整合SIFT與平面轉換整體平差於歷史航照影像自動對位

Integration of SIFT and 2D-Transformation Global Adjustment for Automatic Registration in Historical Aerial Images

饒芳如(Fang-Ju Jao) ； 朱宏杰(Hone-Jay Chu) ； 曾義星(Yi-Hsing Tseng)

《航測及遙測學刊》 22卷2期 (2017/06) Pp. 117-126

https://doi.org/10.6574/JPRS.2017.22(2).4

摘要

多數歷史航照影像通常缺乏相機資訊無法直接進行影像對位，導致無法作為後續研究使用。而以往研究文獻為解決歷史影像匹配及接對位之問題，皆需透過較耗費人力與時間的方式進行手動影像拼接。有鑑於尺度不變特徵轉換演算法所提取的特徵點具有抗影像尺度、方向及亮度改變之特性，亦可得到大量匹配點，故本研究欲利用尺度不變特徵轉換(Scale-Invariant Feature Transform, SIFT)作為影像特徵點提取，並對所有匹配點進行多重共軛點位偵測，再透過最小二乘整體平差，解算各張影像仿射轉換(Affine Transformation)參數來完成歷史航照影像對位。

關鍵字

尺度不變特徵轉換 ； 最小二乘平差 ； 歷史航照影像對位

並列摘要

Due to lack of camera information, automatic procedures of photogrammetric triangulation and image registration can't be applied to the historical images. Manual image stitching which is required to register the images is time consuming. Scale invariant feature transform (SIFT) algorithm is widely used to perform feature extraction because the SIFT feature descriptor is invariant to uniform scaling, orientation, and partially invariant to illumination changes. This study employed the SIFT algorithm to automate image matching and then used least-squares adjustment to calculate the six parameters of each image on the basis of affine transformation. Through a further study, the transformation parameters were transformed from the image coordinate system to the ground coordinate system for image registration.

並列關鍵字

SIFT ； Least Squares ； Historical Aerial Image Registration