Nowadays autonomous vehicles are becoming more and more popular, which will take the place of traditional vehicles in the near future. As an important part of autonomous control of vehicles, computer vision (image processing) is the key point this thesis mainly focused on. In this thesis we want to design an algorithm to do lane detection and obstacle detection based on the real time RGB frames acquired by the video camera installed on the vehicle. We would do experiments using simulated car model and a simulated road with two lanes (printed on A3 papers). And we are going to introduce the concept of camera calibration and one of the most commonly used computer vision libraries, OpenCV (Open Source Computer Vision Library). The results we obtained from the experiment allow us to see the performance of lane detection and obstacle detection algorithm according to different environments.
Al giorno d'oggi i veicoli a guida autonoma sono sempre piu' popolari, e sostituiranno presto i veicoli tradizionali. La computer vision (per processare le immagini), in quanto componente fondamentale del controllo dei veicoli a guida autonoma, sara' il soggetto principale di questa tesi. In questa tesi vogliamo progettare un algoritmo per il riconoscimento della corsia (stradale) e per l'aggiramento di ostacoli, basato su immagini a colori (profilo RGB) acquisite in tempo reale da una videocamera installata sul veicolo. Condurremo degli esperimenti usando modelli simulati di macchine e una simulazione di una strada a due corsie (stampata su un foglio A3). Ed andremo ad introdurre il concetto di calibrazione della video camera, oltre ad una tra le librerie piu' comunemente usate: OpenCV (Open source Computer Vision library). I risultati ottenuti dagli esperimenti ci permettono di vedere le prestazioni degli algoritmi di identificazione della corsia e di aggiramento degli ostacoli, sotto differenti condizioni.
Lane detection and obstacle avoidance for autonomous vehicles
LI, DEHUI
2017/2018
Abstract
Nowadays autonomous vehicles are becoming more and more popular, which will take the place of traditional vehicles in the near future. As an important part of autonomous control of vehicles, computer vision (image processing) is the key point this thesis mainly focused on. In this thesis we want to design an algorithm to do lane detection and obstacle detection based on the real time RGB frames acquired by the video camera installed on the vehicle. We would do experiments using simulated car model and a simulated road with two lanes (printed on A3 papers). And we are going to introduce the concept of camera calibration and one of the most commonly used computer vision libraries, OpenCV (Open Source Computer Vision Library). The results we obtained from the experiment allow us to see the performance of lane detection and obstacle detection algorithm according to different environments.File | Dimensione | Formato | |
---|---|---|---|
Thesis-LiDehui.pdf
non accessibile
Dimensione
3 MB
Formato
Adobe PDF
|
3 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/147328