在智慧型運輸系統中,釵h關於碰撞警示和避免碰撞的系統被開發來降低意外事故發生的機率;故精確的測量出障礙物的位置及車輛本身在道路中的相對位置是非常重要的。本系統係以影像視覺為基礎,將路面影像經由裝置於車內照後鏡下方之攝影機傳入,取出影像中欲處理的區域,透過影像處理的方法,將道路標線和路面區分出來。利用一套全新的逆透視轉換公式將影像座標系統轉換回真實世界中的座標。再利用均方值誤差法把邊界偵測出的物體,將非車道標線的部分除去,僅保留車道標線的部分。從找出之車道線近似方程式和車輛之相對位置,及利用影像計算出的車速,判斷車輛是否行使於車道之安全範圍內,作警示駕駛者之用,此為本論文之主要目的。
Recently, intelligent vehicle systems have been developed to reduce and/ or avoid car collisions. Accurate detections of the objects on the road and the relative positions of vehicles to the lane are very important in the development of intelligent vehicle. This system is vision-based. The road images are captured by a camera mounted in a car. The domain of interest will be decided first from each image, and then the lane markers on roads are detected by image processing. An Inverse Perspective Mapping(IPM)is used to transform the image coordinates into world coordinates. Then edge detection is utilized to eliminate the portions of non-markers from the image and to save the portions of lane markers. Finally, the relative position between the lane and the vehicles can be determined to see if the vehicle is safe or not. If the situation is probably dangerous, then a warning is provided to the driver.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。