數位地圖的應用日新月異，在Google開始進行拍攝街景圖的計畫時，許多有趣的應用也隨之產生，因此我們發想一個系統讓使用者可以透過自動導航探索世界。 Google Map街景圖在這幾年普及度越來越高，使用Google街景圖衍伸出來的創意也如雨後春筍般不斷的被發想出來。其中有使用Google街景圖實現從甲地到乙地的導航影片，不過都是基於串接街景圖來完成一段導航影片，難免會造成些許的延遲而且無法與使用者做更多的互動。因此我們發想了一個應用Google 街景圖與Google Earth來實現的導航系統。使用者在網頁上輸入一組有效起訖位置後，本系統會自動切換到起點位置的街景圖，並在網頁上即時運算行進動畫，在行進間使用者可以隨時調整前進速度與前進方向。本系統主要分為兩大部分，第一部分：透過Google Map API傳回兩地間之導航資訊，把資訊傳遞給Google Earth之後再即時運算前進的位置。第二部分：由於第一部分所產生的影片會有部分圖像產生扭曲(如汽車、機車)，因此我們使用影像修補的方式，先偵測道路消失點，再依據消失點判斷道路線範圍，最後將道路線範圍內非馬路的地方修改成馬路的顏色再輸出一段影片。

Since Google launched street view with Google Maps, many impressive creations based on street view have been brought to us. Thus, we developed a system to discover the street view through an automatic navigation. As Google Map street view has been well known in recent years, some creative innovations have been developed. For instance, we can create a navigation video from point A to point B. However, it is based on displaying continuous images; therefore it cannot interact with a user and will cause some latency. To address this, we developed a navigation system based on Google Earth and Google Map. After a user submits the starting and ending addresses, our system will automatically switch to the street view of the starting point and generating a moving animation in real time. The user can change the speed and direction anytime during the navigation. Our system can be divided into two parts. First, we retrieve the navigation information from Google Map API and send it to Google Earth for animation processing. Second, as some parts of the video will be distorted (e.g. cars, bikes), we use image inpainting to inpaint these area. We first detect vanishing points on the roads and identify road area. Finally we inpaint those areas which are not parts of a road by the color similar to the road to create a video as the output.

[1] Billy Chen, Boris Neubert, Eyal Ofek, Oliver Deussen, Michael F. Cohen, Integrated Videos and Maps for Driving Direction. UIST 2009.
[2] 沈信良,曾定章,林君威,胡紹俊,劉志強, Versatile Vision Detection for Lanes and Vehicles. NCS 2007.
[3] 賴孟庭,楊傳凱, Automatic Video Generation from Google Street View Following a Path.
[4] Chi Peng, Bing-Yu Chen, Chi-Hung Tsai, Integrated Google Maps and Smooth Street View Videos for Route Planning. ICS 2010.
[5] Dorin Comaniciu, Peter Meer, Mean Shift: A Robust Approach Toward Feature Space Analysis. TPAMI 2002.
[6] Haifeng Huang, Wei Liu, Development of Three Dimensional Digital Tourism Presentation System Based on Google Earth API. ICSDM 2011
[7] Jongin Lee, John Kim, KwanHing Lee, Hidden View Game: Designing Human Computation Games to Update Maps and Street Views. WWW Companion Volume 2013
[8] Kotaro Hara,Victoria Le, Jon E. Froehlich, Combining Crowdsourcing and Google Street View to Identify Street-level accessibility Problems. CHI 2013
[9] Aaron Wetzler, Bresenham optimized for Matlab. July 2010
[10] Marcelo Bertalmio, Guillermo Sapiro, Vicent Caselles, Coloma Ballester, Image Inpainting. SIGGRAPH 2000.
[11] Manuel M. Oliveira, Brian Bowen, Richard Mckenna, Yu-Sung Chang, Fast Digital Image Inpainting. VIIP 2001

[12] B. Vidhya, S. Valarmathy, Novel Video Inpainting Using Patch Sparsity. ICRTIT 2011
[13] D.H. Ballard. 1981. Generalizing the Hough Transform to Detect Arbitrary Shapes, Pattern Recognition, vol. 13, no. 2, pp. 111-122.
[14] D. Anguelov, C. Dulong and D. Filip. 2010. Google Street View: Capturing the World at Street Level, Computer, vol. 43 .
[15] T. Gao and H. Aghajan. 2009. Lane Assignment Using Egocentric Smart Mobile Camera For Intelligent GPS Navigation. CVPR 2009.