was created. The algorithm’s purpose is to detect the center of each car in every frame aiming to compare the centers for each car among all the frames in order to extract the information of the car’s moving direction. The algorithm was performed with the help of an embedded system. For this purpose, I used Xilinx’s Zedboard zynq 7000 evaluation and development kit. My thesis so as the development of the system separated into main four parts: image processing, writing code in python for algorithm development, writing the code for the algorithm in c and finally performing the c code in Vitis software for the Xilinx’s Zedboard. First of all, for the system’s implementation, it’s needed to perform video processing in order to implement some procedures so that the video is suitable for the application of the algorithm. In addition, the algorithm’s main idea is the detection of the regions with white pixels so as find every car’s center. Specifically, the algorithm’s goal is to compare car’s centers to find its direction in the street. The algorithm development is realized in Python & C programming languages. For the performance of the algorithm Xilinx’s tools, Vitis for software and Vivado for hardware design have been utilized. In more details, in Vitis the C code was applied in order to perform the algorithm in the embedded system with the Xilinx’s Zedboard zynq 7000. The algorithm covers many cases with the purpose to detect exactly the limits of the vehicle and finally to be effective. For the evaluation and the operation verification of the system and finally the thesis different videos were used to perform the algorithm.
|