A Lightweight Augmented Reality System to See-through Car

• Main Authors: Bing-Jie Yuan, 袁炳杰
• Other Authors: Yan-Ann Chen
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/hm435c

碩士 === 元智大學 === 資訊工程學系 === 107 === Internet of Vehicles (IoV) enables advanced driver assistance systems (ADAS) to acquire the sensing information of nearby vehicles for driving safety. One of the most dangerous driving maneuvers is overtaking, where the leading vehicle may occlude the view of the following vehicle’s driver. To assist the driver, providing the capability of see-through car is a promising solution. ADAS with IoV can obtain the videos from the leading vehicle to compensate the occluded view. Therefore, we formulate an image synthesis problem and propose a lightweight solution. Compared to other works, we expect that our system can easily apply on a mobile device with a camera or a smartphone. For implementation, we utilize Unity with Vuforia to make pose estimation, rather than GPS or 3D reconstruction. We also validate the performance of our system by the prototyping result. It shows that the stitching performance is well and we can achieve this AR system with low costs at price and computation.