標題: 以智慧型視覺為基礎之快速公路車輛防擦撞資訊警示系統
Intelligent Vision Based Vehicle Side Collision Warning System on Highway
作者: 梅庭瑋
Ting-Wei Mei
林進燈
Chin-Teng Lin
電控工程研究所
關鍵字: 視覺死角;電腦視覺;車側影像;智慧型運輸系統;blind spot;computer vision;ITS
公開日期: 2005
摘要: 近年來,由於車輛數目的快速成長而帶來愈來愈嚴重的交通問題,因此,世界各地皆投入了龐大的人力及物力在智慧型運輸系統( ITS )的研究上;而智慧型運輸系統最主要的研究目的就是為了發展更先進的科技來改善及提昇交通運輸系統的安全和效率。 智慧型運輸系統所涵蓋的範圍很廣泛,而先進車輛的控制和安全系統( AVCSS )是其中非常重要的一個環節;先進車輛控制和安全系統發展的主要目的,就是在避免駕駛者避免於駕駛時分心而導致車禍的發生。為了達到這個目的,車輛本身就必須具備智慧型感測的能力,因此,在先進車輛控制和安全系統的研究中,研究人員通常利用了各種的感測器來實現感測的功能。 通常駕駛者要變換車道時,常利用的是以車側後視鏡中的影像資訊做為判斷的依據,但由於駕駛者自身的疏忽和車側視覺死角區域的存在,側撞及擦撞是最容易發生的車禍類型之一,為了要避免這類型的意外發生,我們便想要發展出一套能自動偵測車側視覺死角區域是否有車輛存在的系統;在此系統中,我們利用攝影機做為偵測用的感測器,首先,會利用攝影機中的資訊去尋找路面邊線的特徵,以此定義出車側視覺死角的偵測範圍,之後便是對偵測範圍判斷是否有其它車輛存在的可能性和可能的位置,找出位置後,便利用演算法判斷出車輛與雜訊( 如:影子、道路間隙…等 ),若系統偵測出有其它車輛的存在,系統便會發出警示以提示駕駛者注意車側視覺死角區域。
In recent years, Intelligent Transportation System ( ITS ) has been researched all over the world. Because of the high growth of popularization of vehicles, ITS has become an important scope research and industrial development in order to apply advanced technologies to improve the efficiency and safety of the transportation system. ITS consists of a lot of sections, and AVCSS (Advanced Vehicle Control and Safety System) is a part of these sections. The purpose of this system is to assist to prevent the traffic accidents resulted from the negligence of drivers. In order to achieve this goal, researchers make use many kinds of sensors. Rearview mirror is usually used to assist the driver to determine if it is safe for lane-change maneuvers. But side collision accident might still happen because of the existence of the blind spot and the negligence of the driver. In order to prevent the occurrence of side collision accident resulted from the above situation, we develop a algorithm to detect objects at a moving vehicle. In our system, we use the camera as the sensor to detect if there is any other vehicle in the blind spot. At first, we make use of the feature of road to define the region of interesting (ROI ). Then, we find the area where other vehicles may exist in ROI. Furthermore, our algorithm will judge that the object is a vehicle or a noise (such as: shadow, gap, etc). Finally, if our system detects other vehicles in the ROI, the system will give some alarm to warn the driver that lane-changing is not allowed.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009312602
http://hdl.handle.net/11536/78292
Appears in Collections:Thesis


Files in This Item:

  1. 260201.pdf