Real-time Updated Street Navigation System

• Main Authors: Chi-Wei Lu, 呂紀緯
• Other Authors: Fay Huang Ph. D.
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/97112643706333624155

碩士 === 國立宜蘭大學 === 資訊工程學系碩士班 === 102 === In recent years, Google Street View has been used more and more frequently for understanding each county’s shops and stores locations and exploring each country’s geographic characteristic and architectural style. However, due to the high cost and time required for capturing and generating the street view images, the street view database is unable to be frequently updated. In this thesis, we develop an algorithm which uses event data recorder commonly installed in vehicles to record street view images and then update the street view database. The objective of this thesis is to develop a real-time updated street navigation system based on the images captured by event data recorders, which are commonly used nowadays and with affordable prices. Two main contributions are: first, a novel image warping algorithm is proposed to merge the captured images and the street view panoramic images from database. Since there are usually many straight-line features in street images, our image warping method is able to preserve straight lines and thus is particularly useful for such a street view application. Second, a navigation platform is implemented to allow users to navigate the streets and see the different appearances of the same location at different times. Keywords : Warping、Street View Guide System、Panoramic Image