智慧型車輛之自動視覺導航系統

Abstract

由於近年來計算機的處理速度愈來愈快，使得機器視覺（Machine Vision）的應用更為廣泛，本研究即為機器視覺應用在智慧型車輛之自動視覺導航系統。利用影像辨識技術及照相機成像原理，將道路的彎曲變化情形推測出來，以提供車輛動態系統追尋的軌跡。 然而，由於影像本身為龐大的資訊使得影像處理需耗費較長得時間，而較難應用於高速運動的控制系統中。因此，為了減少影像處理的時間，我們利用了控制理論的觀念，並進一步結合了道路模型與車輛動態模型，預測出影像中特徵物體的位置，以減輕計算機在即時影像處理上的負擔，而加快特徵抽取與辨識的時間，依此方式，我們可成功的將機器視覺應用於高速運動的控制系統中。

As the speed of computer becomes faster, the application of the machine vision is also broadened. In this work, we apply the machine vision to the automatically visual guidance of the intelligent vehicle. By the image process and the camera calibration, we can find the curvature of the road and provide this information to the vehicle dynamic system. Thus the vehicle can trace the road automatically. Since the data of the image is too large and image process takes much time. So it is hard to apply the machine vision to the high-speed control system. In order to shorten the time of image process, we use the control theorem, combine the road model and the vehicle model to predict the position of the feature object on the image. In this way, it makes possible to apply the machine vision to the high-speed control system.