影像接合及其於空拍影像應用

Image stitching and application of UAV image

吳柏漢(Po-Han Wu)

指導教授 ： 黃詰琳

長榮大學/資訊暨工程學院/資訊管理學系(所)/碩士(2016年)

摘要

無人機的技術逐漸成熟，在現代空拍機越來越頻繁出現在電視新聞或學校的教學課程裡面。早期是無人機運用在軍事用途上，而現在無人機更延伸當作空拍機來使用，如同電影看見台灣，無人機空拍可以讓我們欣賞平常看不到的風景。也能夠拿來探查地形、科學探討，而空拍影像也在災難現場發揮很大的作用，透過拍攝建築物可以加速救災的進行。有了這些空拍影像的幫助，生活會變得更不一樣，發展性也是無可限量的，但是空拍影像是相對龐大，數量很多，在運算上或者是接合上是相對不容易的，如果演算法拼接出的效果又不是很好，等同於浪費了許多時間。因此，本論文的目的是要研究一般的影像接合將其運用於空拍影像，一開始利用邊緣偵測以及角點偵測，接著擷取影像的特徵點，並且能夠順利的將特徵點匹配，進而順利完成影像的拼接。最後利用Pix4D軟體探討影像間的重疊率以及張數多寡與處理效率的影響。

關鍵字

空拍影像 ； 影像接合 ； 特徵點 ； Pix4D

並列摘要

As The UAV technology matures, in a modern, drone more frequently appeared on TV news or school curricula, previously UAV used in the military. UAV more extended used as drone now, as a movie Beyond Beauty – TAIWAN FROM ABOVE, drone allow us to appreciate rare landscape can also be used to probe the terrain, scientific inquiry, and empty UAV images also play a significant role in the disaster scene, though shot buildings, disaster relief can be accelerated. With the help of UAV images, life would be far different, and the development of UAV images is immeasurable. But UAV images is very large, a lot of quantity. On the operation or stitch is not easy, if the stitching algorithms are not very good effect, equivalent to waste a lot of time. Therefore, the purpose of this paper is to study the general image stitching of its application to UAV images, start using edge detection and corner detection, and then capture the feature point of the image, and smooth the matching feature points, Further successful completion of image stitching. In the final, explore the images overlap rate between the number of sheets and the impact of the amount of processing efficiency and utilization Pix4D software.

並列關鍵字

UAV images ； image stitching ； feature point ； Pix4D

參考文獻

[3] J. Canny , "A computational approach to edge detection" , IEEE Trans. Pattern Anal. Machine Intelligence , vol. PAMI-8 , no. 6 , pp.679 -698 , 1986.

[6] D. G. Lowe, "Object recognition from local scale-invariant features. " Proceedings of the International Conference on Computer Vision: 1150–1157. 1999.

[7] B. Herbert, E. Andreas, T. Tinne, V. G. Luc, "SURF: Speeded Up Robust Features", Computer Vision and Image Understanding (CVIU), Vol. 110, No. 3, pp. 346–359, 2008

[8] E. Rublee, V. Rabaud, K. Konolige , G. Bradski, "ORB: an efficient alternative to SIFT or SURF", IEEE International Conference on Computer Vision (ICCV)

[9] M. A. Fischler and R.C. Bolles. "Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography. " Comm. of the ACM. Vol. 24 p.381–395, 1981.