The aim of this diploma thesis is the study and the implementation of modern theories and methods of computer vision in a real robotic vehicle. In this framework, the robotic vehicle SRV-1 of ACS Lab was used. Its communication with a computer was achieved and the wireless transfer of commands, data, images, videos and sensor data makes possible the realization. By using the received data, the robotic vehicle is capable of navigating autonomously in the real world. It realizes obstacles and understands the visual content of images. This perception is due to the creation of visual content descriptor synthesized in real time. The descriptor's components are the local features of the image, which are in the sequel used for image retrieval. Furthermore, another method that is able to perform the robotic vehicle is automatic panorama composition.
For more details please contact Antonio Arvanitidis <antoarva<at>gmail.com>